Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
11)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 54758)
Posted 4149 days ago by  Odicin
Yeah, this one works! Thx@Lennart.
But the x86 application crashes also on my system half the way because of ntdll.dll:
Name der fehlerhaften Anwendung: wwwwcl.exe, Version: 0.0.0.0, Zeitstempel: 0x4fb91fb0
Name des fehlerhaften Moduls: ntdll.dll, Version: 6.1.7601.17725, Zeitstempel: 0x4ec49b60
Ausnahmecode: 0xc0000264
Fehleroffset: 0x000a1742
ID des fehlerhaften Prozesses: 0x1ab8
Startzeit der fehlerhaften Anwendung: 0x01cd36c686883323
Pfad der fehlerhaften Anwendung: C:\prpnet\prpclient-gpu\wwwwcl.exe
Pfad des fehlerhaften Moduls: C:\Windows\SYSTEM32\ntdll.dll
Berichtskennung: 65c5cfcb-a2ba-11e1-98d8-404e57434401
Regards Odi
|
12)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 54755)
Posted 4149 days ago by Odicin
Same problem here with prpclient 5.0.8 and latest wwwwcl 2.1.5:
[2012-05-20 21:59:00 MS] PRPNet Client application v5.0.8 started
[2012-05-20 21:59:00 MS] User name odicin at email address is XX@XXXXX.de
[2012-05-20 21:59:01 MS] WIEFERICH: Getting work from server prpnet.mine.nu at port 13000
[2012-05-20 21:59:01 MS] WIEFERICH: ERROR: The client must run wwww to use this server.
[2012-05-20 21:59:02 MS] WALLSUNSUN: Getting work from server prpnet.mine.nu atport 13001
[2012-05-20 21:59:02 MS] WALLSUNSUN: ERROR: The client must run wwww to use this server.
[2012-05-20 21:59:02 MS] Could not connect to any servers and no work is pending. Pausing 3 minutes
Could x86 version of wwwwcl a problem because former version were limit on x64?
Regards Odi
|
13)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 54510)
Posted 4154 days ago by Odicin
I tested the new one on my GT440.
wwwwcl 2.0:
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b4000
Elapsed time: 590.40 sec. (3.79 init + 586.40 sieve) at 4295290 p/sec.
Processor time: 612.65 sec. (3.65 init + 609.00 sieve).
Seconds spent in CPU and GPU: 600.78 (cpu), 2140.41 (gpu)
Percent of time spent in CPU vs. GPU: 0.22 (cpu), 0.78 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.72 (devices)
wwwwcl 2.1.2:
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b4000
Elapsed time: 744.01 sec. (3.71 init + 728.88 sieve) at 3461158 p/sec.
Processor time: 350.16 sec. (2.67 init + 347.49 sieve).
Seconds spent in CPU and GPU: 11.32 (cpu), 1353.72 (gpu)
Percent of time spent in CPU vs. GPU: 0.01 (cpu), 0.99 (gpu)
CPU/GPU utilization: 0.30 (cores), 35.56 (devices)
The throughput falled rapidly but gpu is used better than before...
Regards Odi
|
14)
Message boards :
Proth Prime Search :
Faulty batch PPS_Sr2 W.U.s
(Message 54032)
Posted 4163 days ago by Odicin
Incidentally Collatz is the only BOINC project I have found for pre HD3xxx cards.
And with Moowrap the last for HD3XXX cards, since milkyway dropped HD3XXX support with opencl migration.
I know, there is also a HD3XXX version of lunatics for sah, but they use a lot of cpu resources.
Regards Odi
|
15)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 53688)
Posted 4167 days ago by Odicin
It would be interesting if you vary -t and -b more.
I tested the x86 win app also on a nvidia GT440 with different parameters:
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b2000
Elapsed time: 655.17 sec. (18.86 init + 636.19 sieve) at 3870698 p/sec.
Processor time: 616.72 sec. (19.22 init + 597.50 sieve).
Seconds spent in CPU and GPU: 664.37 (cpu), 2071.85 (gpu)
Percent of time spent in CPU vs. GPU: 0.24 (cpu), 0.76 (gpu)
CPU/GPU utilization: 0.21 (cores), 0.64 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b4000
Elapsed time: 590.40 sec. (3.79 init + 586.40 sieve) at 4295290 p/sec.
Processor time: 612.65 sec. (3.65 init + 609.00 sieve).
Seconds spent in CPU and GPU: 600.78 (cpu), 2140.41 (gpu)
Percent of time spent in CPU vs. GPU: 0.22 (cpu), 0.78 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.72 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b8000
Elapsed time: 623.09 sec. (3.08 init + 619.79 sieve) at 4069956 p/sec.
Processor time: 627.64 sec. (3.01 init + 624.63 sieve).
Seconds spent in CPU and GPU: 634.60 (cpu), 2315.38 (gpu)
Percent of time spent in CPU vs. GPU: 0.22 (cpu), 0.78 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.73 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t5 -b16000
Elapsed time: 622.91 sec. (3.54 init + 618.97 sieve) at 4071112 p/sec.
Processor time: 648.48 sec. (3.39 init + 645.10 sieve).
Seconds spent in CPU and GPU: 632.87 (cpu), 2383.03 (gpu)
Percent of time spent in CPU vs. GPU: 0.21 (cpu), 0.79 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.76 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t2 -b4000
Elapsed time: 877.02 sec. (2.77 init + 874.16 sieve) at 2891575 p/sec.
Processor time: 834.92 sec. (2.65 init + 832.27 sieve).
Seconds spent in CPU and GPU: 886.77 (cpu), 966.86 (gpu)
Percent of time spent in CPU vs. GPU: 0.48 (cpu), 0.52 (gpu)
CPU/GPU utilization: 0.50 (cores), 0.55 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t8 -b4000
Elapsed time: 596.75 sec. (4.50 init + 591.96 sieve) at 4249589 p/sec.
Processor time: 607.87 sec. (4.23 init + 603.65 sieve).
Seconds spent in CPU and GPU: 607.30 (cpu), 3813.20 (gpu)
Percent of time spent in CPU vs. GPU: 0.14 (cpu), 0.86 (gpu)
CPU/GPU utilization: 0.13 (cores), 0.79 (devices)
C:\prpnet\prpclient-gpu>wwwwcl -p1334568e11 -P1334569e11 -TWieferich -t10 -b4000
Elapsed time: 614.21 sec. (6.13 init + 607.78 sieve) at 4128787 p/sec.
Processor time: 619.71 sec. (5.76 init + 613.96 sieve).
Seconds spent in CPU and GPU: 624.95 (cpu), 4768.02 (gpu)
Percent of time spent in CPU vs. GPU: 0.12 (cpu), 0.88 (gpu)
CPU/GPU utilization: 0.10 (cores), 0.77 (devices)
@rogue: Is it possible to add a function which do some short tests with settings for -t and -b and give an output to determine the optimal setting/throughput?
Regards Odi
|
16)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 53687)
Posted 4167 days ago by Odicin
The error with clGetKernelWorkGroupInfo is due to the requirement for OpenCL 1.1. The code in wwwwcl is not supported by OpenCL 1.0. Can anyone tell me the minimum driver needed to support OpenCL 1.1? I'll add that information to the readme.
According the nvidia developer forum since 280.13/280.19 beta and the official build was the 280.26 from 2011-08-09. Source: http://forums.nvidia.com/index.php?showtopic=206951
Edith says: too late ;)
Regards Odi
|
17)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 53610)
Posted 4168 days ago by Odicin
I tested the compiled win versions. After an error because of missing MSVCP100.dll I installed the Microsoft Visual C++ 2010 Redistributable Package and the the application started.
Test machine was an 8core server with 3 HD5450, 6 cpu cores running but not other gpu work. I tested the same range linke rouge:
x86:
C:\prpnet\prpclient-gpu1>wwwwcl_x86.exe -p1334568e11 -P1334569e11 -TWieferich -t5 -b2000
wwwwcl v2.0
Sieve started: 133456800000000000 <= p < 133456900000000000
p=133456803694770223, 1.542M p/sec, 2.34 CPU cores, 3.7% done. ETA 01 May 12:33
[...]
p=133456896430784287, 1.544M p/sec, 2.35 CPU cores, 96.4% done. ETA 01 May 12:33
Sieve complete: 133456800000000001 <= p < 133456900000000000 2535955284 primes tested
Elapsed time: 1653.19 sec. (7.24 init + 1645.72 sieve) at 1533977 p/sec.
Processor time: 3835.47 sec. (10.75 init + 3824.72 sieve).
Seconds spent in CPU and GPU: 1663.88 (cpu), 6831.04 (gpu)
Percent of time spent in CPU vs. GPU: 0.20 (cpu), 0.80 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.82 (devices)
x64:
C:\prpnet\prpclient-gpu1>wwwwcl_x64.exe -p1334568e11 -P1334569e11 -TWieferich -t5 -b2000
wwwwcl v2.0
Sieve started: 133456800000000000 <= p < 133456900000000000
p=133456803664459411, 1.537M p/sec, 2.29 CPU cores, 3.7% done. ETA 01 May 13:02
[...]
p=133456899903374813, 1.541M p/sec, 2.34 CPU cores, 99.9% done. ETA 01 May 13:02
Sieve complete: 133456800000000001 <= p < 133456900000000000 2535955284 primes tested
Elapsed time: 1654.47 sec. (6.33 init + 1647.78 sieve) at 1532791 p/sec.
Processor time: 3811.70 sec. (10.03 init + 3801.67 sieve).
Seconds spent in CPU and GPU: 1664.52 (cpu), 6872.51 (gpu)
Percent of time spent in CPU vs. GPU: 0.19 (cpu), 0.81 (gpu)
CPU/GPU utilization: 0.20 (cores), 0.83 (devices)
Regards Odi
|
18)
Message boards :
Project Staging Area :
Call for wwwwcl beta testers (OpenCL)
(Message 53598)
Posted 4169 days ago by Odicin
Some HD 5450, one old Mobility HD4330 and a couple of different Nvidia cards available for testing.
Regards Odi
|
19)
Message boards :
Project Staging Area :
The extended Sierpinski Problem
(Message 53550)
Posted 4170 days ago by Odicin
The port seems down at the moment...
Regards Odi
|
20)
Message boards :
Sieving :
Cullen/woodall sieve
(Message 53453)
Posted 4173 days ago by Odicin
Look at here: http://www.primegrid.com/forum_thread.php?id=2557#23939
This error occured a time ago. Maybe the app also need some optimization.
Regards Odi
|
Next 10 posts
|
