| Author |
Message |
|---|
 RogerVolunteer developer
Volunteer tester
Send message
Joined: 27 Nov 11
Posts: 1137
ID: 120786
Credit: 267,535,355
RAC: 0
                    
|
|
Welcome to the Winter Solstice Challenge
The Final Challenge of the 2018 Challenge series is upon us. The Winter Solstice Challenge will be a 10 day celebration on account of the winter solstice. The challenge is being offered on PrimeGrid's Generalized Fermat Prime Search n=21 application. Come join us and warm yourself by the Solstice fires!
The Winter Solstice...a predictable event in nature that has occurred for billions of years. A way point in the cyclic motion of the Universe. Many of the today's major winter festivals and celebrations can trace their origins back to this event...the longest night of the year (or the shortest day).
We'd like to wish everyone a Happy Holiday...whatever festival or celebration you may observe.
NOTE: 88% of the world's population lives in the Northern Hemisphere; therefore, we'll observe winter. However, the Southern Hemisphere will be experiencing the Summer Solstice.
This GFN-project is part of the challenge:
GFN-21
These GFN-projects are NOT part of the challenge:
GFN-15
GFN-16
GFN-17 LOW
GFN-17 MEGA
GFN-18
GFN-19
GFN-20
GFN-22
To participate in the Challenge, please select only the (GFN-21) project in your PrimeGrid preferences section. The challenge will begin 11th December 2018 22:23 UTC and end 21st December 2018 22:23 UTC. Only work units issued AFTER the start time and received BEFORE the finish time will be counted.
Application Builds
IMPORTANT: Overclocking -- including factory overclocking -- on Nvidia GPUs is very strongly discouraged. Even if your GPU can run other tasks without difficulty, it may be unable to run GFN tasks when overclocked.
Supported platforms:
- Windows: Nvidia GPU (OpenCL): 32 bit, AMD/ATI GPU (OpenCL): 32 bit, CPU: 64 bit, 32 bit
- Linux: Nvidia GPU (OpenCL): 32 bit, 64 bit, AMD/ATI GPU (OpenCL): 32 bit, 64 bit, CPU: 64 bit, 32 bit
- Mac: Nvidia GPU (OpenCL): 64 bit, AMD/ATI GPU (OpenCL): 64 bit, CPU: 64 bit, 32 bit
A Cautionary Reminder
ATTENTION: The primality programs (both GPU and CPU) are computationally intensive; so, it is vital to have a stable system with good cooling. It does not tolerate "even the slightest of errors." Please see this post for more details on how you can "stress test" your CPU, and please see this post for tips on running GFN on your GPU successfully.
As with all number crunching, excessive heat can potentially cause permanent hardware failure. Please ensure your cooling system is sufficient.
It is strongly suggested that you run at least one Work Unit before the challenge to make sure your computer can handle it. Note that "Overclocking" includes GPUs that come overclocked from the factory, and lowering clock rates to the reference speeds is sometimes necessary. In the case of one particular GPU, the GTX 550 Ti, lowering the memory clock below the reference speed is often necessary.
WU's are currently averaging 11 days on CPU and 19 hours on GPU. For a general idea of how your GPU stacks up, you can have a look at the fastest GPU list.
If your CPU is highly overclocked, please consider "stress testing" it. Overclocking your GPU is not recommended at all for OpenCL. Sieving is an excellent alternative for computers that are not able to run Genefer. :)
Please, please, please make sure your machines are up to the task.
Multi-threading optimisation instructions
Those looking to maximise their computer's performance during this challenge, or when running GFN in general, may find this information useful. The latest release of genefer provides multithreading and performance improvements on the CPU for n=21. The client will download the new app automatically, no action is required on your part unless you are using app_info.xml (if you don't know what this means, don't worry). The GPU app versions remain unchanged.
- Your mileage may vary. Before the challenge starts, take some time and experiment and see what works best on your computer.
- If you have an Intel CPU with hyperthreading, either turn off the hyperthreading in the BIOS, or set BOINC to use 50% of the processors.
- If you're using a GPU for other tasks, it may be beneficial to leave hyperthreading on in the BIOS and instead tell BOINC to use 50% of the CPU's. This will allow one of the hyperthreads to service the GPU.
- Use GFN's multithreaded mode. It requires a little bit of setup, but it's worth the effort. Follow these steps:
Time zone converter:
The World Clock - Time Zone Converter
NOTE: The countdown clock on the front page uses the host computer time. Therefore, if your computer time is off, so will the countdown clock. For precise timing, use the UTC Time in the data section to the left of the countdown clock.
Scoring information
Scores will be kept for individuals and teams. Only work units issued AFTER December 11th 22:23:00 UTC and received BEFORE December 21st 22:23:00 UTC will be considered for credit.
In this challenge we will be using the same scoring algorithm as we use to determine BOINC credits.
At the Conclusion of the Challenge
We kindly ask users "moving on" to ABORT their WU's instead of DETACHING, RESETTING, or PAUSING.
ABORTING WU's allows them to be recycled immediately; thus a much faster "clean up" to the end of a challenge. DETACHING, RESETTING, and PAUSING WU's causes them to remain in limbo until they EXPIRE. Therefore, we must wait until WU's expire to send them out to be completed.
Please consider either completing what's in the queue or ABORTING them. Thank you. :)
For those of you who wish to continue to help with GFN sieving, we are currently sieving n=18, n=19, n=21 and n=22. Please see New automated manual sieving system for more information.
More information on Generalized Fermat Numbers and the Genefer program
PrimeGrid Challenge Series Results --- Last update: From Halloween (Oct 2018)
Good luck!
____________
|
|
|
|
|
|
hi
You say that GFN-21 tasks are currently averagiing over 11 days: whiich of course implies an "average" CPU will not reach the scoreboard for this challenge.
Of course "average" means average of those CPUs that have recently returned work, and an average over their recent configurations, including the use (or not) of multithreading.
Do you have a feeling for how many in that average would have been using multi-threading as described in your post?
I am inferring that with 4 CPUs multi-threaded you expect a better than /4 on the task durations (because otherwise throughput would be better with separate tasks). Or were you suggesting multitasking simply to make the tasks fit into the 10day constraint?
Many thanks for your thoughts on the above.
My aim is to get onto the scoreboard at all - so will be running just one task and allow it to run to completion even if it misses the challenge deaadline.
R~~
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
RogerVolunteer developer
Volunteer tester
Send message
Joined: 27 Nov 11
Posts: 1137
ID: 120786
Credit: 267,535,355
RAC: 0
|
hi
You say that GFN-21 tasks are currently averagiing over 11 days: whiich of course implies an "average" CPU will not reach the scoreboard for this challenge.
A lot of people, like me, haven't run GFN-21 on their CPU, like ever. So I have no idea if I'd be able to complete a task per core on the CPU. I am going to try multithreading in the next few days to see what the performance is. Maybe 2 cores per task is better than 5 cores per task, I'll find out before the challenge so I have a plan in place. For me most tasks will be by GPU, so CPU tasks will be icing on the cake.
There is more discussion on this new GFN-21 CPU multithreading feature in the thread: Genefer 3.3.4 now live . Have a read for more in depth discussion. Great that this development has happened before the challenge. Can we find a World Record GFN Prime?
Leading edge currently at b=304 ,082. Only 58,004 WU's completed. |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
                   
|
hi
You say that GFN-21 tasks are currently averagiing over 11 days: whiich of course implies an "average" CPU will not reach the scoreboard for this challenge.
First of all, everyone's computer is different, so just go and test it yourself. You don't need to run an actual BOINC task. Just get the genefer program and run it from the command line. After a short while it will tell you how long it will take to run, and its estimates are *VERY* accurate. If it's estimating less than 240 hours, it will be able to complete the task within the challenge. Be sure to test in the same configuration as you'll be actually be running during the challenge, since it matters what else is running.
If you want a general rule of thumb for what should work:
Divide CPUs into two classes: Those that have at least AVX, and those that do not. If your CPU does not have AVX, it can't use multi-threading, and is also probably an old and slow CPU, and I'm guessing it won't make it.
If it has AVX (or FMA), and you use multithreading (remember, it's -nt #, not -t #), and it's not a terribly slow computer, I'm guessing you can get at least one task done. My 4 core dsektop Haswell can do a GFN21 in about 38 hours using 4 cores and about 44 hours using 3 cores. Obviously, running on a 2 core machine will take longer, as will running on a laptop or other low power CPU. I think AMD CPUs with AVX do count as having AVX, so they should be able to use multi-threading, but per-core they will be a lot slower than Intel.
Bottom line: AVX or better and using multithreading: very likely. Otherwise, probably not.
P.S. Multi-threading is only usable with GFN-21, so there are not a lot of people who have real experience using it in a live environment.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Thanks Michael,
Both your "try it" and "rule of thumb" tips are VERY helpful, as usual.
Due to the rule of thumb I will exclude the Vista era laptop from the "try it" test ;)
Warmly
R~~
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
|
|
|
I did run the multi-threaded CPU version on two computers.
4-threaded task
3-threaded task
Only 3/4 of cores were utilized and other stuff run on the machines, so results are sub-optimal. Still something.
|
|
|
|
|
I think AMD CPUs with AVX do count as having AVX, so they should be able to use multi-threading, but per-core they will be a lot slower than Intel.
Based on LLR, a current AMD core is about half the performance of an Intel core. I have no test data yet how it might scale. Suppose I should do it at some point soon... |
|
|
|
|
|
Well, genefer isn't LLR :) I just done quick tests on 4 systems using command -q "304208^2097152+1" -nt x
Test value was leading edge at the time I looked. Following CPUs are at stock clocks (varies depending on cores active), but have faster ram than officially supported. I did test fewer than max threads per core, and saw a gradual drop in efficiency except in one case I'll get to later. Results below are for running -nt with a value matching the number of cores. Power is that reported in software averaged over 30 seconds minimum. Intel CPUs were pretty stable, the Ryzen CPU varied more.
6700k, 24h, 68W
5930k, 25h, 117W
8086k, 26h, 76W
2600, 30h, 61W average, 75W peak
I wouldn't have guessed it, but the 6700k is my fastest tested system even though it has fewest cores. It does have the fastest ram of these systems, running dual channel dual rank 3200.
The 5930k is showing its age, performing similarly but at much more power cost. Due to system limitations this was only running 2666 quad channel ram.
The 8086k was the only one showing limiting of some form, with 4 to 6 cores performing essentially the same. I wonder if the dual channel single rank 3000 ram is limiting hard? Even if I drop to 4 cores, it is higher clocked than the 6700k but still runs slower. Running 4 cores power drops to 69W, so essentially the same as 6700k at that point but slightly lower performance.
Ryzen 2600 was better than I'd expect, not that far behind the other systems. Then again, it is ball park comparable to the 6700k/8086k at 3 cores, so the factor of two is still there. This was also running dual channel single rank 3000 ram (low latency B-die).
For comparison, my RTX 2070 does 3 units in just over 24h, and takes reported 150W constantly. That puts it a fair bit more power efficient than CPU. I don't have data for older generation GPUs yet. |
|
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 644
ID: 164101
Credit: 305,010,093
RAC: 0
|
|
Winter Solstice Challenge warm-up lap: a new CPU record!?
Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz - 17 threads
27,416 sec = 7 h 36 mn 56 sec
https://www.primegrid.com/result.php?resultid=951710455 |
|
|
|
|
Winter Solstice Challenge warm-up lap: a new CPU record!?
Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz - 17 threads
27,416 sec = 7 h 36 mn 56 sec
https://www.primegrid.com/result.php?resultid=951710455
Wow |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
                  
|
Winter Solstice Challenge warm-up lap: a new CPU record!?
Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz - 17 threads
27,416 sec = 7 h 36 mn 56 sec
https://www.primegrid.com/result.php?resultid=951710455
Wow, think that's twice as fast as my fastest gpu!
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
Scott BrownVolunteer moderator
Project administrator
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 05
Posts: 2165
ID: 1178
Credit: 8,777,295,508
RAC: 0
                                    
|
Winter Solstice Challenge warm-up lap: a new CPU record!?
Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz - 17 threads
27,416 sec = 7 h 36 mn 56 sec
https://www.primegrid.com/result.php?resultid=951710455
That's a little faster (by 800 seconds) than my RTX 2070 FE!
|
|
|
SkyFallVolunteer tester
Send message
Joined: 26 Mar 15
Posts: 13
ID: 388081
Credit: 56,709,166
RAC: 0
              
|
|
Looks like the GFN21 (manual-)sieve is complete.
Are the "results" of the sieve automatically incorporated into the GFN21 subproject? |
|
|
|
|
|
If I only use this I get the error;
Your app_config.xml file refers to an unknown application 'cpuGFN21'
<app_config>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
But this does not get the error
<app_config>
<app>
<name>genefer</name>
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.1</cpu_usage>
</gpu_versions>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app>
</app_config>
BTW the RTX 2080 Ti Run Time Less that 5 Hours
https://www.primegrid.com/result.php?resultid=951772295 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
If I only use this I get the error;
Your app_config.xml file refers to an unknown application 'cpuGFN21'
<app_config>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
But this does not get the error
<app_config>
<app>
<name>genefer</name>
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.1</cpu_usage>
</gpu_versions>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app>
</app_config>
BTW the RTX 2080 Ti Run Time Less that 5 Hours
https://www.primegrid.com/result.php?resultid=951772295
Does that second one actually work and run Genefer with 4 threads? It's not supposed to work according to the BOINC documentation. Specifically, the <app_version> block is supposed to be outside of the <app> block.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Now I get this error "Your app_config.xml file refers to an unknown application ''"
Like this then?
<app_config>
<app_version>
<app>
<name>genefer</name>
<app_name>cpuGFN21</app_name>
<cmdline>-nt 12</cmdline>
<avg_ncpus>12</avg_ncpus>
</app>
</app_version>
</app_config>
I still only show 7% CPU Usage |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
If I only use this I get the error;
Your app_config.xml file refers to an unknown application 'cpuGFN21'
<app_config>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
But this does not get the error
<app_config>
<app>
<name>genefer</name>
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.1</cpu_usage>
</gpu_versions>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app>
</app_config>
BTW the RTX 2080 Ti Run Time Less that 5 Hours
https://www.primegrid.com/result.php?resultid=951772295
Does that second one actually work and run Genefer with 4 threads? It's not supposed to work according to the BOINC documentation. Specifically, the <app_version> block is supposed to be outside of the <app> block.
Like this then?
<app_config>
<app_version>
<app>
<name>genefer</name>
<app_version>
<app_name>cpuGFN21</app_name>
<cmdline>-nt 12</cmdline>
<avg_ncpus>12</avg_ncpus>
</app_version>
</app>
</app_version>
</app_config>
No.
The <app> and <app_version> sections are separate. They're supposed to follow each other rather than be inside of each other. Furthermore, there's actually no need for the <app> section. All you need is the <app_version> part. This is what I use:
<app_config>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
EDIT: That's for 4 cores. Change the numbers as necessary.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Working now, thank you Michael
Using the Below for One CPU Task and 2 GPU
<app_config>
<app>
<name>genefer</name>
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.1</cpu_usage>
</gpu_versions>
</app>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config> |
|
|
|
|
Winter Solstice Challenge warm-up lap: a new CPU record!?
Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz - 17 threads
27,416 sec = 7 h 36 mn 56 sec
https://www.primegrid.com/result.php?resultid=951710455
W...O...W...!
|
|
|
|
|
|
One tip I would add to Michael's advice on using the command line to run genefer to get the time estimates.
tl;dr: run the test with a system that has been warmed up running genefer under the same number of cores for at least five minutes, then kill the task and immediately restart.
Detail
At first I found that the results were very erratic -- the same command line could get an estimate of 83 hours or 118 hours and various figures between. I was wondering how reliable those estimates were and yet Michael, who is usually reliable, had said they are "surprisingly accurate".
The solution is that genefer is not only sensitive to what is running, it is also very sensitive to what HAS BEEN running recently before it makes the estimate. My guess is that this is because my system is undercooled for this work and the cpu is slowing itself down to protect itself. So starting with a cold system I get an estimate of 83 hours because the system is still running at full speed when genefer makes its estimate. Running genefer for 5mins then killing it and immediately restarting I get an estimate of 118 hours consistently running under LinuxMint, or 113 hours under Debian 9. These figures seem more likely to pan out into the real result, in my opinion and it will be interesting if that is reflected in the challenge: I have not had time to run a genefer to completion on it, but I have checked that boinc runs on 2 cores OK.
My tests also showed that using 2 real cores is good, using 3 or 4 cores (including hyperthreads) is unhelpful on that machine. No surprises there then.
This is not only adequate, it is sufficient that I am now reasonably hopeful to get two tasks onto the scoreboard from that machine, and a borderline maybe-will maybe-wont from another.
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 916
ID: 107307
Credit: 974,514,092
RAC: 0
     
|
Looks like the GFN21 (manual-)sieve is complete.
Are the "results" of the sieve automatically incorporated into the GFN21 subproject?
Yes, of course they are |
|
|
|
|
|
I advised
tl;dr: run the test with a system that has been warmed up running genefer under the same number of cores for at least five minutes, then kill the task and immediately restart.
and please see the message this is replying to for the longer detail.
Two tasks demonstrating the temperature effect running from Boinc.
In both cases boinc had been suspended immediately before the run. Notice how when genefer is restarted the new estimate is considerably more than the initial one. When it is restarted a few times running the estimates then vary by a few percent.
In case the output is not available to you, here is the stderr from the two-threaded task
genefer 3.3.4 (Linux/CPU/64-bit)
Copyright 2001-2018, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
Copyright 2010-2012, Shoichiro Yamada, Ken Brazier
Copyright 2011-2014, Michael Goetz, Ronald Schneider
Copyright 2011-2018, Iain Bethune
Genefer is free source code, under the MIT license.
Supported transform implementations: fma3 avx sse4 sse2 x87
Command line: ../../projects/www.primegrid.com/primegrid_genefer_3_3_4_3.20_x86_64-pc-linux-gnu__cpuGFN21 -boinc -q 306390^2097152+1 -nt 2
Low priority change succeeded.
Using FMA3 transform (2 threads)
Starting initialization...
Initialization complete (14.910 seconds).
Testing 306390^2097152+1...
Estimated time for 306390^2097152+1 is 80:50:00
genefer 3.3.4 (Linux/CPU/64-bit)
Copyright 2001-2018, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
Copyright 2010-2012, Shoichiro Yamada, Ken Brazier
Copyright 2011-2014, Michael Goetz, Ronald Schneider
Copyright 2011-2018, Iain Bethune
Genefer is free source code, under the MIT license.
Supported transform implementations: fma3 avx sse4 sse2 x87
Command line: ../../projects/www.primegrid.com/primegrid_genefer_3_3_4_3.20_x86_64-pc-linux-gnu__cpuGFN21 -boinc -q 306390^2097152+1 -nt 2
Low priority change succeeded.
Using FMA3 transform (2 threads)
Resuming 306390^2097152+1 from a checkpoint (37446049 iterations left)
Estimated time remaining for 306390^2097152+1 is 123:00:00
genefer 3.3.4 (Linux/CPU/64-bit)
Copyright 2001-2018, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
Copyright 2010-2012, Shoichiro Yamada, Ken Brazier
Copyright 2011-2014, Michael Goetz, Ronald Schneider
Copyright 2011-2018, Iain Bethune
Genefer is free source code, under the MIT license.
Supported transform implementations: fma3 avx sse4 sse2 x87
Command line: ../../projects/www.primegrid.com/primegrid_genefer_3_3_4_3.20_x86_64-pc-linux-gnu__cpuGFN21 -boinc -q 306390^2097152+1 -nt 2
Low priority change succeeded.
Using FMA3 transform (2 threads)
Resuming 306390^2097152+1 from a checkpoint (37446049 iterations left)
Estimated time remaining for 306390^2097152+1 is 121:00:00
genefer 3.3.4 (Linux/CPU/64-bit)
Copyright 2001-2018, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
Copyright 2010-2012, Shoichiro Yamada, Ken Brazier
Copyright 2011-2014, Michael Goetz, Ronald Schneider
Copyright 2011-2018, Iain Bethune
Genefer is free source code, under the MIT license.
Supported transform implementations: fma3 avx sse4 sse2 x87
Command line: ../../projects/www.primegrid.com/primegrid_genefer_3_3_4_3.20_x86_64-pc-linux-gnu__cpuGFN21 -boinc -q 306390^2097152+1 -nt 2
Low priority change succeeded.
Using FMA3 transform (2 threads)
Resuming 306390^2097152+1 from a checkpoint (37446049 iterations left)
Estimated time remaining for 306390^2097152+1 is 122:00:00
Terminating because BOINC client requested that we should abort.
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
|
|
|
I have posted some Linux specific thoughts here rather than annoying any windows folk in this thread.
They looked for a while like they could make the difference between one task in the 240 hours and two. But the progression of the leading edge has worked against me, even as I tweak the extra performance needed. This will be all the more so during the challenge, of course and my current guess is that my second task is likely to complete an hour or two after solstice :(
Anyway good luck to all my competitors - my goal is only to get onto the scoreboards at all, using elderly cpu-only machines, so I expect most of you to totally whizz ahead of me.
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
|
|
Looks like the GFN21 (manual-)sieve is complete.
Are the "results" of the sieve automatically incorporated into the GFN21 subproject?
Yes, of course they are
Out of curiosity, that answer triggered new questions from me. Nothing depends on these, just my ongoing desire to know absolutely everything.... ;)
While a sieve is operational, how often does the cron run to apply the results to future workunits?
And does that update apply to WU that are already created but unsent, or conversely does it only apply as from the next batch of WU (which themselves are automatically generated)?
I am wondering if two crunchers on the same WU could be using different sieves, it the tasks were downloaded at times that differ a lot (like a _3 or _8 task (*)). My feeling is that that might make validation problematic, but that is only my inexpert guess. I am also guessing that Boinc is set up so that all crunchers on the same WU get the same files.
(*) yes I have seen an _8 task -- generated when I aborted the _6 from that WU and it turned out the earlier 5 people had aborted or errored out their tasks too -- so I guess I am not the only person downloading and aborting work to fine tune my racing start tonight...
I am trusting that this will not give Michael any headaches as all the high tasknumber jobs will get validated as the challenge gets under way. It is at the END of a challenge that overhangs give him grief, if I understand correctly.
@Michael but please do say if you want us all to stop doing this. Especially on your shiny new server...
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
I'll let Jim answer the sieve questions.
I am wondering if two crunchers on the same WU could be using different sieves, it the tasks were downloaded at times that differ a lot (like a _3 or _8 task (*)). My feeling is that that might make validation problematic, but that is only my inexpert guess. I am also guessing that Boinc is set up so that all crunchers on the same WU get the same files.
For LLR or Genefer tasks, that question makes no sense. You just get a candidate to test. "Sieve file" makes no sense in this context. If you're asking whether we might suspend additional tasks if a candidate gets sieved out while someone's already testing it with Genefer (essentially canceling the WU), I don't believe we do that.
For sieve tasks that use a sieve file, yes, every task sent out in that WU will use the same sieve file.
(*) yes I have seen an _8 task -- generated when I aborted the _6 from that WU and it turned out the earlier 5 people had aborted or errored out their tasks too -- so I guess I am not the only person downloading and aborting work to fine tune my racing start tonight...
I am trusting that this will not give Michael any headaches as all the high tasknumber jobs will get validated as the challenge gets under way. It is at the END of a challenge that overhangs give him grief, if I understand correctly.
@Michael but please do say if you want us all to stop doing this. Especially on your shiny new server...
Neither Mike, Jim, nor the server care how many tasks you cancel. It is much preferred that you cancel them rather than just abandon them and let them time out. Cancel good, abandon bad.
____________
My lucky number is 75898524288+1 |
|
|
dthononVolunteer tester
Send message
Joined: 6 Dec 17
Posts: 393
ID: 957147
Credit: 1,424,257,210
RAC: 0
                     
|
Cancel good, abandon bad.
Finishing task: very good
Computing error: not too good, but it happens
Timeout after too slow progress: very bad, due to delay extension
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
In case anyone was wondering whether it's better to use the CPU Genefer in multi-threaded mode as opposed to the OpenCL Genefer running on the imbedded Intel GPU, the answer is probably the multi-threaded CPU.
I tested a Kaby Lake i5 laptop (2c/4) running "-mt 2", versus manually running GeneferOCL with the -intel switch. 99 hours for multi-threaded CPU on 2 cores vs. 229 hours for the Intel GPU. Heck, even running just one CPU core would still beat the GPU.
Although I didn't test it, previous tests have shown that running both the CPU and it's imbedded GPU is not a good idea. There's too much memory contention.
So, if you've got at least AVX, run the multi-threaded CPU app on the CPU. Don't bother with trying to run the GPU app on an internal Intel GPU.
____________
My lucky number is 75898524288+1 |
|
|
|
|
In case anyone was wondering whether it's better to use the CPU Genefer in multi-threaded mode as opposed to the OpenCL Genefer running on the imbedded Intel GPU, the answer is probably the multi-threaded CPU.
I tested a Kaby Lake i5 laptop (2c/4) running "-mt 2", versus manually running GeneferOCL with the -intel switch. 99 hours for multi-threaded CPU on 2 cores vs. 229 hours for the Intel GPU. Heck, even running just one CPU core would still beat the GPU.
Although I didn't test it, previous tests have shown that running both the CPU and it's imbedded GPU is not a good idea. There's too much memory contention.
So, if you've got at least AVX, run the multi-threaded CPU app on the CPU. Don't bother with trying to run the GPU app on an internal Intel GPU.
GPU also ? and CPU ? |
|
|
Scott BrownVolunteer moderator
Project administrator
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 05
Posts: 2165
ID: 1178
Credit: 8,777,295,508
RAC: 0
|
In case anyone was wondering whether it's better to use the CPU Genefer in multi-threaded mode as opposed to the OpenCL Genefer running on the imbedded Intel GPU, the answer is probably the multi-threaded CPU.
I tested a Kaby Lake i5 laptop (2c/4) running "-mt 2", versus manually running GeneferOCL with the -intel switch. 99 hours for multi-threaded CPU on 2 cores vs. 229 hours for the Intel GPU. Heck, even running just one CPU core would still beat the GPU.
Although I didn't test it, previous tests have shown that running both the CPU and it's imbedded GPU is not a good idea. There's too much memory contention.
So, if you've got at least AVX, run the multi-threaded CPU app on the CPU. Don't bother with trying to run the GPU app on an internal Intel GPU.
I am curious if anyone has tried this with an AMD mobile CPU/GPU. I'd be very curious to see the times for a Vega 10 or Vega 11.
|
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 916
ID: 107307
Credit: 974,514,092
RAC: 0
|
|
River~~ wrote: Looks like the GFN21 (manual-)sieve is complete.
Are the "results" of the sieve automatically incorporated into the GFN21 subproject?
Yes, of course they are
Out of curiosity, that answer triggered new questions from me. Nothing depends on these, just my ongoing desire to know absolutely everything.... ;)
While a sieve is operational, how often does the cron run to apply the results to future workunits?
It's not a cron, it's me. Every time I process sieving the programs involved spit out a list of entries corresponding to currently loaded candidates at or above the smallest unfinished candidate up to the maximum of what's been loaded. So the answer to that question is at least once a day but as high as how many times I process factors (at the end of the GFN21 sieving I did it four or more times in the same day).
And does that update apply to WU that are already created but unsent, or conversely does it only apply as from the next batch of WU (which themselves are automatically generated)?
It applies to everything. If the candidate has not yet been turned into a workunit, it's deleted entirely from the system. If it's a workunit, it's cancelled. Any unsent results won't ever be sent out. If any result has already been sent out the workunit is changed to quorum 1 so that the first result will always be valid. Meanwhile in the science table the residue field for that candidate is set to "factor found", which the validator will not overwrite. So, any result is valid. If a result is aborted, a replacement will not be sent out. If one result is already back, it's immediately validated.
Yes, two users could theoretically return different results and in the event of a mismatch the first one back will be considered the valid one. Given how few candidates match work already loaded and how small the window is that the workunit is actually in progress as opposed to completed or not yet sent, I don't worry about it.
And of course all this is logged by the system. Here's a list of what's been removed in December so far:
2018-12-01 03:33:27 DELETED 177926808920376475649 | 6222666 ^ 262144 + 1
2018-12-02 03:01:48 DELETED 180007194393592201217 | 6475776 ^ 262144 + 1
2018-12-02 03:02:58 DELETED 268173557754071351297 | 2599094 ^ 524288 + 1
2018-12-03 03:49:31 DELETED 180158366734683734017 | 6358436 ^ 262144 + 1
2018-12-03 03:49:31 DELETED 180163671864106811393 | 6474722 ^ 262144 + 1
2018-12-03 03:52:29 CANCELED 595390791947219107841 | 301210 ^ 2097152 + 1
2018-12-04 03:37:38 DELETED 180268903588063346689 | 6419592 ^ 262144 + 1
2018-12-04 03:39:56 DELETED 580455962646969057281 | 304062 ^ 2097152 + 1
2018-12-04 19:14:28 DELETED 179480832383978045441 | 6296050 ^ 262144 + 1
2018-12-05 03:02:31 DELETED 178005078926700838913 | 6379318 ^ 262144 + 1
2018-12-05 03:02:31 DELETED 180463112043020419073 | 6305712 ^ 262144 + 1
2018-12-05 18:40:55 CANCELED 616373047341787119617 | 301476 ^ 2097152 + 1
2018-12-06 03:34:02 DELETED 270421335622425772033 | 2537998 ^ 524288 + 1
2018-12-06 03:34:02 DELETED 270863604221939286017 | 2580640 ^ 524288 + 1
2018-12-06 03:34:02 DELETED 271067970231780507649 | 2512474 ^ 524288 + 1
2018-12-07 02:04:11 DELETED 178534551862615474177 | 6422708 ^ 262144 + 1
2018-12-07 02:04:11 DELETED 181314329692403662849 | 6473894 ^ 262144 + 1
2018-12-07 02:07:05 DELETED 601762618606234894337 | 324746 ^ 2097152 + 1
2018-12-07 15:29:08 DELETED 180623761788362555393 | 6382390 ^ 262144 + 1
2018-12-07 15:29:57 DELETED 272043225154299363329 | 2551528 ^ 524288 + 1
2018-12-07 15:32:24 DELETED 608929624886316367873 | 308074 ^ 2097152 + 1
2018-12-08 02:12:26 DELETED 178698980196415438849 | 6491944 ^ 262144 + 1
2018-12-08 02:12:26 DELETED 178768413182851022849 | 6403358 ^ 262144 + 1
2018-12-08 02:13:50 DELETED 272885450356405829633 | 2553986 ^ 524288 + 1
2018-12-08 16:18:57 DELETED 272448994108663398401 | 2571892 ^ 524288 + 1
2018-12-08 16:20:54 DELETED 611452311016635891713 | 315706 ^ 2097152 + 1
2018-12-08 16:22:31 DELETED 453443186919871610881 | 146558 ^ 4194304 + 1
2018-12-08 16:29:35 DELETED 456827269156492541953 | 147544 ^ 4194304 + 1
2018-12-09 02:56:06 DELETED 181059337314884386817 | 6311914 ^ 262144 + 1
2018-12-09 02:56:47 DELETED 274250578952392278017 | 2510280 ^ 524288 + 1
2018-12-09 02:56:47 DELETED 274652930427952562177 | 2539816 ^ 524288 + 1
2018-12-09 18:29:40 DELETED 184220624501510504449 | 6435698 ^ 262144 + 1
2018-12-09 18:30:41 DELETED 270731358544365355009 | 2597982 ^ 524288 + 1
2018-12-09 18:30:41 DELETED 270735112080819486721 | 2553378 ^ 524288 + 1
2018-12-10 03:14:23 DELETED 182909763903383142401 | 6378790 ^ 262144 + 1
2018-12-10 11:12:34 DELETED 276502229314786295809 | 2567030 ^ 524288 + 1
2018-12-10 11:14:26 DELETED 613138860445460856833 | 410550 ^ 2097152 + 1
2018-12-10 17:01:51 DELETED 183631710291857768449 | 6448830 ^ 262144 + 1
2018-12-10 17:01:51 DELETED 184362248104753758209 | 6295834 ^ 262144 + 1
2018-12-11 02:32:21 DELETED 277497845918591877121 | 2587598 ^ 524288 + 1
Anything marked as DELETED was completely removed from PrimeGrid, while those marked as CANCELED were already in progress. Those candidates that had already been run don't show up at all as no action is taken. If we were ever to rerun that candidate range (like if the genefer program was found to have some sort of flaw), then those candidates, not being in the then-current sieve, wouldn't be run again in any case.
By the way, the two CANCELED workunits were these:
301210^2097152+1: http://www.primegrid.com/workunit.php?wuid=589177236
301476^2097152+1: http://www.primegrid.com/workunit.php?wuid=589216305 |
|
|
|
|
|
I see the winter challenge has begun. Perfect, that means it's time for my winter power outage to begin. An unfortunate coincidence that has occurred during at least one challenge a year for the past 3 or 4 years. The city just to the south is out and the wind is blowing. I'm shutting down everything not UPS'd. :(
____________
Eating more cheese on Thursdays. |
|
|
|
|
|
What a drag!
Personally, I'm off to an interesting start.
Did set max_concurrent to 1 and had to notice that this means one GPU-task OR one CPU-task.
Really good I sat in front of my computers and noticed this.
____________
Greetings, Jens
92914140^65536+1 |
|
|
|
|
|
That is interesting to see, JimB. /JeppeSN |
|
|
|
|
|
Is the newest nvidia driver suitable for this challenge or is another driver recommended like with the previous challenge? |
|
|
|
|
|
All my GFN-21 work units run by my GTX-1080 errored too, what is up with the new driver? |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
All my GFN-21 work units run by my GTX-1080 errored too, what is up with the new driver?
Download the driver directly from Nvidia and install that driver. The driver that Windows Update gave you doesn't include the OpenCL interface.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
If you don't have an RTX 20x0 card (the new ones), 399 is still the last suitable driver.
____________
Eating more cheese on Thursdays. |
|
|
|
|
If you don't have an RTX 20x0 card (the new ones), 399 is still the last suitable driver.
Thanks. That dropped my estimated runtime on a 1080 from 11 hours to 6 hours. |
|
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 644
ID: 164101
Credit: 305,010,093
RAC: 0
|
If you don't have an RTX 20x0 card (the new ones), 399 is still the last suitable driver.
No, genefer is not AP27. All drivers can be used, any 4xx is as fast as 399.
You can download from NVIDIA the 417.22. |
|
|
|
|
|
The Open CL install did the trick I hope, will be running some GFN this morning! |
|
|
|
|
|
Thank you windows update.... grrr, cost me about 6 hrs of computing for this contest on at least one of my machines. Rebooted overnight, it didn't restart BOINC, and then I had no internet access... just got it working again
Back to crunching.
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
Thank you windows update.... grrr, cost me about 6 hrs of computing for this contest on at least one of my machines. Rebooted overnight, it didn't restart BOINC, and then I had no internet access... just got it working again
Back to crunching.
It's a full new version of Windows 10. On a Kaby Lake laptop with an SSD it took about 3 hours to do. It took a lot longer on a system running off a HDD.
Someone realized yesterday was "Update Tuesday" and I immediately started the updates. The KL finished 3 minutes before the challenge started.
Windows 7 was a fairly small update, but 10's update is huge.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Well, something changed with the video card stuff..
I'm only able to run one GPU work unit now in BOINC...
One primegrid GPU task using only 1 GPU. I can't get it to get work for the 2nd GPU now :( WTH!?
Even log says this:
12/12/2018 8:46:39 AM | PrimeGrid | Sending scheduler request: Requested by user.
12/12/2018 8:46:39 AM | PrimeGrid | Not requesting tasks: don't need (CPU: ; NVIDIA GPU: )
But it does need a GPU task as it does see both GPUs, it WAS crunching on both, it still says Device 0 and about 45 min ago said Device 0 on one and Device 1 on the other...
but now, I can't get two to run :(
Ideas? |
|
|
|
|
|
Seems isolated to Primegrid.. i was able to get tasks for the 2nd GPU from another Boinc project :(
Hmmm
The only thing that happened was this update. Last night, crunching happlily on 2 GPUs.
Things I just tried:
Exiting and restarting BOINC
Rebooting
Setting it to get a minimum or 1 day of work instead of .1 days
Removed cc_config.xml restrictions
Still, can't get it to download a 2nd GPU task on this machine for Primegrid but other projects it gets work from.... I don't get it. |
|
|
|
|
Seems isolated to Primegrid.. i was able to get tasks for the 2nd GPU from another Boinc project :(
Did you set max_concurrent at some point?
Even if the app_config.xml is deleted some values seem to still linger on in Boinc's behaviour.
You would have to set new values then.
____________
Greetings, Jens
92914140^65536+1 |
|
|
|
|
Seems isolated to Primegrid.. i was able to get tasks for the 2nd GPU from another Boinc project :(
Did you set max_concurrent at some point?
Even if the app_config.xml is deleted some values seem to still linger on in Boinc's behaviour.
You would have to set new values then.
max concurrent not set anywhere for primegrid |
|
|
|
|
Seems isolated to Primegrid.. i was able to get tasks for the 2nd GPU from another Boinc project :(
Did you set max_concurrent at some point?
Even if the app_config.xml is deleted some values seem to still linger on in Boinc's behaviour.
You would have to set new values then.
max concurrent not set anywhere for primegrid
but, now I will set it to 2 for genefer and see if that does anything... |
|
|
|
|
|
ok, so I aborted the one task it was crunching, lost about 33% done on it, but now, with max concurrent set to 2 for genefer in app_config, it has 1 for each GPU again...
I really dunno if it was the app_config that helped because it didn't get any new work until I aborted the currently running task.
So, another 2 hrs of crunching lost. 8 hrs total lost :( arg. And, with a 2080! Double ugh and arg.
Hope this helps anyone who can't get 2 tasks running, abort the one and see if it resets itself. |
|
|
|
|
|
If anyone should want to crunch Genefer-21 on CPU as well they should keep in mind that the max_concurrent for 'genefer' works on Genefer-21 GPU AND CPU.
Btw.: Someone on my team mentioned that two tasks on one GPU would not have any benefit.
____________
Greetings, Jens
92914140^65536+1 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
If anyone should want to crunch Genefer-21 on CPU as well they should keep in mind that the max_concurrent for 'genefer' works on Genefer-21 GPU AND CPU.
Btw.: Someone on my team mentioned that two tasks on one GPU would not have any benefit.
I personally never use <max_concurrent>.
____________
My lucky number is 75898524288+1 |
|
|
|
|
If anyone should want to crunch Genefer-21 on CPU as well they should keep in mind that the max_concurrent for 'genefer' works on Genefer-21 GPU AND CPU.
Btw.: Someone on my team mentioned that two tasks on one GPU would not have any benefit.
I personally never use <max_concurrent>.
Yeah, me either. In fact I'm going to remove it from the app_config because it ran fine without it. |
|
|
|
|
It's a full new version of Windows 10. On a Kaby Lake laptop with an SSD it took about 3 hours to do. It took a lot longer on a system running off a HDD.
To clarify, the updates released yesterday were the normal small ones. What might have happened is that in some cases the major update that was released earlier may also have been pushed, as that is being rolled out gradually.
In case of any suspected GPU problems after the major update, reinstall the GPU driver as a first step. |
|
|
|
|
It's a full new version of Windows 10. On a Kaby Lake laptop with an SSD it took about 3 hours to do. It took a lot longer on a system running off a HDD.
To clarify, the updates released yesterday were the normal small ones. What might have happened is that in some cases the major update that was released earlier may also have been pushed, as that is being rolled out gradually.
In case of any suspected GPU problems after the major update, reinstall the GPU driver as a first step.
I got a small one that rebooted me automatically that caused crunching issues... Now I went to double check to make sure they were done with updates and I am getting, and installing, a feature update which is several GB. It's big. Its been going for over 2 hrs already.
I already lost some time for this contest from these updates so I want to be sure I'm good to go b4 heading to work shortly.
And yes, I'm being prepared to reinstall the graphics drivers if need be. |
|
|
|
|
I personally never use <max_concurrent>.
I use it because I don't want to fiddle with my settings in CPU-challenges.
Other CPU-projects off, one work-unit from the challenged sub-project running beside whatever is crunching on the GPU, plus half of my threads are unused.
Works fine.
____________
Greetings, Jens
92914140^65536+1 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
It's a full new version of Windows 10. On a Kaby Lake laptop with an SSD it took about 3 hours to do. It took a lot longer on a system running off a HDD.
To clarify, the updates released yesterday were the normal small ones. What might have happened is that in some cases the major update that was released earlier may also have been pushed, as that is being rolled out gradually.
In case of any suspected GPU problems after the major update, reinstall the GPU driver as a first step.
I manually did a "check for updates". It's possible you won't get the big Windows update if you don't do that.
____________
My lucky number is 75898524288+1 |
|
|
|
|
I manually did a "check for updates". It's possible you won't get the big Windows update if you don't do that.
Yup, I did the manual on 4 win10 systems and got the full new package on each system, plus had to re-install the nvidia drivers. |
|
|
|
|
|
Does anyone know what the ACTUAL power draw of a 1080 Ti (FE) running at stock clocks is?
____________
 |
|
|
|
|
Does anyone know what the ACTUAL power draw of a 1080 Ti (FE) running at stock clocks is?
Under max load it's about 250W
Source: https://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-ti,4972-6.html
Edited to add source |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
Day 1 is done!
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-12 22:46:02 UTC)
10216 tasks have been sent out. [CPU/GPU/anonymous_platform: 3910 (38%) / 6306 (62%) / 0 (0%)]
Of those tasks that have been sent out:
4149 (41%) came back with some kind of an error. [1158 (11%) / 2991 (29%) / 0 (0%)]
803 (8%) have returned a successful result. [39 (0%) / 764 (7%) / 0 (0%)]
5264 (52%) are still in progress. [2713 (27%) / 2551 (25%) / 0 (0%)]
Of the tasks that have been returned successfully:
659 (82%) are pending validation. [30 (4%) / 629 (78%) / 0 (0%)]
144 (18%) have been successfully validated. [9 (1%) / 135 (17%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=323414. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 5.00% as much as it had prior to the challenge!
Normal GFN-21 traffic is about 200 tasks per day. Accounting for errors, we're sending out tasks at about 20x times that rate. We'll see how many get completed by the time the challenge ends.
This is the largest challenge we've ever done. The candidates being tested are 11 million digits in length. Even an SoB challenge at the leading edge would be only 9 million digits, and every GFN challenge we've run in the past included smaller GFN tasks.
With everyone running such large tasks on both CPUs and GPUs, the server is sending out very few tasks compared to a normal day. Normally on challenges we see most server metrics jump way up. This time, the charts look like the server usage fell off a cliff.
I've personally completed one task so far. My GTX 1060 can do about 1 task every 20 hours, and the 3 CPUs, combined, will do less than one task per day. This challenge really is a challenge!
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Actually pretty close to the TDP. That's nice, for a change.
____________
|
|
|
|
|
|
I'm having the same problem as earlier. Primegrid or BOINC won't allow a 2nd task to be sent for the 2nd GPU. Obviously the 2080 finishes twice as fast as the 1070, but, once the 2080 finishes, the only one running is the 1070 task. It just won't get a new task for that 2nd GPU. I'm at a loss as to why. I've been crunching with two GPUs just fine previously. Seems a shame to only be able to do half the work on a 2080 now while waiting for the 1070. Or, I could just tell it to only use the 2080, but then the 1070 is useless.
Looking for ideas... thought I'd do well in this challenge but I'm losing a lot of GPU computing time.
Is anyone else in the contest running 2 different GPU models and having issues like this? |
|
|
|
|
This is the largest challenge we've ever done. The candidates being tested are 11 million digits in length. Even an SoB challenge at the leading edge would be only 9 million digits, and every GFN challenge we've run in the past included smaller GFN tasks. I am not sure. As I can remember, we have run GFN 22 during the 2014's Stallion Edition 15 days challenge. |
|
|
|
|
I'm having the same problem as earlier. Primegrid or BOINC won't allow a 2nd task to be sent for the 2nd GPU. Obviously the 2080 finishes twice as fast as the 1070, but, once the 2080 finishes, the only one running is the 1070 task. It just won't get a new task for that 2nd GPU. I'm at a loss as to why. I've been crunching with two GPUs just fine previously. Seems a shame to only be able to do half the work on a 2080 now while waiting for the 1070. Or, I could just tell it to only use the 2080, but then the 1070 is useless.
Looking for ideas... thought I'd do well in this challenge but I'm losing a lot of GPU computing time.
Is anyone else in the contest running 2 different GPU models and having issues like this?
My fix, hopefully a good one this time, was to set resource share to 100 instead of 0, and set the minumum work queue to 1 full day. My theory is that it was only getting work for the 1 GPU and not any additional. So now, it gets more work since there are only X hrs left for the first task and it needs to get more work to fill the queue of 24hrs. Then, when it gets the 2nd task, it immediately starts running on the free 2nd GPU. Why it doesn't know to queue up work for the 2nd GPU too I have no idea. But hopefully, fingers crossed, this keeps them both crunching. A 1st is not needed, only continuous work for both GPUs... we shall see.
Wish a Penguin luck :) |
|
|
|
|
4149 (41%) came back with some kind of an error. [1158 (11%) / 2991 (29%) / 0 (0%)]
803 (8%) have returned a successful result. [39 (0%) / 764 (7%) / 0 (0%)]
5264 (52%) are still in progress. [2713 (27%) / 2551 (25%) / 0 (0%)]
I got a late start with my newly-working GPU (finishing off the gazillion PPS SV WUs that were in my queue).
My first WU should complete in five and a half hours; hopefully I will not add to the error total!
____________
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
This is the largest challenge we've ever done. The candidates being tested are 11 million digits in length. Even an SoB challenge at the leading edge would be only 9 million digits, and every GFN challenge we've run in the past included smaller GFN tasks. I am not sure. As I can remember, we have run GFN 22 during the 2014's Stallion Edition 15 days challenge.
Right, but that was simultaneous with the GFN-short challenge, which was either GFN-19 or GFN-20 back then. We've never had a challenge where you had to run tasks of this size.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
This works to keep Windows Updates from updating the Graphics Drivers.
https://forums.evga.com/FindPost/2893374 |
|
|
dukebgVolunteer tester
Send message
Joined: 21 Nov 17
Posts: 238
ID: 950482
Credit: 23,670,125
RAC: 0
             
|
|
I have something weird going on with cpuGFN21 MT on many threads.
Initially I wanted to run on all 8 threads (i7-9700K). This is how the first task concluded.
Estimated time remaining for 310594^2097152+1 is 24:20:00
...
310594^2097152+1 is complete. (11517964 digits) (err = 0.0781) (time = 25:16:18)
This is fine. It ran a bit slower than initial estimation, but that's irrelevant to the point of this post.
The second task has already gone to about 30-31%, when I decided to switch to 4 threads.
Using FMA3 transform (4 threads)
Resuming 322604^2097152+1 from a checkpoint (26342606 iterations left)
Estimated time remaining for 322604^2097152+1 is 16:10:00
But... I expected twice as long. 70% of 25hrs on 8 threads is ~17 hours. And this is on 4? While another task is starting on other 4 cores?
I switched back to 8 (restarting the client & task)
Using FMA3 transform (8 threads)
Resuming 322604^2097152+1 from a checkpoint (26267841 iterations left)
Estimated time remaining for 322604^2097152+1 is 17:00:00
Actually expected value. Went back to 4 threads, thinking it was a fluke.
Using FMA3 transform (4 threads)
Resuming 322604^2097152+1 from a checkpoint (26221558 iterations left)
Estimated time remaining for 322604^2097152+1 is 16:10:00
Nope. Here it is again.
I'll see how fast it will actually finish.
N.B. The second task on other 4 threads starting from the beginning had the "expected" twice-longer estimation:
Using FMA3 transform (4 threads)
Starting initialization...
Initialization complete (6.141 seconds).
Testing 323976^2097152+1...
Estimated time for 323976^2097152+1 is 56:00:00
...
Using FMA3 transform (8 threads)
Resuming 323976^2097152+1 from a checkpoint (38358471 iterations left)
Estimated time remaining for 323976^2097152+1 is 24:10:00
...
Using FMA3 transform (4 threads)
Resuming 323976^2097152+1 from a checkpoint (38331746 iterations left)
Estimated time remaining for 323976^2097152+1 is 56:20:00
EDIT: nevermind, the estimation was done before the second task kicked in. Made another test making sure that the half-started task is starting when the second is well on the way and the estimation is 31 hours as expected. |
|
|
dukebgVolunteer tester
Send message
Joined: 21 Nov 17
Posts: 238
ID: 950482
Credit: 23,670,125
RAC: 0
|
EDIT: nevermind, the estimation was done before the second task kicked in. Made another test making sure that the half-started task is starting when the second is well on the way and the estimation is 31 hours as expected.
Anyway, the weird thing then is that the estimation seems to be based on how many threads are free at the moment, not how many it's supposed to work on?
Or is that just the memory / cache size bottleneck that ends up like this... |
|
|
SkyFallVolunteer tester
Send message
Joined: 26 Mar 15
Posts: 13
ID: 388081
Credit: 56,709,166
RAC: 0
|
|
Solved, a suspended Task blocked the queue.
______________________
I have a weird issue that I can't get tasks for my secondary AMD GPU.
(Host: http://www.primegrid.com/show_host_detail.php?hostid=938706)
I also have very strange runtime estimates, for 4 core CPU tasks it estimates 471days and for the nVidia GPU tasks 46 days.
Also, when I try to update the primegrid project, it just postpones the communication for 10 seconds and then still is not able to get any new workunits. |
|
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 644
ID: 164101
Credit: 305,010,093
RAC: 0
|
Anyway, the weird thing then is that the estimation seems to be based on how many threads are free at the moment, not how many it's supposed to work on?
Or is that just the memory / cache size bottleneck that ends up like this...
The estimation is the result of a benchmark: let the program run few seconds and estimate the time it will take.
Note that larger is the ratio turbo frequency / normal frequency, less accurate is the estimation.
There is a memory bottleneck with 9th generation i7s and i9s.
L3 cache size of i7-9700K is 12 MB and cpuGFN21 memory size is 16 MB R/W + 8 MB Ro.
8 threads run at about the same speed than 4.
The optimal setting may be 2 x 4-thread cpuGFN21??? |
|
|
|
|
8 threads run at about the same speed than 4.
The optimal setting may be 2 x 4-thread cpuGFN21???
Worth a test to confirm, but I would suspect that if the 1st task already is able to consume all the bandwidth, running the 2nd task would make both run at half the speed. |
|
|
|
|
|
Indeed.
But if the first task is not consuming all of the bandwidth, two tasks might benefit from running parallel.
____________
Greetings, Jens
92914140^65536+1 |
|
|
dukebgVolunteer tester
Send message
Joined: 21 Nov 17
Posts: 238
ID: 950482
Credit: 23,670,125
RAC: 0
|
8 threads run at about the same speed than 4.
The optimal setting may be 2 x 4-thread cpuGFN21???
Worth a test to confirm, but I would suspect that if the 1st task already is able to consume all the bandwidth, running the 2nd task would make both run at half the speed.
Yes, this is my impression from the logs posted above. 4-threaded task working alone gives the same estimation (even slightly better actually) as a 8-threaded task (working alone). But if another 4-threaded task is underway, a 4-threaded task gives estimation that is about [slightly less than] double in time.
So, theoretically, I can leave 1 4-threaded task alone and it will work about as fast as 1 8-threaded task alone. While leaving me with less heat/power consumed and "space" for feeding GPU.
I regret not doing any benchmarks in preparation to the challenge now. |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
This works to keep Windows Updates from updating the Graphics Drivers.
https://forums.evga.com/FindPost/2893374
Cool. Thanks for this.
I had five boxes reboot overnight, grrrr, thankfully no driver issues........ (I hope).
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
|
|
But if another 4-threaded task is underway, a 4-threaded task gives estimation that is about [slightly less than] double in time.
Question then is, if the "slightly less than double in time" is worth it for the throughput increase, vs completing units faster at slightly lower throughput. That's a personal choice.
I regret not doing any benchmarks in preparation to the challenge now.
I did a bunch, but even then maybe I could have done more. With my 8086k (6 core) I saw similar. Up to 4 threads saw some degree of scaling, but 4-6 were practically the same. I didn't try 2x3. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
Mid-day update...
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-13 19:27:43 UTC)
14023 tasks have been sent out. [CPU/GPU/anonymous_platform: 5000 (36%) / 9023 (64%) / 0 (0%)]
Of those tasks that have been sent out:
6325 (45%) came back with some kind of an error. [1999 (14%) / 4326 (31%) / 0 (0%)]
2114 (15%) have returned a successful result. [211 (2%) / 1903 (14%) / 0 (0%)]
5584 (40%) are still in progress. [2790 (20%) / 2794 (20%) / 0 (0%)]
Of the tasks that have been returned successfully:
1408 (67%) are pending validation. [153 (7%) / 1255 (59%) / 0 (0%)]
703 (33%) have been successfully validated. [58 (3%) / 645 (31%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
3 (0%) are inconclusive. [0 (0%) / 3 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=327842. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 6.44% as much as it had prior to the challenge!
I want to say something about the large number of errors.
We need to consider the GPU and CPU tasks separately.
The 4326 GPU errors are mostly what I call "3 second errors", i.e., the task fails before it can even start. By far, the most likely cause for this is a Windows 10 update which updated the Nvidia driver with one that doesn't support OpenCL. Without OpenCL, the GeneferOCL (aka GeneferOpenCL) can't find a useable GPU and immediately fails. Rinse and repeat every two or three seconds, and a single computer with a borked video driver can easily rack up 1000 errors in an hour. Meanwhile, a correctly running computer will correctly complete one of these tasks in several hours or days.
Even though there's about 50% errors, that represents only a handful of bad computers.
On the CPU side, there's 1999 errors. We don't see "3 second" failures on CPUs that much. For the most part, most computers are able to at least start running LLR. The large number of errors here has a completely different cause: these tasks were aborted by the user. There's two common reasons for doing this. The first is that BOINC downloaded way too many tasks, for whatever reason. The user then cancels the extra tasks. The second is that the user set the computer up for multi-threading, which will cause BOINC to download extra tasks at the beginning. Many people cancel those extra tasks. Put it all together and over 75% of the CPU tasks with "some sort of error" were in fact aborted by the user.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Thank you for the clarification, Michael.
____________
"Accidit in puncto, quod non contingit in anno."
Something that does not occur in a year may, perchance, happen in a moment. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
After 2 days:
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-13 22:45:56 UTC)
14615 tasks have been sent out. [CPU/GPU/anonymous_platform: 5127 (35%) / 9488 (65%) / 0 (0%)]
Of those tasks that have been sent out:
6652 (46%) came back with some kind of an error. [2053 (14%) / 4599 (31%) / 0 (0%)]
2310 (16%) have returned a successful result. [243 (2%) / 2067 (14%) / 0 (0%)]
5653 (39%) are still in progress. [2831 (19%) / 2822 (19%) / 0 (0%)]
Of the tasks that have been returned successfully:
1519 (66%) are pending validation. [176 (8%) / 1343 (58%) / 0 (0%)]
788 (34%) have been successfully validated. [67 (3%) / 721 (31%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
3 (0%) are inconclusive. [0 (0%) / 3 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=328340. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
____________
My lucky number is 75898524288+1 |
|
|
|
|
I had five boxes reboot overnight, grrrr, thankfully no driver issues........ (I hope).
I hear ya. One here only, but I woke up and was like why is it so cool in here? GPUs were idling not crunching.
But now, between 3 PCs, a 1050 Ti, a 1060 6gb, a 1070, and a 2080 I'm fully back up and crunching smoothly now for the contest. |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
I had five boxes reboot overnight, grrrr, thankfully no driver issues........ (I hope).
I hear ya. One here only, but I woke up and was like why is it so cool in here? GPUs were idling not crunching.
But now, between 3 PCs, a 1050 Ti, a 1060 6gb, a 1070, and a 2080 I'm fully back up and crunching smoothly now for the contest.
Best of luck! :)
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
Somewhere between 10% and 20% is probably more realistic. The huge surge of tasks sent out at the beginning of the challenge skews the statistics somewhat, so we won't see quite as much growth for the remaining days as we did in the first 2.
That's still incredible progress, however.
____________
My lucky number is 75898524288+1 |
|
|
|
|
But now, between 3 PCs, a 1050 Ti, a 1060 6gb, a 1070, and a 2080 I'm fully back up and crunching smoothly now for the contest.
Have you had any problems with the 2080? There are a lot of complaints about RTX cards not working (and I'm not talking about ray tracing slowing down games).
I ask because I was looking at the RTX specs, and the 2080 Ti has pretty close to the performance of the Volta (14889 GFLOP/s v 13448, FP32; 107.6 TFLOP/s v 110.0, Tensor) and only falls short on FP64 (7449.6 GFLOP/s vs 420.2 GFLOP/s) - for $1,800 less, making it a somewhat more reasonable proposition for a second high end cruncher, but I certainly wouldn't want to build a new machine just to have the star of the show be a dud.
____________
|
|
|
|
|
|
Just a curious thought. An aborted task counts as an errored task? So when I aborted tasks as you said because I didn't want them they are now errored tasks. Why then is there two categories on the task list of error and aborted?
Of course I may have read the explanation in-correctly. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
Just a curious thought. An aborted task counts as an errored task? So when I aborted tasks as you said because I didn't want them they are now errored tasks. Why then is there two categories on the task list of error and aborted?
Of course I may have read the explanation in-correctly.
It's just semantics. Aborted tasks are different from other errors, but for the challenge statistics I lumped them all together in the same basket. For every task sent out, either a result was returned, the task ended without returning a result, or the task is still in progress.
For the important stuff, there's still no penalty to aborting tasks and you should feel free to do so whenever you wish for whatever reason you wish. Aborting tasks does not hurt anything.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Thanks. Makes sense now. Was just hoping I wasnt making too big a mess aborting units. |
|
|
|
|
Have you had any problems with the 2080? There are a lot of complaints about RTX cards not working (and I'm not talking about ray tracing slowing down games).
I ask because I was looking at the RTX specs, and the 2080 Ti has pretty close to the performance of the Volta (14889 GFLOP/s v 13448, FP32; 107.6 TFLOP/s v 110.0, Tensor) and only falls short on FP64 (7449.6 GFLOP/s vs 420.2 GFLOP/s) - for $1,800 less, making it a somewhat more reasonable proposition for a second high end cruncher, but I certainly wouldn't want to build a new machine just to have the star of the show be a dud.
What complaints? I've got a 2070 early on, and it's done folding for a month elsewhere, and is now powering through this challenge. The speed and power efficiency over Pascal is a nice boost. It will be interesting to see how efficiency scales as we see more of the range. Also don't fixate purely on the compute specs, the ram bandwidth also plays a part and that will take more testing. |
|
|
AzmodesVolunteer tester
Send message
Joined: 30 Dec 16
Posts: 184
ID: 479275
Credit: 1,564,594,480
RAC: 0
     
|
|
I have a 2070 and a 2080 crunching for the challenge and they're performing well. 28,000 and 24,000 seconds for one task, respectively (cf. 33,000 and 48,000 secs on a 1080 Ti and 1070 Ti). The only issue I've had recently was the 2080 starting to choose the OCL3 transform all of a sudden, which added another hour to runtimes, but I forced it to use OCL4 through the app_config and now it's back to normal.
So I can't speak for gaming performance, but for PrimeGrid (and crunching in general) they're pretty great. They're absolutely wiping the floor with Pascal cards at manual sieving especially.
As for aborted tasks, I went through a couple of these for two reasons. One, BOINC downloaded a whole bunch of them for my multi-threaded CPUs despite zero work cache. Two, I didn't realize Genefer CPU multi-threading requires AVX, so I had to abort tasks for some of my older CPUs which would never have finished single-threaded tasks in time.
____________
Long live the sievers.
+ Encyclopaedia Metallum: The Metal Archives + |
|
|
|
|
I had five boxes reboot overnight, grrrr, thankfully no driver issues........ (I hope).
I hear ya. One here only, but I woke up and was like why is it so cool in here? GPUs were idling not crunching.
But now, between 3 PCs, a 1050 Ti, a 1060 6gb, a 1070, and a 2080 I'm fully back up and crunching smoothly now for the contest.
Best of luck! :)
Thanks Robish. I see all the usual big names in the lead. If I can come in around top 25% I'd be a smiling penguin :) Most of the leaders are some of the biggest sievers with some impressive hardware :) Good luck to all! |
|
|
|
|
But now, between 3 PCs, a 1050 Ti, a 1060 6gb, a 1070, and a 2080 I'm fully back up and crunching smoothly now for the contest.
Have you had any problems with the 2080? There are a lot of complaints about RTX cards not working (and I'm not talking about ray tracing slowing down games).
I ask because I was looking at the RTX specs, and the 2080 Ti has pretty close to the performance of the Volta (14889 GFLOP/s v 13448, FP32; 107.6 TFLOP/s v 110.0, Tensor) and only falls short on FP64 (7449.6 GFLOP/s vs 420.2 GFLOP/s) - for $1,800 less, making it a somewhat more reasonable proposition for a second high end cruncher, but I certainly wouldn't want to build a new machine just to have the star of the show be a dud.
I've had no problems here at Primegrid with the 2080. It runs everthing I throw at it, and fast. I notice now suddenly there seems to be a 5% slowdown both on the 1070 and the 2080. Not sure why. From just over 6hrs on the 2080 to just over 6 and a half hours now. Not sure why, didn't change anything on my side.
It's fast, sucks down some power and heats up like a mofo though. GFN21 has it running at 80c with 250watts and 95% fan speed and I have the Windforce one with 3 fans. It hasn't faultered though, it's handling it fine 24/7. |
|
|
|
|
I have a 2070 and a 2080 crunching for the challenge and they're performing well. 28,000 and 24,000 seconds for one task, respectively (cf. 33,000 and 48,000 secs on a 1080 Ti and 1070 Ti). The only issue I've had recently was the 2080 starting to choose the OCL3 transform all of a sudden, which added another hour to runtimes, but I forced it to use OCL4 through the app_config and now it's back to normal.
I wonder if this is why I am seeing a 5% slowdown.... I'm not at all familiar with the OCL3 or OCL4 transforms...
What do you add to app_config to set it for OCL4?
|
|
|
|
|
Have you had any problems with the 2080? There are a lot of complaints about RTX cards not working (and I'm not talking about ray tracing slowing down games).
I've had problems with other BOINC projects not working with the 2080. Also my video encoding software doesn't work with it... Problems not resolved yet. But, as I mentioned, Primegrid = zero problems. The 2080 tears work units up.
If you keep you older graphics card in the same machine, you can use cc_config in the BOINC data directory to tell BOINC to ignore a certain GPU for each project, and even for each app within a project. |
|
|
|
|
|
Just skimmed through my systems, I'm not seeing any significant variation in unit times throughout this challenge. Could there be another factor? e.g. an increase in ambient temperature could result in reduced core clocks.
Also I'd be cautious about changing the transform. I would think there is a reason for it, and you risk a bad result if the transform is not suitable. I had noticed during pre-testing one of my Maxwell cards would swap between two transforms due to round off error I think it was. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
Also I'd be cautious about changing the transform. I would think there is a reason for it...
There is always "a reason", but it's not always a "good reason".
It will always be safe during the this challenge to manually select OCL4. It won't break anything.
The code is written to do a quick benchmark of each transform to see which one runs fastest. That turns out to have been a really, really good decision, since different GPUs run different transforms at different speeds. Putting those automatic speed tests into the code has worked very well over the years.
But it's not perfect. The benchmarks can be tricked into choosing a slower transform, if, for example, you're doing something on the GPU while OCL4 is being benchmarked but not when OCL5 is being benchmarked. It's not a perfect system.
Also, if two transforms are very close in speed, it may be close to random which one gets selected.
If you're confident that OCL4 is the best choice, then there's no downside to forcing it via the command line. Just remember that when GFN21 reaches b=1011306 OCL4 is no longer usable.
____________
My lucky number is 75898524288+1 |
|
|
|
|
Have you had any problems with the 2080? There are a lot of complaints about RTX cards not working (and I'm not talking about ray tracing slowing down games).
AP is the Only PG App that needs a little Tender Loving OpenCL Coding
And not on the RTX GPU but using the 4xx.xx Drivers on the older GTX GPU.
There are some other BOINC Projects that do not yet support RTX like GPUGRID and a few others.
But the RTX GPU Kicks But in all other BOINC GPU Projects.
http://www.primegrid.com/forum_thread.php?id=8321&nowrap=true#123421
It should Read OpenCL and Not CUDA But I cannot change the Topic Title
AP26 - AP27 Search : Geforce Drivers 4xx.xx Drop more than 2/3 in CUDA Performance from the 3xx.xx Drvers. |
|
|
|
|
It will always be safe during the this challenge to manually select OCL4. It won't break anything.
Good to know. My caution was from warnings in the past about certain transforms exceeding b limits, so that people who had fixed the transform would need to update. Which b limit for which transform on which devices... is a bit much for me to process. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
We had some brief unplanned downtime...
A problem developed with the networking in our server and we had to reboot the server. Although this only takes a few minutes, we don't like to do this during a challenge. This time, however, it was necessary.
My apologies to anyone who may have been affected either by the network problems (lag and/or difficulty connecting) and by the outage while we rebooted. Fortunately, with these long GFN-21 tasks, computers don't have to connect to the server very often so hopefully very few people even noticed there was a problem.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
And here I thought my system was having network issues of it's own. Thanks for the update.
That just goes to show how stable PG is. |
|
|
|
|
|
Michael Goetz wrote: We had some brief unplanned downtime...
Thank you for the explanation and the quick fix !
____________
"Accidit in puncto, quod non contingit in anno."
Something that does not occur in a year may, perchance, happen in a moment. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
Michael Goetz wrote: Furthermore, there's actually no need for the <app> section. All you need is the <app_version> part. This is what I use:
<app_config>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
EDIT: That's for 4 cores. Change the numbers as necessary.
After a bit of experimentation, I'm going to change that.
This is what I now recommend:
<app_config>
<report_results_immediately/>
<app>
<name>genefer</name>
<fraction_done_exact/>
</app>
<app_version>
<app_name>genefer</app_name>
<cmdline>-nt 4</cmdline>
<plan_class>cpuGFN21</plan_class>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>
I added two things.
I added the app section because it seems we need to explicitly specify <fraction_done_exact/> for each app in app_config. The server itself sets this flag, but it appears that if an app appears in app_config it ignores what the server says. This flag improves the accuracy of BOINC's time estimates.
I also added <report_results_immediately/>, which goes outside of any <app> or <app_version> block. You only need this once. I'm not actually sure that this is necessary, but it won't hurt. This too is set on the server, but if the BOINC client is ignoring the server's <fraction_done_exact/> perhaps it's also ignoring the server's <report_results_immediately/>. By including it explicitly you can insure it's turned on.
____________
My lucky number is 75898524288+1 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
After day 3 of 10:
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-14 22:28:21 UTC)
18376 tasks have been sent out. [CPU/GPU/anonymous_platform: 5619 (31%) / 12757 (69%) / 0 (0%)]
Of those tasks that have been sent out:
8497 (46%) came back with some kind of an error. [2178 (12%) / 6319 (34%) / 0 (0%)]
4004 (22%) have returned a successful result. [574 (3%) / 3430 (19%) / 0 (0%)]
5875 (32%) are still in progress. [2867 (16%) / 3008 (16%) / 0 (0%)]
Of the tasks that have been returned successfully:
2183 (55%) are pending validation. [354 (9%) / 1829 (46%) / 0 (0%)]
1808 (45%) have been successfully validated. [220 (5%) / 1588 (40%) / 0 (0%)]
1 (0%) were invalid. [0 (0%) / 1 (0%) / 0 (0%)]
12 (0%) are inconclusive. [0 (0%) / 12 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=333264. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 8.20% as much as it had prior to the challenge!
So far, two GFN-18s have been found during the challenge!?!
Can we find a GFN-21?
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Michael Goetz wrote: I also added <report_results_immediately/>, which goes outside of any <app> or <app_version> block. You only need this once. I'm not actually sure that this is necessary, but it won't hurt.
This works fine even without that line.
You can see it missing from other projects which have their tasks uploaded and only get them reported an hour later.
If something like that had occured on PrimeGrid you'd have noticed this a long time ago. ;-)
____________
Greetings, Jens
92914140^65536+1 |
|
|
AzmodesVolunteer tester
Send message
Joined: 30 Dec 16
Posts: 184
ID: 479275
Credit: 1,564,594,480
RAC: 0
|
I wonder if this is why I am seeing a 5% slowdown.... I'm not at all familiar with the OCL3 or OCL4 transforms...
What do you add to app_config to set it for OCL4?
<app_version>
<app_name>genefer</app_name>
<cmdline>-x ocl4</cmdline>
<plan_class>OCLcudaGFN</plan_class>
</app_version>
Also, you can check which transform was used in the task's details.
____________
Long live the sievers.
+ Encyclopaedia Metallum: The Metal Archives + |
|
|
|
|
|
Azmodes wrote: BOINC downloaded a whole bunch of them for my multi-threaded CPUs despite zero work cache.
When BOINC requests tasks, it evidently ignores the possibility that tasks may occupy more than one logical CPU.
An effective workaround against this is almost obvious: Tell the client that multithreaded tasks occupy only one CPU each, and that the client may only use as many CPUs as you want concurrent CPU tasks. (If you run GPU tasks in the same client instance, tell the client that these use only negligible CPU percentage which vanishes in boinc-client's rounding down of overall CPU usage.)
That is,
- in app_config.xml, set avg_ncpus = 1 for CPU tasks. (And while you are at it, set avg_ncpus = 0.1 or less for GPU tasks.)
- In "Computing preferences...", reduce the number of CPUs used to the number of tasks to run simultaneously.
- Re-read config files. (Or restart the client, if you like boincmgr's display of already downloaded tasks updated with the new settings.)
This recipe is not applicable at projects which use avg_ncpus to control the number of logical CPUs per task, such as Cosmology@Home.
Other means to influence the number of tasks that the client starts simultaneously, for example max_concurrent or project_max_concurrent in app_config, do not directly influence how boinc-client requests tasks, hence do not safely prevent the downloading of too many tasks.
This trick obviously makes it harder to properly run a mix of multithreaded and singlethreaded projects. Personally, whenever I run such a mix, I use separate boinc-client instances for this purpose. |
|
|
|
|
I had five boxes reboot overnight, grrrr, thankfully no driver issues........ (I hope).
I hear ya. One here only, but I woke up and was like why is it so cool in here? GPUs were idling not crunching.
Sincere commiserations to everyone who has been slowed down by this. Penguin posted about 48hiurs into the challenge -- and I just picked his post to quote at random.
I can afford to be sincere, as a CPU-only participant in just one day anyone with a GPU will be way ahead of me, so far that it makes not difference.
This strikes me as being like Ferrari doing a recall on its racing cars half way through Le Mans. ;)
Microsoft does not help its own machines by leaving you with no wiggle room on updates. (And yeas I do understand why they think that is a good idea)
When I was perfomance testing on the 9th, to see how many threads to run concurrently, I started to do the same tests on Windows. OK OK I am a Linux fanboy, but if Windows had run faster I would have used it for the challenge, I am not that one-track-minded
The first machine I tried Windows on locked me out for windows update (probably that gradual roll-out that stream mentioned, because it was a long time. After twenty minutes of configuring updates and no clue how much longer, I forced a reboot of the same hardware running Linux and that is the slower of the two machines I have in the challenge.
After that experience I did not even try to test GFN21 on the faster Windows machine -- that is representing me and Linux too.
My suggestion for the next challenge is to dual boot your machine. Get BOINC running under each OS.
If Windows is your preferred OS, or if it runs faster for the tasks you are running, go with that.
But you will have a tested, working alternate system to fall back on if the Windows BOINC falls over, whether that is caused by Microsoft, or by a virus, or whatever.
(Ditto if it is a Mac -- get it ready to run under Linux as an alternative)
If you feel inclined to take my advice, one further annoyance is that Windows Update usually prevents you from booting into installed alternatives.
My tip is that immediately after you install the dual boot, as soon as you are in your installed Linux system, from the command line run grub-install again, this time installing onto a dedicated USB. This will only install a very few files onto the USB, but will set up the bootloader.
That means you can get back into Linux when Windows Update next deletes the Linux boot info. WU does not trash your whole install, just the bootloader, and that USB gets you back in.
hope that seems useful.
R~~
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
|
|
|
Where does one find the current Overall Points standing for 2018?
The challenge page for 2018 shows little, the link takes you to an external site which only shows 2016 and older.
The 2017 page works correctly though, linking to a Google.docs spreadsheet. |
|
|
HonzaVolunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 15 Aug 05
Posts: 1893
ID: 352
Credit: 3,142,312,174
RAC: 0
                  
|
|
See All Challenge Series Results thread or this link
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186 |
|
|
|
|
|
Thanks Honza, and xii5ku (who replied via other means), for the https://whereismy.coffee/primegrid/ link. Hmmm, where IS my coffee? Ah, there it is, still warm too!
After day 3 of 10:
..........
Of those tasks that have been sent out:
8497 (46%) came back with some kind of an error.......
Is it safe to assume these errors were nearly instant, within seconds or minutes, or have thousands of hours of compute time been lost? |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
Thanks Honza, and xii5ku (who replied via other means), for the https://whereismy.coffee/primegrid/ link. Hmmm, where IS my coffee? Ah, there it is, still warm too!
After day 3 of 10:
..........
Of those tasks that have been sent out:
8497 (46%) came back with some kind of an error.......
Is it safe to assume these errors were nearly instant, within seconds or minutes, or have thousands of hours of compute time been lost?
http://www.primegrid.com/forum_thread.php?id=8326&nowrap=true#123494
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
That is good to know, which I would have already known, had I read just a bit.
///insert big red embarassed face emoticon here/// :D
I want to say something about the large number of errors.
We need to consider the GPU and CPU tasks separately.
The 4326 GPU errors are mostly what I call "3 second errors", i.e., the task fails before it can even start. By far, the most likely cause for this is a Windows 10 update which updated the Nvidia driver with one that doesn't support OpenCL. Without OpenCL, the GeneferOCL (aka GeneferOpenCL) can't find a useable GPU and immediately fails. Rinse and repeat every two or three seconds, and a single computer with a borked video driver can easily rack up 1000 errors in an hour. Meanwhile, a correctly running computer will correctly complete one of these tasks in several hours or days.
Even though there's about 50% errors, that represents only a handful of bad computers.
On the CPU side, there's 1999 errors. We don't see "3 second" failures on CPUs that much. For the most part, most computers are able to at least start running LLR. The large number of errors here has a completely different cause: these tasks were aborted by the user. There's two common reasons for doing this. The first is that BOINC downloaded way too many tasks, for whatever reason. The user then cancels the extra tasks. The second is that the user set the computer up for multi-threading, which will cause BOINC to download extra tasks at the beginning. Many people cancel those extra tasks. Put it all together and over 75% of the CPU tasks with "some sort of error" were in fact aborted by the user. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
After four days:
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-15 22:27:13 UTC)
22192 tasks have been sent out. [CPU/GPU/anonymous_platform: 6103 (28%) / 16089 (72%) / 0 (0%)]
Of those tasks that have been sent out:
10396 (47%) came back with some kind of an error. [2360 (11%) / 8036 (36%) / 0 (0%)]
5663 (26%) have returned a successful result. [862 (4%) / 4801 (22%) / 0 (0%)]
6133 (28%) are still in progress. [2881 (13%) / 3252 (15%) / 0 (0%)]
Of the tasks that have been returned successfully:
2629 (46%) are pending validation. [455 (8%) / 2174 (38%) / 0 (0%)]
3019 (53%) have been successfully validated. [407 (7%) / 2612 (46%) / 0 (0%)]
5 (0%) were invalid. [0 (0%) / 5 (0%) / 0 (0%)]
10 (0%) are inconclusive. [0 (0%) / 10 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=338176. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 9.79% as much as it had prior to the challenge!
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
I'm glad to see the leading edge progress during the challenge, but am sad to say I am on holiday and not participating.
____________
 |
|
|
|
|
|
Michael Goetz wrote: ...
I also added <report_results_immediately/>, which goes outside of any <app> or <app_version> block. You only need this once. I'm not actually sure that this is necessary, but it won't hurt. This too is set on the server, but if the BOINC client is ignoring the server's <fraction_done_exact/> perhaps it's also ignoring the server's <report_results_immediately/>. By including it explicitly you can insure it's turned on.
I am using app config to run GFN multithreaded, and without the <report results immediately> flag.
My first CPU GFN of this challenge completed computation at 01:07:20 and by 01:07:27 had been reported. My impression is that this is about as "immediate" as the upload and report process ever gets.
Boinc client versions may vary on how they treat the interaction between server settings and app config file settings -- the above was with client v 7.6.33 on Linux, if that makes a difference.
As you say, no harm in making the flag explicit, and it's a wise precaution. But, like, don't panic if you left it out...
In other news:
I got up to check at 2am and found it had uploaded almost an hour before, and am now going back to bed happy that I achieved my basic goal of getting at least one task onto the scoreboard, and happy that it completed sooner than I expected (Current projection is that I will have a total of 3 across 2 computers).
And delighted that my laptop with its passive-cooled CPU has beaten a wingmate with a task on a GPU who is yet to return their task. Unexpected surprise there.
There are currently just over 1/3 way through and just over fifty of us at the foot of the scoreboard with one task each (all slightly different scores of course), or about one in six individual participants. If I do get a total of 3 tasks by full time, I am likely to have some company in the three-task zone :)
I have put in a delay before the next task starts, hoping to get a higher b and therefore a few extra points to put me nearer the top of that zone. Tricky decision is how much delay: push it too far and one task will fall off the end, and I have no feeling for how the times will increase.
I will achieve this delay by running a script to allow new work at a preset time. (linux "at" command runs script at appointed time, script calls boinccmd).
Sporting good wishes to fellow CPU crunchers
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
Somewhere between 10% and 20% is probably more realistic. The huge surge of tasks sent out at the beginning of the challenge skews the statistics somewhat, so we won't see quite as much growth for the remaining days as we did in the first 2.
That's still incredible progress, however.
Right now it looks like we're on a course for about 20%, or maybe a bit more. Wow.
____________
My lucky number is 75898524288+1 |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
Somewhere between 10% and 20% is probably more realistic. The huge surge of tasks sent out at the beginning of the challenge skews the statistics somewhat, so we won't see quite as much growth for the remaining days as we did in the first 2.
That's still incredible progress, however.
Right now it looks like we're on a course for about 20%, or maybe a bit more. Wow.
That is very cool. What year did 21 start Mike? Are we looking at a year worth of (normal) work in 10 days? We wouldn't be long finding one if this pace were kept up.
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
Robish wrote: After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
Somewhere between 10% and 20% is probably more realistic. The huge surge of tasks sent out at the beginning of the challenge skews the statistics somewhat, so we won't see quite as much growth for the remaining days as we did in the first 2.
That's still incredible progress, however.
Right now it looks like we're on a course for about 20%, or maybe a bit more. Wow.
That is very cool. What year did 21 start Mike? Are we looking at a year worth of (normal) work in 10 days? We wouldn't be long finding one if this pace were kept up.
September 15th, 2015, as per GFN Status by n-Range (17-22).
It would be really nice to find a GFN-21 during the challenge. Given your luck, especially with GFNs, I fully expect you to be the one who finds it. No pressure. :)
____________
My lucky number is 75898524288+1 |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
Robish wrote: After 2 days:
Since the challenge started, the leading edge has advanced 6.60% as much as it had prior to the challenge!
So if we keep this up we're looking at over 33% advance in the 10 days. That IS truely awesome!
Somewhere between 10% and 20% is probably more realistic. The huge surge of tasks sent out at the beginning of the challenge skews the statistics somewhat, so we won't see quite as much growth for the remaining days as we did in the first 2.
That's still incredible progress, however.
Right now it looks like we're on a course for about 20%, or maybe a bit more. Wow.
That is very cool. What year did 21 start Mike? Are we looking at a year worth of (normal) work in 10 days? We wouldn't be long finding one if this pace were kept up.
September 15th, 2015, as per GFN Status by n-Range (17-22).
It would be really nice to find a GFN-21 during the challenge. Given your luck, especially with GFNs, I fully expect you to be the one who finds it. No pressure. :)
:) somehow I think I may have used up my quota of luck already, but no harm trying :) fingers still crossed.
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
At the halfway point:
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-16 22:41:04 UTC)
26070 tasks have been sent out. [CPU/GPU/anonymous_platform: 6619 (25%) / 19451 (75%) / 0 (0%)]
Of those tasks that have been sent out:
12260 (47%) came back with some kind of an error. [2547 (10%) / 9713 (37%) / 0 (0%)]
7542 (29%) have returned a successful result. [1272 (5%) / 6270 (24%) / 0 (0%)]
6268 (24%) are still in progress. [2800 (11%) / 3468 (13%) / 0 (0%)]
Of the tasks that have been returned successfully:
3093 (41%) are pending validation. [593 (8%) / 2500 (33%) / 0 (0%)]
4431 (59%) have been successfully validated. [679 (9%) / 3752 (50%) / 0 (0%)]
6 (0%) were invalid. [0 (0%) / 6 (0%) / 0 (0%)]
12 (0%) are inconclusive. [0 (0%) / 12 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=343266. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 11.45% as much as it had prior to the challenge!
On the leaderboards, there's been quite the battle raging between the #1 and #2 team positions as well as the #3 and #4 team positions. They keep flipping back and forth. With tasks worth 150K points each, a million point lead can vanish in a heartbeat.
On the individual side, Ryan Propper has a decent lead in first, but #2, #3, and #4 are very close together.
With half the challenge to go, and with server-class Tesla V100-SXM2 GPUs available to rent that can run circles around an RTX 2080 Ti, completing GFN-21 tasks in only 3 hours and 20 minutes, nothing is close to being decided yet.
____________
My lucky number is 75898524288+1 |
|
|
tngSend message
Joined: 29 Aug 10
Posts: 398
ID: 66603
Credit: 22,925,088,044
RAC: 2
                    
|
|
Something else I think worth mentioning is that the leading edge is now searching for primes more than 70 thousand digits larger than at the beginning of the challenge.
____________
 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
Something else I think worth mentioning is that the leading edge is now searching for primes more than 70 thousand digits larger than at the beginning of the challenge.
(log10308008)221 = 11510349
(log10343266)221 = 11609060
11609060 - 11510349 = 98711
____________
My lucky number is 75898524288+1 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
Several off topic posts relating to the recent Mersenne prime have been removed. Please keep all forum posts on topic. Thank you!
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Looks like my goal of being in the top 25% for this challenge is going well... Penguin is currently ranked 74 out of 372 :) Yay! |
|
|
Dad
Send message
Joined: 28 Feb 18
Posts: 284
ID: 984171
Credit: 182,080,291
RAC: 0
      
|
|
OK... Message to Roger...
'We' in the antipodes object to the WINTER solstice challenge
32*C here today and my PCs are melting
Just sayin
Dad
____________
Tonight's lucky numbers are
555*2^3563328+1 (PPS-MEGA)
and
58523466^131072+1 (GFN-17 MEGA) |
|
|
Dad
Send message
Joined: 28 Feb 18
Posts: 284
ID: 984171
Credit: 182,080,291
RAC: 0
|
|
Well done Penguin
You just shot by me on total credit as well!!!!
____________
Tonight's lucky numbers are
555*2^3563328+1 (PPS-MEGA)
and
58523466^131072+1 (GFN-17 MEGA) |
|
|
|
|
|
Dad wrote: ... 32*C here today and my PCs are melting ...
For us, Dad, it's the Summer Solstice Challenge :)
____________
"Accidit in puncto, quod non contingit in anno."
Something that does not occur in a year may, perchance, happen in a moment. |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1774
ID: 126266
Credit: 5,066,569,935
RAC: 0
|
OK... Message to Roger...
'We' in the antipodes object to the WINTER solstice challenge
32*C here today and my PCs are melting
Just sayin
Dad
Want to house swap/country swap dad?, 5*C in Ireland and freezing windy wet and miserable. :)
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
RogerVolunteer developer
Volunteer tester
Send message
Joined: 27 Nov 11
Posts: 1137
ID: 120786
Credit: 267,535,355
RAC: 0
|
OK... Message to Roger...
'We' in the antipodes object to the WINTER solstice challenge
32*C here today and my PCs are melting
Just sayin
Dad
I hear you. I've drafted the 2019 Challenge Calendar and it's just called the Solstice Challenge.
Anyway, off Hiking today. 41 C forecast here in Ipswich QLD later in the week. Luckily Father in Law we're visiting has a pool. |
|
|
|
|
I hear you. I've drafted the 2019 Challenge Calendar and it's just called the Solstice Challenge.
Please don't leave us hanging with the 2019 Challenge Calendar! |
|
|
|
|
|
We have had borealocentrism for so long. As someone living only 3800 km from the North Pole, I cannot see why we should change that 😉 /JeppeSN |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
After 6 days:
Challenge: Winter Solstice
App: 16 (GFN-21)
(As of 2018-12-17 23:15:13 UTC)
29738 tasks have been sent out. [CPU/GPU/anonymous_platform: 7248 (24%) / 22490 (76%) / 0 (0%)]
Of those tasks that have been sent out:
13801 (46%) came back with some kind of an error. [2735 (9%) / 11066 (37%) / 0 (0%)]
9448 (32%) have returned a successful result. [1632 (5%) / 7816 (26%) / 0 (0%)]
6489 (22%) are still in progress. [2881 (10%) / 3608 (12%) / 0 (0%)]
Of the tasks that have been returned successfully:
3377 (36%) are pending validation. [664 (7%) / 2713 (29%) / 0 (0%)]
6034 (64%) have been successfully validated. [964 (10%) / 5070 (54%) / 0 (0%)]
11 (0%) were invalid. [1 (0%) / 10 (0%) / 0 (0%)]
26 (0%) are inconclusive. [3 (0%) / 23 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is b=349120. The leading edge was at b=308008 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 13.35% as much as it had prior to the challenge!
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
There should be a badge for participating in a challenge whilst suffering from a fever that is a prime number... :(
____________
|
|
|
|
|
|
I have noticed a consistent near-coincidence throughout this challenge, so far.
I wonder if it will go away now I mention it.
At the time of this post there are just under 400 different individuals on the scoreboard (392 to be exact) and the current top scorer, Ryan Propper, has returned either 401 or 402 workunits, judging by their score and the largest/smallest WU scores I have seen.
This means that over the challenge so far, the servers have been receiving work back from first-time reporters at almost the same rate as receiving work from Ryan's 56 machines.
That is just a quick look at the figures. After the challenge is over I might do a further analysis about the distribution of work over the scoreboard. Dividing competitors into cohorts of the number of WU returned, which cohort achieved most?
Let me know (here or PM) if you would be interested in that?
Intuition suggests it would be neither Ryanat the top, nor the cohort containing me near the bottom, but somewhere mid table.
Warmly
R~~
____________
My computers found:
9831*21441403+1 is a quadhectokilo prime prime, ie >400,000 digits ;)
252031090528237591 + 65521*149*23*19*17*13*11*7*5*3*2*n is prime for every n in { 0..20 } (an arithemtic progression of 21 primes) |
|
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 644
ID: 164101
Credit: 305,010,093
RAC: 0
|
That is just a quick look at the figures. After the challenge is over I might do a further analysis about the distribution of work over the scoreboard. Dividing competitors into cohorts of the number of WU returned, which cohort achieved most?
Let me know (here or PM) if you would be interested in that?
The "score of the challenge" is currently 1.528.794.800.
The score of the top 15 is 508.435.896 (~ 1/3).
The score of the top 30 is 762.911.924 (~ 1/2).
The score of the top 55 is 1.017.185.044 (~ 2/3).
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
|
I decided to see how fast the multithreaded CPU tasks are taking. In each chart below, p_model is the CPU type, count is the number of validated tasks, and elapsed and cpu are average times are in hours, with the fastest at the top. Each table is for a particular number of threads. I checked from 32 to 2 threads. If a thread count isn't shown, there's no validated tasks that were run with that number of threads.
(Note: a task that was stopped and continued with a different number of threads will show up in both tables.)
32 threads:
+--------------------------------------------------------------------------+-------+--------------------+-------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------------+-------+--------------------+-------------------+
| Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz [Family 6 Model 62 Stepping 4] | 5 | 14.449555555555555 | 461.0215555555556 |
+--------------------------------------------------------------------------+-------+--------------------+-------------------+
31 threads:
+--------------------------------------------------------------------+-------+-------------------+--------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------+-------+-------------------+--------------------+
| Genuine Intel(R) CPU 0000 @ 1.60GHz [Family 6 Model 79 Stepping 0] | 12 | 82.70534722222222 | 506.32962962962966 |
+--------------------------------------------------------------------+-------+-------------------+--------------------+
30 threads:
+-------------------------------------------------------------------------+-------+-------------------+-------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------+-------+-------------------+-------------------+
| Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz [Family 6 Model 45 Stepping 7] | 2 | 31.67222222222222 | 865.5891666666666 |
+-------------------------------------------------------------------------+-------+-------------------+-------------------+
24 threads:
+---------------------------------------------------------------------------+-------+-------------------+--------------------+
| p_model | count | elapsed | cpu |
+---------------------------------------------------------------------------+-------+-------------------+--------------------+
| Intel(R) Xeon(R) CPU E5-2650L v3 @ 1.80GHz [Family 6 Model 63 Stepping 2] | 4 | 91.91930555555555 | 254.00059722222224 |
+---------------------------------------------------------------------------+-------+-------------------+--------------------+
20 threads:
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Xeon(R) CPU E5-2696 v4 @ 2.20GHz [Family 6 Model 79 Stepping 1] | 3 | 15.053796296296296 | 177.41039814814812 |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
19 threads:
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Xeon(R) CPU E5-2696 v4 @ 2.20GHz [Family 6 Model 79 Stepping 1] | 1 | 13.998333333333333 | 240.51594444444444 |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
18 threads:
+------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz [Family 6 Model 85 Stepping 4] | 6 | 11.951944444444445 | 178.51872685185185 |
| Intel(R) Xeon(R) CPU E5-2696 v4 @ 2.20GHz [Family 6 Model 79 Stepping 1] | 4 | 16.78875 | 245.05666666666667 |
| Intel(R) Core(TM) i7-6950X CPU @ 3.00GHz [Family 6 Model 79 Stepping 1] | 1 | 18.399722222222223 | 219.58427777777777 |
+------------------------------------------------------------------------------+-------+--------------------+--------------------+
17 threads:
+--------------------------------------------------------------------------+-------+-------------------+-------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------------+-------+-------------------+-------------------+
| Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz [Family 6 Model 85 Stepping 4] | 13 | 7.643247863247863 | 7.643247863247863 |
+--------------------------------------------------------------------------+-------+-------------------+-------------------+
16 threads:
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Core(TM) i7-5960X CPU @ 3.00GHz [Family 6 Model 63 Stepping 2] | 6 | 22.965694444444445 | 257.2378703703704 |
| Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz [Family 6 Model 45 Stepping 7] | 1 | 31.60527777777778 | 272.02802777777777 |
| Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz [Family 6 Model 45 Stepping 6] | 2 | 32.92305555555556 | 494.1025 |
| Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz [Family 6 Model 79 Stepping 1] | 2 | 39.80319444444444 | 396.2631944444444 |
| AMD Ryzen 7 2700X Eight-Core Processor [Family 23 Model 8 Stepping 2] | 8 | 42.797777777777775 | 359.4068402777778 |
| AMD Ryzen 7 1700 Eight-Core Processor [Family 23 Model 1 Stepping 1] | 2 | 51.21625 | 554.2172222222222 |
| AMD Ryzen Threadripper 1950X 16-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 88.05583333333334 | 414.13694444444445 |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
15 threads:
+---------------------------------------------------------------------------+-------+-------------------+-----------+
| p_model | count | elapsed | cpu |
+---------------------------------------------------------------------------+-------+-------------------+-----------+
| Intel(R) Xeon(R) CPU E5-2687W v2 @ 3.40GHz [Family 6 Model 62 Stepping 4] | 4 | 18.37034722222222 | 224.12275 |
+---------------------------------------------------------------------------+-------+-------------------+-----------+
14 threads:
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Core(TM) i9-7940X CPU @ 3.10GHz [Family 6 Model 85 Stepping 4] | 8 | 8.151527777777778 | 8.151527777777778 |
| Intel(R) Xeon(R) CPU E5-2690 v4 @ 2.60GHz [Family 6 Model 79 Stepping 1] | 11 | 10.371641414141413 | 144.4276868686869 |
| Intel(R) Xeon(R) CPU E5-2683 v3 @ 2.00GHz [Family 6 Model 63 Stepping 2] | 2 | 24.872777777777777 | 240.52897222222222 |
| AMD Ryzen 7 2700X Eight-Core Processor [Family 23 Model 8 Stepping 2] | 2 | 36.61763888888889 | 344.38458333333335 |
+--------------------------------------------------------------------------+-------+--------------------+--------------------+
13 threads:
+-------------------------------------------------------------------------+-------+--------------------+-------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------+-------+--------------------+-------------------+
| Intel(R) Core(TM) i9-7940X CPU @ 3.10GHz [Family 6 Model 85 Stepping 4] | 10 | 19.829416666666663 | 84.30315277777777 |
+-------------------------------------------------------------------------+-------+--------------------+-------------------+
12 threads:
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Core(TM) i9-7940X CPU @ 3.10GHz [Family 6 Model 85 Stepping 4] | 1 | 15.932777777777778 | 82.96869444444444 |
| Intel(R) Core(TM) i9-7960X CPU @ 2.80GHz [Family 6 Model 85 Stepping 4] | 2 | 29.91208333333333 | 145.00077777777778 |
| Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz [Family 6 Model 63 Stepping 2] | 3 | 29.962222222222223 | 265.72064814814814 |
| Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz [Family 6 Model 158 Stepping 10] | 2 | 34.97416666666667 | 175.70765277777778 |
| Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz [Family 6 Model 45 Stepping 7] | 1 | 36.590833333333336 | 345.14972222222224 |
| AMD Ryzen 5 1600X Six-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 37.33027777777778 | 223.98166666666665 |
| Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz [Family 6 Model 79 Stepping 1] | 2 | 37.93277777777778 | 314.4769444444444 |
| AMD Ryzen 7 1700 Eight-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 39.158055555555556 | 438.7925 |
| AMD Ryzen 7 2700 Eight-Core Processor [Family 23 Model 8 Stepping 2] | 1 | 57.495555555555555 | 311.49694444444447 |
| AMD Ryzen 7 2700X Eight-Core Processor [Family 23 Model 8 Stepping 2] | 3 | 58.861296296296295 | 307.62555555555554 |
| Intel(R) Xeon(R) CPU E5-2683 v3 @ 2.00GHz [Family 6 Model 63 Stepping 2] | 5 | 59.116611111111105 | 213.17852777777776 |
| Intel(R) Xeon(R) Gold 6126 CPU @ 2.60GHz [Family 6 Model 85 Stepping 4] | 17 | 86.20480228758171 | 117.05547222222222 |
| AMD Ryzen Threadripper 1920X 12-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 91.13361111111111 | 199.01269444444443 |
| Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz [Family 6 Model 63 Stepping 2] | 1 | 116.7611111111111 | 397.1722222222222 |
| AMD Opteron(TM) Processor 6276 [Family 21 Model 1 Stepping 2] | 1 | 141.10972222222222 | 973.5163888888889 |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
11 threads:
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Xeon(R) CPU E5-2690 v4 @ 2.60GHz [Family 6 Model 79 Stepping 1] | 3 | 10.303333333333333 | 143.36740740740743 |
| Intel(R) Xeon(R) CPU E5-2696 v4 @ 2.20GHz [Family 6 Model 79 Stepping 1] | 28 | 14.667797619047619 | 165.24669642857143 |
| Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz [Family 6 Model 63 Stepping 2] | 12 | 18.90696759259259 | 157.32837037037035 |
| Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz [Family 6 Model 63 Stepping 2] | 2 | 79.14430555555556 | 279.27722222222224 |
| Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz [Family 6 Model 158 Stepping 10] | 4 | 90.32048611111111 | 262.44923611111113 |
| AMD Ryzen Threadripper 1920X 12-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 91.13361111111111 | 199.01269444444443 |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
10 threads:
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Core(TM) i9-7920X CPU @ 2.90GHz [Family 6 Model 85 Stepping 4] | 11 | 11.214116161616163 | 93.39731565656565 |
| Intel(R) Xeon(R) CPU E5-2696 v4 @ 2.20GHz [Family 6 Model 79 Stepping 1] | 2 | 19.861527777777777 | 243.6681527777778 |
| Intel(R) Core(TM) i9-7900X CPU @ 3.30GHz [Family 6 Model 85 Stepping 4] | 17 | 24.241781045751637 | 36.699248366013066 |
| Genuine Intel(R) CPU @ 2.30GHz [Family 6 Model 62 Stepping 2] | 5 | 28.203833333333336 | 28.203833333333336 |
| AMD Ryzen Threadripper 1920X 12-Core Processor [Family 23 Model 1 Stepping 1] | 1 | 91.13361111111111 | 199.01269444444443 |
| Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz [Family 6 Model 158 Stepping 10] | 3 | 91.70972222222223 | 234.45200925925926 |
| AMD Ryzen Threadripper 1950X 16-Core Processor [Family 23 Model 1 Stepping 1] | 13 | 97.63420940170941 | 413.488547008547 |
| Genuine Intel(R) CPU @ 2.30GHz [Family 6 Model 62 Stepping 2] | 4 | 108.64020833333333 | 251.14098611111112 |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
9 threads:
+---------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+---------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Xeon(R) CPU E5-2687W v3 @ 3.10GHz [Family 6 Model 63 Stepping 2] | 4 | 16.082916666666666 | 131.68389583333334 |
| Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz [Family 6 Model 63 Stepping 2] | 1 | 33.131388888888885 | 204.57775 |
| Intel(R) Core(TM) i9-7960X CPU @ 2.80GHz [Family 6 Model 85 Stepping 4] | 1 | 35.56944444444444 | 153.51125 |
+---------------------------------------------------------------------------+-------+--------------------+--------------------+
8 threads:
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
| Intel(R) Core(TM) i9-7960X CPU @ 2.80GHz [Family 6 Model 85 Stepping 4] | 2 | 17.666527777777777 | 120.12772222222223 |
| Intel(R) Xeon(R) CPU E5-2667 v3 @ 3.20GHz [Family 6 Model 63 Stepping 2] | 5 | 21.96977777777778 | 152.35529999999994 |
| Genuine Intel(R) CPU @ 2.90GHz [Family 6 Model 45 Stepping 5] | 5 | 28.62622222222222 | 206.42066111111112 |
| Intel(R) Xeon(R) Silver 4110 CPU @ 2.10GHz [Family 6 Model 85 Stepping 4] | 2 | 29.27402777777778 | 213.537125 |
| Intel(R) Core(TM) i7-9700K CPU @ 3.60GHz [Family 6 Model 158 Stepping 12] | 2 | 33.581805555555555 | 146.63056944444446 |
| Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz [Family 6 Model 158 Stepping 9] | 1 | 34.746944444444445 | 246.72105555555558 |
| Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz [Family 6 Model 45 Stepping 6] | 7 | 36.846666666666664 | 73.69333333333333 |
| AMD Ryzen 7 2700 Eight-Core Processor [Family 23 Model 8 Stepping 2] | 8 | 39.66486111111111 | 228.76832291666668 |
| AMD Ryzen Threadripper 1950X 16-Core Processor [Family 23 Model 1 Stepping 1] | 4 | 42.22909722222222 | 246.79736111111112 |
| Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz [Family 6 Model 60 Stepping 3] | 2 | 45.12861111111111 | 286.6089861111111 |
| AMD Ryzen 7 1700 Eight-Core Processor [Family 23 Model 1 Stepping 1] | 11 | 47.100353535353534 | 239.89435101010102 |
| Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz [Family 6 Model 60 Stepping 3] | 1 | 50.64222222222222 | 354.705 |
| Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz [Family 6 Model 58 Stepping 9] | 5 | 51.330777777777776 | 383.1412777777778 |
| Intel(R) Xeon(R) CPU E5-2623 v3 @ 3.00GHz [Family 6 Model 63 Stepping 2] | 2 | 51.56402777777778 | 279.3659166666667 |
| AMD Ryzen 7 1700X Eight-Core Processor [Family 23 Model 1 Stepping 1] | 12 | 51.703402777777775 | 203.82899768518516 |
| Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz [Family 6 Model 63 Stepping 2] | 1 | 54.178333333333335 | 189.17105555555557 |
| Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz [Family 6 Model 45 Stepping 7] | 8 | 56.06520833333333 | 338.27804513888884 |
| Intel(R) Xeon(R) CPU E5-2690 0 @ 2.90GHz [Family 6 Model 45 Stepping 7] | 3 | 56.75694444444444 | 196.06321296296298 |
| Intel(R) Core(TM) i7-7820X CPU @ 3.60GHz [Family 6 Model 85 Stepping 4] | 53 | 59.40561949685535 | 79.30693186582812 |
| Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz [Family 6 Model 158 Stepping 9] | 2 | 64.8411111111111 | 371.56058333333334 |
| Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz [Family 6 Model 79 Stepping 1] | 3 | 66.98064814814815 | 190.51285185185188 |
| Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz [Family 6 Model 45 Stepping 7] | 1 | 68.62722222222222 | 239.24005555555553 |
| Intel(R) Xeon(R) CPU E5-2680 0 @ 2.70GHz [Family 6 Model 45 Stepping 7] | 2 | 71.75722222222223 | 368.1675 |
| AMD Ryzen 7 2700X Eight-Core Processor [Family 23 Model 8 Stepping 2] | 10 | 72.46552777777778 | 213.03988333333334 |
| Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz [Family 6 Model 45 Stepping 6] | 2 | 79.37361111111112 | 389.5473611111111 |
| Intel(R) Core(TM) i7-4700MQ CPU @ 2.40GHz [Family 6 Model 60 Stepping 3] | 2 | 97.77527777777777 | 451.6144444444444 |
| AMD EPYC 7281 16-Core Processor [Family 23 Model 1 Stepping 2] | 2 | 99.49833333333333 | 532.5823611111111 |
| Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz [Family 6 Model 79 Stepping 1] | 2 | 106.34958333333333 | 197.87354166666665 |
| Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7] | 2 | 110.27194444444444 | 472.6101388888889 |
| Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz [Family 6 Model 94 Stepping 3] | 1 | 122.11305555555556 | 287.6327777777778 |
| AMD Opteron(TM) Processor 6276 [Family 21 Model 1 Stepping 2] | 2 | 129.31402777777777 | 950.5358333333334 |
| AMD Ryzen 5 2400G with Radeon Vega Graphics [Family 23 Model 17 Stepping 0] | 1 | 137.0986111111111 | 387.97555555555556 |
| AMD FX(tm)-8350 Eight-Core Processor [Family 21 Model 2 Stepping 0] | 1 | 138.74277777777777 | 829.2311111111111 |
| AMD Opteron(tm) Processor 6344 [Family 21 Model 2 Stepping 0] | 1 | 154.45888888888888 | 463.37666666666667 |
+-------------------------------------------------------------------------------+-------+--------------------+--------------------+
7 threads:
+---------------------------------------------------------------------------------+-------+--------------------+--------------------+
| p_model | count | elapsed | cpu |
