| Author |
Message |
|---|
|
|
|
First PRPNet Challenge of 2015 is Weiferich, starting at 09:37 UTC on January 20th and running for one week.
More info here:
http://www.primegrid.com/forum_thread.php?id=6083#82385
I've added the Challenge to the Google calendar here:
http://www.primegrid.com/forum_thread.php?id=4788
Also, note that there is a new PRPNet version (5.3.1) as of November 2014. This version fixes a crashing issue for wwww work, so be sure to get this if you're going to crunch WFS in the challenge:
http://www.primegrid.com/forum_thread.php?id=1215
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
 Neo Volunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
                   
|
|
I read the "more info" link...But no info was given as to whether that was a GPU or CPU project...
You can count me in regardless. ATP has dominated Prpnet ports from the dawn of mathematics... (ok... maybe not that far back ;) Prpnet is Rick Reynolds old stomping grounds... :)
Neo
AtP |
|
|
|
|
I read the "more info" link...But no info was given as to whether that was a GPU or CPU project...
It can be either, or both.
In the prpclient.ini file, use this for OpenCL GPU crunching:
// This is the name of the executable used for Wieferich and Wall-Sun-Sun searches.
// Available ONLY to 64bit OS. Uncomment if you have 64bit OS.
//wwwwexe=wwww.exe
wwwwexe=wwwwcl64.exe
or use this for CPU crunching:
// This is the name of the executable used for Wieferich and Wall-Sun-Sun searches.
// Available ONLY to 64bit OS. Uncomment if you have 64bit OS.
wwwwexe=wwww.exe
//wwwwexe=wwwwcl64.exe
You could run both (and even multiples of both) simultaneously from separate folders for GPU + CPU crunching.
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
|
|
|
You'll want a file called "wwww.ini" in your prpclient-N directory. Uncomment the "blocks=XXX" and "threads=XXX" lines. Try values like 8000 for blocks and 2 for threads... but test it on your own box to optimize performance vs. screen lag induced. If you have multiple GPUs in the same box, the "device=X" line can be used in different prpclient-N directories to force wwwwcl to run on different GPUs (devices are numbered starting from 0, on up). I think if you have ATI and Nvidia in the same box, then you use "platform" rather than "device", but I can't test that. You can get wwwwcl to describe the hardware it sees by running wwwwcl from the command line with the "-l" (lower-case 'L') option.
Without wwww.ini, you'll get the default settings for "blocks" and "threads", which yield poor performance. Note the file name: wwww.ini, *not* wwwwcl.ini.
Personally I wouldn't bother running on CPU if you've got a capable GPU, for performance reasons. It's the proverbial "bringing a knife to a gunfight" scenario.
--Gary |
|
|
NeoVolunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
|
|
I read the prior wwww prpnet challenge thread.
Looks like fun. Can't wait to test out the 750 and 770 :)
I checked out the nvidia card wikisheet for "SM"'s....(threads?) both cards look like performers..
Prpnet used to be my old stomping grounds back when proths were 60K digits :)
I love watching the screen output in comparison to the BOINC screen....
A lot of talk on the prior prpnet challenge about this project taking all the cpu cores to feed the gpu's... I'm betting my i5 can handle it... ;)
Neo |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
                     
|
|
I'm just refining my settings after setting up from scratch as per another thread.
Got a wwww.ini. Temps on Speedfan rose. Will play with threads + blocks until it pushes out from my braincell all the postage action from work today (+ tomorrow).
Does it re-read the prpclient.ini at the start of each unit or does it require a ctrl-c each time? Btw I have stopoption=6 so it completes what it's doing.
Where's the link to one's own stats? This should be on the main results page I think. |
|
|
|
|
|
Here's what one of my wwww.ini files look like. You can specify just the info such as nthreads=?? or make it task dependent from what I've read. If you have multiple cards in your system the platform= is important. And if you only have one card make sure you assign it to the correct one. On one of my systems the card is 0 and on the other it is 1.
For those who might be new to PRPNET there is also a prpclient.ini file that contains lots of info on running the correct port and what not. Not least is the team name as it needs to be exact and I've copied what I have here as well.
//wieferich_threads=2
//wieferich_blocks=4000
nthreads=4
blocks=8000
platform=1
team name from prpclient.ini file
// Tests completed by this "team" will be rolled-up as part of team stats. This
// will be recorded on tests that are pending and then updated when tests are
// completed. Since it is stored on the server per test, it is possible for a
// single user to be a member of multiple teams. If no value is specified for
// this field, then completed tests and primes will not be awarded to any teams.
// The teamid cannot have any embedded spaces.
teamid=Aggie_The_Pew
|
|
|
|
|
|
Link to your own stats will look like:
http://u-g-f.de/PRPNet/user_search?uid=Gary_Craig
There's no link on the main page. If you look at the overall stats for a particular port, the user names (leftmost column) are clickable and will take you to the same place as above.
I believe prpclient.ini is only read when you start, not each WU. I further believe that wwww.ini is re-read each WU. Anyone correct me if I'm wrong, but that's what seems correct to me.
With very little recent fine tuning, I was able to get 19.7M p/sec on WFS on my 770. That works out to about 127 seconds per WU. I'll have to re-run some experiments soon to see if I can squeeze more out. One thing I note is that wwwwcl doesn't run the GPU very hot, so it might be possible to get a lot more out by running a second instance in parallel (different prpclient-N directory), if that's a sign of the GPU being under-utilized. On Linux at least wwwwcl is a massive CPU hog. Screen lag can be a problem, so balance that with crunching performance.
--Gary |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
Thanks
Fiddling with 4 instances (2 per GPU) but requires no BOINC as eats CPU time + I want to continue with 321s. Think I'll go for 2 instances with 50% BOINC.
I can 'hear' what it's doing over the RFI through the cheap USB speakers so as long as it sounds "bippy" it seems busier. |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
...So the wwww.exe is a bad idea yes? Stick to cl64? |
|
|
|
|
...So the wwww.exe is a bad idea yes? Stick to cl64?
Yes, I believe the common wisdom is that GPU wwwwcl is about 20 times faster than wwww (many variables there, so hard to give a precise number, but that's probably the right order of magnitude). Leave the CPU free to service the GPU and use anything left over just to run standard BOINC LLR to hunt for primes... that will be my scheme.
G |
|
|
|
|
I believe prpclient.ini is only read when you start, not each WU. I further believe that wwww.ini is re-read each WU. Anyone correct me if I'm wrong, but that's what seems correct to me.
I think prpclient.ini is read right before every server transaction, because that's how prpclient.exe picks up instructions from the stopasapoption variable.
If you are getting one WU at a time, it will get re-read between each WU. If you are getting multiple WUs at a time (I get WFS WUs in blocks of 20) it will get re-read every time an entire block is returned to the prpnet server.
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
|
|
|
+1 to mattozan. I stand corrected.
G |
|
|
|
|
+1 to mattozan. I stand corrected.
G
How come no one ever says "I sit corrected" |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
Metaphorical "stand"point (as opposed to "sitpoint"...). |
|
|
|
|
Metaphorical "stand"point (as opposed to "sitpoint"...).
What about a metaphysical "setpoint"?
Ok, back to normal programming |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
Metaphorical "stand"point (as opposed to "sitpoint"...).
What about a metaphysical "setpoint"?
Ok, back to normal programming
Set point.
Rick Reynolds to serve.
Quiet please ladies & gentleman. |
|
|
|
|
Metaphorical "stand"point (as opposed to "sitpoint"...).
What about a metaphysical "setpoint"?
Ok, back to normal programming
Set point.
Rick Reynolds to serve.
Quiet please ladies & gentleman.
Why is it that golf and tennis (and probably others...) spectators have to be quiet during play? Is the player's genius for the game that fragile? I imagine that some usher standing in a football (either kind) stadium during a game with a big "QUIET" sign would be pelted with rocks and garbage. Similarly, we taunt the Cows during challenges. I just don't get it.
OK, I've gone waaayyyyyy off-topic.
--Gary |
|
|
NeoVolunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
|
|
Whoot!
Just finished my FIRST world record GFN task!! 625K credit...
393,000 seconds for GPU and 392,000 seconds for CPU.... wowsers.
That was crunched on the 770GTX ;)
Going to GFN short for now so I can tinker with the upcoming prpnet challenge this weekend.
Neo
AtP |
|
|
|
|
Whoot!
Just finished my FIRST world record GFN task!! 625K credit...
393,000 seconds for GPU and 392,000 seconds for CPU.... wowsers.
That was crunched on the 770GTX ;)
Going to GFN short for now so I can tinker with the upcoming prpnet challenge this weekend.
Neo
AtP
Did I do my math right? 4 1/2 days? If so that's really good ! (or at least I think that's really good!) |
|
|
|
|
Why is it that golf and tennis (and probably others...) spectators have to be quiet during play? Is the player's genius for the game that fragile? I imagine that some usher standing in a football (either kind) stadium during a game with a big "QUIET" sign would be pelted with rocks and garbage. Similarly, we taunt the Cows during challenges. I just don't get it.
OK, I've gone waaayyyyyy off-topic.
--Gary
I understand why in golf. I don't agree with it so much as understand. Being a golfer, when you are trying to concentrate on your swing external noise makes that difficult and because it started out as a "gentleman's" game and etiquette was the name of the game. I play in a group that has none and yet at least once a round someone gets upset because we were making noise during their swing (results can be very humorous at times - a wrestling match on the next tee box as an example).
However, surely you've noticed the clothing worn by golfers. Talk about loud and in bad taste! If you can't make noise you can distract your opponents eye balls :)
As for the Cows - I dearly miss our banter during challenges and what not. I suppose the horde just out distanced the herd to such a point they no longer feel able to participate. Sad so sad. (Cows if you are listening, buck up and throw down the glove!) |
|
|
|
|
|
Yeah, well said.
Come on Cows... or should I say... CHICKENS? |
|
|
|
|
Whoot!
Just finished my FIRST world record GFN task!! 625K credit...
393,000 seconds for GPU and 392,000 seconds for CPU.... wowsers.
That was crunched on the 770GTX ;)
Going to GFN short for now so I can tinker with the upcoming prpnet challenge this weekend.
Neo
AtP
Did I do my math right? 4 1/2 days? If so that's really good ! (or at least I think that's really good!)
Yeah, that's good. Mine take 10 days on three-year-old GTX 460's.
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
TylerProject administrator
Volunteer tester
Send message
Joined: 4 Dec 12
Posts: 1077
ID: 183129
Credit: 1,280,170,555
RAC: 7,054
               
|
|
I'll be in for this challenge...
On my GTX 760, I run 2 instances (blocks 8192 threads 4) And they take between 115 and 120 seconds each. So 2 WUs every 2 minutes or so. CPU stays at 90- 95% utilization, with just WWWcl64 running.
____________
275*2^3585539+1 is prime!!! (1079358 digits)
 Proud member of Aggie the Pew
 |
|
|
|
|
I'll be in for this challenge...
On my GTX 760, I run 2 instances (blocks 8192 threads 4) And they take between 115 and 120 seconds each. So 2 WUs every 2 minutes or so. CPU stays at 90- 95% utilization, with just WWWcl64 running.
What is your GPU utilization? And how many processor cores do you have?
My system seems CPU-bound for Weifrich. I have two GTX 460's. but a single core CPU. With CPU utilization at 100%, my GPUs are only at about 10-20% utilization, and the WUs are taking ~10 min each.
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
Like a total pansy I strted 2 GFN-WRs today.
O well they'll have to continue from the 27th. |
|
|
|
|
I'll be in for this challenge...
On my GTX 760, I run 2 instances (blocks 8192 threads 4) And they take between 115 and 120 seconds each. So 2 WUs every 2 minutes or so. CPU stays at 90- 95% utilization, with just WWWcl64 running.
What is your GPU utilization? And how many processor cores do you have?
My system seems CPU-bound for Weifrich. I have two GTX 460's. but a single core CPU. With CPU utilization at 100%, my GPUs are only at about 10-20% utilization, and the WUs are taking ~10 min each.
You sound like me. My 770 takes 5 minutes. Horrible usage of the GPU due to the cpu not being fast enough or the cache or the memory. So I just crank up 2 instances and let it hobble along. I was hoping the latest version might help but haven't taken time to install it yet. Waiting on my last GFN shorts to complete before I move over. I've tested all kinds of block sizes and other settings but found that old equipment is just old. |
|
|
|
|
You sound like me. My 770 takes 5 minutes. Horrible usage of the GPU due to the cpu not being fast enough or the cache or the memory. So I just crank up 2 instances and let it hobble along. I was hoping the latest version might help but haven't taken time to install it yet. Waiting on my last GFN shorts to complete before I move over. I've tested all kinds of block sizes and other settings but found that old equipment is just old.
I read elsewhere in the forums that WFS is much more CPU-intensive than WSS. My poor Sempron can't keep up. I tried replacing it with a quad-core Phenom II, but my mobo wouldn't accept it.
I'm just running PPS-Sieve on top of WFS. At least the GPUs aren't underutilized that way. It only slows down each process slightly, and I still get to participate in the challenge...
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
TylerProject administrator
Volunteer tester
Send message
Joined: 4 Dec 12
Posts: 1077
ID: 183129
Credit: 1,280,170,555
RAC: 7,054
|
I'll be in for this challenge...
On my GTX 760, I run 2 instances (blocks 8192 threads 4) And they take between 115 and 120 seconds each. So 2 WUs every 2 minutes or so. CPU stays at 90- 95% utilization, with just WWWcl64 running.
What is your GPU utilization? And how many processor cores do you have?
My system seems CPU-bound for Weifrich. I have two GTX 460's. but a single core CPU. With CPU utilization at 100%, my GPUs are only at about 10-20% utilization, and the WUs are taking ~10 min each.
GPU Utilization is erratic, bouncing around the upper 90s. I have a quad core CPU, i5-2500k @ 4.5GHz. 1866MHz RAM.
It is also noteworthy that they use a lot of RAM, each instance using up to 1.25GB at times..
____________
275*2^3585539+1 is prime!!! (1079358 digits)
Proud member of Aggie the Pew
|
|
|
mackerelVolunteer tester
Send message
Joined: 2 Oct 08
Posts: 2460
ID: 29980
Credit: 442,802,854
RAC: 10,291
                 
|
|
Setting up to join in now. Is the challenge actually getting it to run at all in the first place???
Found I had to get a download from MS to run it at all, then I had to copy the wwww.ini into the instance folder before it was recognised. Now I'm finally playing with the block/thread settings...
Edit for reference:
Q6600 >3 cores, GTX560Ti 70%+, 121s, 4T 8000B
i7 2600 <2 cores, HD6850 90%, 151s if I'm not using PC, 6T 480B
2nd PC is my main desktop and I can't take too much graphical slowdown. I found a small blocksize helps with smoother video and general screen updates. Threads don't seem to matter as much. |
|
|
|
|
|
I'm up and running on this, though my contribution will be somewhat reduced for the first day or two due to working off SoB and PSP tasks after the BOINC challenge.
Memory usage on Linux appears to vary, but it is large. 375M to 475M resident is what I'm seeing. LLR resident usage generally doesn't exceed 100M.
G |
|
|
NeoVolunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
|
Found I had to get a download from MS to run it at all, then I had to copy the wwww.ini into the instance folder before it was recognised. Now I'm finally playing with the block/thread settings...
I had the same problem. wwww said I was missing a file (can't remember it exactly) but something like "vbl110.dll" ... so I went to microsoft and downloaded two libraries for a visual basic 11 and/or 12...
I haven't had time to re-attempt getting the project to run. Been working late and going to work early.. Barely have enough time to check stats...
I'm excited to learn that wwww uses sooo much memory.... I think my memory gets lonely and asks why I never utilize it. :) I've got 2400 mhz memory :)
Maybe tonight I can jump in the fray of things..
Neo |
|
|
mackerelVolunteer tester
Send message
Joined: 2 Oct 08
Posts: 2460
ID: 29980
Credit: 442,802,854
RAC: 10,291
|
|
I was happy to see I was in 11th place this morning, until I realised there were only 26 participants at the time. Now I'm down to 13th... if I have time tonight I might be able to get one more GPU on it but it will not be an easy fight as I think the spare system I had was one which didn't like turning on. It is fine when it does eventually turn on though!
I tried both 32 and 64 bit versions and had different missing files. I think it was something like MSVC++ 2010 something or other that got it working for me.
Ram usage on my slower CPU box is just over 1GB, and not sure what it is on my fast CPU slower GPU one but I did see over 2GB used at one point while I was playing with the settings. I think my slower box is CPU limited as it takes almost all of 4 cores, and I can't get GPU usage beyond about 70% regardless of the settings and instances. |
|
|
|
|
|
Got 2 instances running on my 770 and 2 on my 660ti but those are taking 10 minutes + to run. That pc is so old and slow it hates WFS. Now the WSS jobs run great. And one instance running on my 460.
And I've notice a new file call memory leakage or something. Not sure how to read it but I'm sure it tells me that my system is bleeding out !
Go AtP |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
<Slowly turns dial upwards>
Anyone else noticed we appear to have as much credit as everyone else put together (@ this early stage)? |
|
|
|
|
<Slowly turns dial upwards>
Anyone else noticed we appear to have as much credit as everyone else put together (@ this early stage)?
Yep, we're doing well so far. Mr. Z. is in the lead and there are a lot of other Rats in the chase. I'm running two instances on my 770 and two on 570, both ubuntu 14.04. I did try running on my macbook, but screen lag was intolerable so I'd have to "de-tune" in wwww.ini so much as to make it worthless... it's pretty slow anyway.
Watch out for the "star people" and other big teams like PST and the Germans. We've a long road ahead.
G |
|
|
|
|
|
I don't think zzcmeyer is on a team but he sure is moving along. Not as well as our Z but not bad either.
I've noticed that AtP takes alot of pride in the PRPNet side :)
edit: not sure turning the dial up slowly is the correct phrase more like jump out of the way or the horde will run right over your backside leaving lots of tiny foot prints. |
|
|
|
|
|
I would normally be in the mix but I'm SoBing away still, and I've still not fixed my 580 properly so it would be virtually worthless. But I thought I'd post to say keep up the good Wieferiching - I would be extremely pleased if AtP were indeed to "go". |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
not sure turning the dial up slowly is the correct phrase more like jump out of the way or the horde will run right over your backside leaving lots of tiny foot prints.
I was overclocking tbe 580s 1% at a time - quit while I was ahead at 110% or 863MHz - 24~25M/sec, 1m05s~1m10s. 2 threads 8000 blocks (couldn't see any difference when I changed these to e.g 4, 16000 etc). |
|
|
|
|
I would normally be in the mix but I'm SoBing away still,...
Pull it together man, you're making a spectacle of yourself!
|
|
|
|
|
|
AtP is currently lapping the team competition. Z and zz however are in a good battle for 1st on the individual side. Go Z ! |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
Zick Zeynolds.
Zrinktastee.
Zave.
This could catch on.
My freezer's been off for nearly a day by accident. |
|
|
|
|
My freezer's been off for nearly a day by accident.
One can only hope nothing thawed out.
|
|
|
NeoVolunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
|
|
wow... joining this challenge has been a pain in the ass.
Anyways,
I finally got it working. Seems to me that something is wrong.
I have my 770 working. Nothing else CPU wise.
Only 1 instance ...
I'm getting 98% cpu usage on my i5 @ 3.7 ghz, 2400 mhz ram, and 99% gpu usage.
220 seconds per workunit. 37Mp/sec
How can this program use my ENTIRE 4 cores?
Should I take a picture "snipit" and post it?
Neo
|
|
|
NeoVolunteer tester
Send message
Joined: 28 Oct 10
Posts: 710
ID: 71509
Credit: 91,178,992
RAC: 0
|
|
Output says:
Clock time 77.60 second at 31863876 p/sec.
Processor time: 225.11 sec. (9.72 init + 215.39 sieve)
Seconds spent in CPU and GPU: 94.81 (cpu), 158.14 (gpu)
CPU/GPU utilization: 2.90 (cores), 1.00 (devices)
Percent of GPU time waiting for GPU: 14.83
Neo |
|
|
|
|
Output says:
Clock time 77.60 second at 31863876 p/sec.
Processor time: 225.11 sec. (9.72 init + 215.39 sieve)
Seconds spent in CPU and GPU: 94.81 (cpu), 158.14 (gpu)
CPU/GPU utilization: 2.90 (cores), 1.00 (devices)
Percent of GPU time waiting for GPU: 14.83
Neo
That look about right.
Its kind of a pig. Takes your whole system to run it. |
|
|
mackerelVolunteer tester
Send message
Joined: 2 Oct 08
Posts: 2460
ID: 29980
Credit: 442,802,854
RAC: 10,291
|
|
My observation: higher GPU speed needs more CPU speed to feed it. And it can take a LOT of CPU. You may or may not be able to optimise it a bit further if you play with the blocks and threads values. On my systems at least, a high block value seems to help reduce CPU a bit, but makes screen draw much laggier so not good on a system that has to be used. |
|
|
TylerProject administrator
Volunteer tester
Send message
Joined: 4 Dec 12
Posts: 1077
ID: 183129
Credit: 1,280,170,555
RAC: 7,054
|
My observation: higher GPU speed needs more CPU speed to feed it. And it can take a LOT of CPU. You may or may not be able to optimise it a bit further if you play with the blocks and threads values. On my systems at least, a high block value seems to help reduce CPU a bit, but makes screen draw much laggier so not good on a system that has to be used.
2 instances uses 80-95% of my CPU (i5-2500k @ 4.5GHz) at any given time. GPU (GTX 760) usage is in the mid 90s.
____________
275*2^3585539+1 is prime!!! (1079358 digits)
Proud member of Aggie the Pew
|
|
|
|
|
|
I run 2 instances each on both my 570 and 770. Blocks/threads are 8192/4. In both cases, this sucks 3 out of 4 available i7 cores, almost. Ubuntu 14.04. I occupy the remaining core(s) with challenge clean-up (hopefully... ESP) work.
I made the mistake of accepting another o/s upgrade on one Linux box during the challenge last night. It wiped out the nvidia driver and OpenCL so I was down for a few hours. I had a black screen with a flashing cursor; that's all. I had to "ssh" in from my iMac just to log in, and then re-install the 304.125 driver. All is working now. What a fuss.
Go AtP! Go zunewantan!
--Gary |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2829
ID: 130544
Credit: 954,747,840
RAC: 5,071
|
|
24M/psec standard 2 8000 wwww.cfg 80%+ utilisation 1 instance 110~120sec on each GTX580. |
|
|
|
|
|
Envious of these high comps/sec and low run times. WFS is just not friendly on old cpu's. Wonder what the big difference is between WFS and WSS as those babies scream. I've heard there's a totally different programming scheme for the 2.
I will probably drop out of this challenge soon and get back to running GFN's and pps sieves at the same time. I might even test dual gfn's just to see if it works.
Rock on AtP! |
|
|
|
|
|
Update on AtP - We are #1 with a bullet and the individual stats are as follows:
rank UID tests done Score
2 zunewantan 63,660 63,660,000
7 brinktastee 9,827 9,827,000
9 Gary_Craig 8,047 8,047,000
13 Dave 5,028 5,028,000
15 Rick_Reynolds 4,589 4,589,000
16 mackerel 3,942 3,942,000
17 1998golfer 3,760 3,760,000
24 Neo 1,563 1,563,000
30 mattozan 453 453,000
ps - if I missed someone please let me know as I don't want to leave anyone out.
Cheers
|
|
|
|
|
|
I think maybe the challenge is over. Congrats to all you folks who ran. AtP kicked again! and Z was 2nd individually.
I see some handwriting on the wall as the star team looks to have some folks interested in these challenges now so we might have some tough competition.
Great job you ratties !!!
Cheers Rick
ps - will leave this sticky for a few (or if you want you can unsticky it) |
|
|
|
|
|
Edit: Ooops, I made a mistake, was thinking about the wrong challenge! |
|
|
|
|
|
Apparently there was some snafu with the challenge scoring in that WFS work continued to be included in the challenge standings for several hours after it was supposed to end. Unsure if it is fixable. It didn't affect my personal placement, stuck in the #9 spot almost the whole way.
Great job, team.
--Gary |
|
|
