log in |
81)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9772)
Posted 18 Mar 2024 by rebirther you need to add support for every cc level explicitly. ok, noticed, thx, we need to change that again |
82)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9770)
Posted 18 Mar 2024 by rebirther was v31 compiled with explicit support for CC_6.0 in the NVCCFLAGS section of your makefile or build script? yes it was but with cc5 |
83)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9767)
Posted 18 Mar 2024 by rebirther Interesting, try to run ldd ./mfaktc |
84)
Message boards :
Number crunching :
Bases loaded
(Message 9765)
Posted 18 Mar 2024 by rebirther S323 - n=100-300k - runtime 8min-1h12min (AVX@3.8Ghz) - 100-150k = 60 credits - 150-200k = 110 credits - 200-250k = 160 credits - 250-300k = 260 credits - Sierpinski Base - deadline 3 days |
85)
Message boards :
News :
Testing new apps started
(Message 9763)
Posted 17 Mar 2024 by rebirther new cuda120 v31 TF mfaktc linux app is up changelog: - reduced the driver requuirement from v30 |
86)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9762)
Posted 17 Mar 2024 by rebirther That fixed one of my PCs. v31 is up, should reduce the driver requirement to 525 |
87)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9759)
Posted 17 Mar 2024 by rebirther All errors with new version. v30 fixed the issue with a wrong and older.so12 file |
88)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9758)
Posted 17 Mar 2024 by rebirther All errors with new version. ok, this was not as planned, reverted back to 0.28, need to check... |
89)
Message boards :
News :
base R815 Magaprime / proven
(Message 9756)
Posted 16 Mar 2024 by rebirther magic_sam, a member of the team Gridcoin found a megaprime for base R815. The prime 8*815^559138-1 has 1.627.740 digits and entered the TOP5000 in Chris Caldwell's The Largest Known Primes Database. With this find it also has proven the base! |
90)
Message boards :
News :
Testing new apps started
(Message 9755)
Posted 16 Mar 2024 by rebirther new cuda120 TF mfaktc app is up changelog: - recompiled linux app from cc5-latest |
91)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9754)
Posted 16 Mar 2024 by rebirther new cuda120 TF mfaktc app is up changelog: - recompiled from cc5-latest |
92)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9753)
Posted 14 Mar 2024 by rebirther there also appears to be a problem with the cuda100 package. The file must be somewhere in another thread. Linux is not my thing so I will ask if we can do something. The package need a recompile too, old stuff. |
93)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9751)
Posted 14 Mar 2024 by rebirther there also appears to be a problem with the cuda100 package. the libcudart.so.10 file you can find in FAQs. If the app was dynamically compiled you dont need them in package. The .exe in linux was also compiled so but running under linux. The .ini file is fixed. |
94)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9749)
Posted 14 Mar 2024 by rebirther cuda 12+ still supports CC 6.0 devices. why not just recompile the cuda120 app with CC 6.0 support? Yes, we will try to recompile, perhaps this will solve a lot of problems. |
95)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9747)
Posted 14 Mar 2024 by rebirther cuda 12+ still supports CC 6.0 devices. why not just recompile the cuda120 app with CC 6.0 support? cuda12 app support cc6.1, see 1070ti I dont know why the Tesla doesnt run cuda120 WUs, anonymous patform is not allowed. |
96)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9745)
Posted 14 Mar 2024 by rebirther I'm trying to say that the situation is worse now for some reason. With your help, I was able to process tasks for a time (on P100 only computers) using the first app_config you posted, but now that one no longer helps, nor does the second app_config with plan_class help either. I can't process tasks on P100s at all now, whether they are alone in the computer, in pairs, or mixed with another GPU. Nothing works now. The server plan_class has the original. You need the app_config in the first post. |
97)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9743)
Posted 14 Mar 2024 by rebirther Also my P100-only machines are having problems again despite having got them working with the initial app_config you posted. Yes, you need the app_config also a 2nd entry for the other card if you have 2 |
98)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9740)
Posted 13 Mar 2024 by rebirther It could be that since the P100 is the second GPU, it's presence isn't being properly detected on the server side. That's just a wild guess based on the fact that the list of hosts always claims the computer has x number of whatever the first GPU is, instead of what's really there. Yes but with the app_config we can reduce the possibilities. |
99)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9738)
Posted 13 Mar 2024 by rebirther I'm sorry, after getting rid of the app_config and resetting the project, the client is still getting cuda120 tasks which error out immediately on the P100. Yeah I see, we have 2 plan_classes and both are working so we need it all in app_config manually to overwrite the settings. <app_config>
<app_version>
<plan_class>cuda100</plan_class>
<host_summary_regex>Tesla P100</host_summary_regex>
<min_gpu_ram_mb>384</min_gpu_ram_mb>
<gpu_ram_used_mb>384</gpu_ram_used_mb>
<gpu_peak_flops_scale>0.22</gpu_peak_flops_scale>
<cpu_frac>0.01</cpu_frac>
</app_version>
<app_version>
<plan_class>cuda120</plan_class>
<host_summary_regex>GeForce GTX 1070 Ti</host_summary_regex>
<min_gpu_ram_mb>384</min_gpu_ram_mb>
<gpu_ram_used_mb>384</gpu_ram_used_mb>
<gpu_peak_flops_scale>0.22</gpu_peak_flops_scale>
<cpu_frac>0.01</cpu_frac>
</app_version>
<app_config>
|
100)
Message boards :
Number crunching :
Nvidia Tesla P100 Problems
(Message 9736)
Posted 13 Mar 2024 by rebirther I have added the cuda100 plan_class for Tesla P100 to the server, better for both sides. You can get rid of the app_config. Lets test it. |