An easy answer in the meantime would be to only validate each brand of gpu against itself, telling people that this could/WILL slow down validation a bit but in the end should produce more valid tasks for each individual user.
That involves the project admins and developers to run three different validator processes to segregate the results. Then still have to compare those results against the other app card types.
That seems a bit involved for that to actually happen then, oh well.
I have installations of BOINC on Linux Pop!_OS (Gnome 44) and Windows 11 Pro on my dual booting computer. The computer has two video cards, an Nvidia RTX 3080 and an Nvidia Titan X Pascal. I have a failure rate of 8.7% (169 of 1935) on GPU tasks when I run Linux. The failure rate of GPU tasks for Windows 11 Pro is 0.46% (1 of 217).
These results were obtained while running the same CPU and GPU speeds, not overclocked, on both installations.
Windows:
38 PM | | Starting BOINC client version 7.20.2 for windows_x86_64
6/1/2023 9:06:38 PM | | log flags: file_xfer, sched_ops, task
6/1/2023 9:06:38 PM | | Libraries: libcurl/7.84.0-DEV Schannel zlib/1.2.12
6/1/2023 9:06:38 PM | | Data directory: C:\ProgramData\BOINC
6/1/2023 9:06:38 PM | | Running under account veste
6/1/2023 9:06:38 PM | | CUDA: NVIDIA GPU 0: NVIDIA GeForce RTX 3080 (driver version 535.98, CUDA version 12.2, compute capability 8.6, 10240MB, 10240MB available, 29768 GFLOPS peak)
6/1/2023 9:06:38 PM | | CUDA: NVIDIA GPU 1: NVIDIA TITAN X (Pascal) (driver version 535.98, CUDA version 12.2, compute capability 6.1, 12288MB, 12288MB available, 10974 GFLOPS peak)
6/1/2023 9:06:38 PM | | OpenCL: NVIDIA GPU 0: NVIDIA GeForce RTX 3080 (driver version 535.98, device version OpenCL 3.0 CUDA, 10240MB, 10240MB available, 29768 GFLOPS peak)
6/1/2023 9:06:38 PM | | OpenCL: NVIDIA GPU 1: NVIDIA TITAN X (Pascal) (driver version 535.98, device version OpenCL 3.0 CUDA, 12288MB, 12288MB available, 10974 GFLOPS peak)
6/1/2023 9:06:38 PM | | [error] bad value 1685678676.998499 of time stats last update; ignoring
6/1/2023 9:06:38 PM | | Windows processor group 0: 20 processors
6/1/2023 9:06:38 PM | | Host name: ANewFreedom
6/1/2023 9:06:38 PM | | Processor: 20 GenuineIntel Intel(R) Core(TM) i9-10850K CPU @ 3.60GHz [Family 6 Model 165 Stepping 5]
6/1/2023 9:06:38 PM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 fma cx16 sse4_1 sse4_2 movebe popcnt aes f16c rdrandsyscall nx lm avx avx2 tm2 pbe fsgsbase bmi1 smep bmi2
6/1/2023 9:06:38 PM | | OS: Microsoft Windows 11: Professional x64 Edition, (10.00.22621.00)
6/1/2023 9:06:38 PM | | Memory: 31.87 GB physical, 36.62 GB virtual
6/1/2023 9:06:38 PM | | Disk: 418.48 GB total, 329.88 GB free
6/1/2023 9:06:38 PM | | Local time is UTC -4 hours
6/1/2023 9:06:38 PM | | No WSL found.
6/1/2023 9:06:38 PM | | VirtualBox version: 6.1.34
6/1/2023 9:06:38 PM | | Config: use all coprocessors
Linux:
Fri 02 Jun 2023 06:01:11 AM EDT | | Starting BOINC client version 7.22.1 for x86_64-pc-linux-gnu
Fri 02 Jun 2023 06:01:11 AM EDT | | log flags: file_xfer, sched_ops, task
Fri 02 Jun 2023 06:01:11 AM EDT | | Libraries: libcurl/7.86.0-DEV OpenSSL/3.0.8 zlib/1.2.13 nghttp2/1.52.0
Fri 02 Jun 2023 06:01:11 AM EDT | | Data directory: /home/vester
Fri 02 Jun 2023 06:01:12 AM EDT | | CUDA: NVIDIA GPU 0: NVIDIA GeForce RTX 3080 (driver version 525.99, CUDA version 12.0, compute capability 8.6, 10009MB, 10009MB available, 29768 GFLOPS peak)
Fri 02 Jun 2023 06:01:12 AM EDT | | CUDA: NVIDIA GPU 1: NVIDIA TITAN X (Pascal) (driver version 525.99, CUDA version 12.0, compute capability 6.1, 12187MB, 12187MB available, 10974 GFLOPS peak)
Fri 02 Jun 2023 06:01:12 AM EDT | | OpenCL: NVIDIA GPU 0: NVIDIA GeForce RTX 3080 (driver version 525.105.17, device version OpenCL 3.0 CUDA, 10009MB, 10009MB available, 29768 GFLOPS peak)
Fri 02 Jun 2023 06:01:12 AM EDT | | OpenCL: NVIDIA GPU 1: NVIDIA TITAN X (Pascal) (driver version 525.105.17, device version OpenCL 3.0 CUDA, 12187MB, 12187MB available, 10974 GFLOPS peak)
Fri 02 Jun 2023 06:01:12 AM EDT | | libc: version 2.35
Fri 02 Jun 2023 06:01:12 AM EDT | | Host name: pop-os
Fri 02 Jun 2023 06:01:12 AM EDT | | Processor: 20 GenuineIntel Intel(R) Core(TM) i9-10850K CPU @ 3.60GHz [Family 6 Model 165 Stepping 5]
Fri 02 Jun 2023 06:01:12 AM EDT | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp pku ospke md_clear flush_l1d arch_capabilities
Fri 02 Jun 2023 06:01:12 AM EDT | | OS: Linux GNOME: GNOME 44 (Flatpak runtime) [6.2.6-76060206-generic|libc 2.35]
Fri 02 Jun 2023 06:01:12 AM EDT | | Memory: 31.20 GB physical, 20.00 GB virtual
Fri 02 Jun 2023 06:01:12 AM EDT | | Disk: 219.31 GB total, 182.90 GB free
Fri 02 Jun 2023 06:01:12 AM EDT | | Local time is UTC -4 hours
Fri 02 Jun 2023 06:01:12 AM EDT | | Config: use all coprocessors
I have installations of BOINC on Linux Pop!_OS (Gnome 44) and Windows 11 Pro on my dual booting computer. The computer has two video cards, an Nvidia RTX 3080 and an Nvidia Titan X Pascal. I have a failure rate of 8.7% (169 of 1935) on GPU tasks when I run Linux. The failure rate of GPU tasks for Windows 11 Pro is 0.46% (1 of 217).
If you are not already go back to a single task at a time and see if that helps, also be sure to give Einstein a full cpu core just for the gpu and see if that's a problem as well.
I now have a small sample of valid results and no invalid results on Pop!_OS. I kept the same settings except "Account | Preferences | Project" setting which I changed to YES.
Run Linux app versions built with LIBC 2.15: YES NO
This ensures compatibility with new Linux systems that have virtual syscalls disabled, but breaks compatibility with older systems with (G)LIBC prior to 2.15
I now have a small sample of valid results and no invalid results on Pop!_OS. I kept the same settings except "Account | Preferences | Project" setting which I changed to YES.
Run Linux app versions built with LIBC 2.15: YES NO
This ensures compatibility with new Linux systems that have virtual syscalls disabled, but breaks compatibility with older systems with (G)LIBC prior to 2.15
I would not have thought of that, good for you and I'm glad it worked!!
Keith Myers wrote: mikey
)
That seems a bit involved for that to actually happen then, oh well.
I have installations of BOINC
)
I have installations of BOINC on Linux Pop!_OS (Gnome 44) and Windows 11 Pro on my dual booting computer. The computer has two video cards, an Nvidia RTX 3080 and an Nvidia Titan X Pascal. I have a failure rate of 8.7% (169 of 1935) on GPU tasks when I run Linux. The failure rate of GPU tasks for Windows 11 Pro is 0.46% (1 of 217).
These results were obtained while running the same CPU and GPU speeds, not overclocked, on both installations.
Linux:
Vester wrote:I have
)
If you are not already go back to a single task at a time and see if that helps, also be sure to give Einstein a full cpu core just for the gpu and see if that's a problem as well.
I now have a small sample of
)
Vester wrote: I now have a
)
I would not have thought of that, good for you and I'm glad it worked!!