opencl no longer works on gts 450 (fedora)

Moderators: Site Moderators, FAHC Science Team

Post Reply
nikitastepanov
Posts: 9
Joined: Fri Oct 09, 2020 5:37 am

opencl no longer works on gts 450 (fedora)

Post by nikitastepanov »

even though opencl worked in 2020

Code: Select all

17:04:33:WU01:FS01:0x22:*********************** Log Started 2021-11-26T17:04:32Z ***********************
17:04:33:WU01:FS01:0x22:*************************** Core22 Folding@home Core ***************************
17:04:33:WU01:FS01:0x22:       Core: Core22
17:04:33:WU01:FS01:0x22:       Type: 0x22
17:04:33:WU01:FS01:0x22:    Version: 0.0.18
17:04:33:WU01:FS01:0x22:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
17:04:33:WU01:FS01:0x22:  Copyright: 2020 foldingathome.org
17:04:33:WU01:FS01:0x22:   Homepage: https://foldingathome.org/
17:04:33:WU01:FS01:0x22:       Date: Sep 28 2021
17:04:33:WU01:FS01:0x22:       Time: 05:43:02
17:04:33:WU01:FS01:0x22:   Revision: cfe3d7d990e8f456e371f8ce63b5fcc6daab2103
17:04:33:WU01:FS01:0x22:     Branch: HEAD
17:04:33:WU01:FS01:0x22:   Compiler: GNU 9.4.0
17:04:33:WU01:FS01:0x22:    Options: -std=c++11 -fsigned-char -ffunction-sections -fdata-sections -O3
17:04:33:WU01:FS01:0x22:             -funroll-loops -fno-pie -DOPENMM_VERSION="\"7.6.0\""
17:04:33:WU01:FS01:0x22:   Platform: linux 5.8.0-1041-azure
17:04:33:WU01:FS01:0x22:       Bits: 64
17:04:33:WU01:FS01:0x22:       Mode: Release
17:04:33:WU01:FS01:0x22:Maintainers: John Chodera <john.chodera@choderalab.org> and Peter Eastman
17:04:33:WU01:FS01:0x22:             <peastman@stanford.edu>
17:04:33:WU01:FS01:0x22:       Args: -dir 01 -suffix 01 -version 706 -lifeline 5939 -checkpoint 15
17:04:33:WU01:FS01:0x22:             -opencl-platform 0 -opencl-device 0 -cuda-device 0 -gpu-vendor
17:04:33:WU01:FS01:0x22:             nvidia -gpu 0 -gpu-usage 100
17:04:33:WU01:FS01:0x22:************************************ libFAH ************************************
17:04:33:WU01:FS01:0x22:       Date: Sep 28 2021
17:04:33:WU01:FS01:0x22:       Time: 05:42:27
17:04:33:WU01:FS01:0x22:   Revision: 44301ed97b996b63fe736bb8073f22209cb2b603
17:04:33:WU01:FS01:0x22:     Branch: HEAD
17:04:33:WU01:FS01:0x22:   Compiler: GNU 9.4.0
17:04:33:WU01:FS01:0x22:    Options: -std=c++11 -fsigned-char -ffunction-sections -fdata-sections -O3
17:04:33:WU01:FS01:0x22:             -funroll-loops -fno-pie
17:04:33:WU01:FS01:0x22:   Platform: linux 5.8.0-1041-azure
17:04:33:WU01:FS01:0x22:       Bits: 64
17:04:33:WU01:FS01:0x22:       Mode: Release
17:04:33:WU01:FS01:0x22:************************************ CBang *************************************
17:04:33:WU01:FS01:0x22:       Date: Sep 28 2021
17:04:33:WU01:FS01:0x22:       Time: 05:42:00
17:04:33:WU01:FS01:0x22:   Revision: 33fcfc2b3ed2195a423606a264718e31e6b3903f
17:04:33:WU01:FS01:0x22:     Branch: HEAD
17:04:33:WU01:FS01:0x22:   Compiler: GNU 9.4.0
17:04:33:WU01:FS01:0x22:    Options: -std=c++11 -fsigned-char -ffunction-sections -fdata-sections -O3
17:04:33:WU01:FS01:0x22:             -funroll-loops -fno-pie -fPIC
17:04:33:WU01:FS01:0x22:   Platform: linux 5.8.0-1041-azure
17:04:33:WU01:FS01:0x22:       Bits: 64
17:04:33:WU01:FS01:0x22:       Mode: Release
17:04:33:WU01:FS01:0x22:************************************ System ************************************
17:04:33:WU01:FS01:0x22:        CPU: Intel(R) Core(TM)2 Duo CPU E7500 @ 2.93GHz
17:04:33:WU01:FS01:0x22:     CPU ID: GenuineIntel Family 6 Model 23 Stepping 10
17:04:33:WU01:FS01:0x22:       CPUs: 2
17:04:33:WU01:FS01:0x22:     Memory: 3.82GiB
17:04:33:WU01:FS01:0x22:Free Memory: 1.91GiB
17:04:33:WU01:FS01:0x22:    Threads: POSIX_THREADS
17:04:33:WU01:FS01:0x22: OS Version: 5.14
17:04:33:WU01:FS01:0x22:Has Battery: false
17:04:33:WU01:FS01:0x22: On Battery: false
17:04:33:WU01:FS01:0x22: UTC Offset: 6
17:04:33:WU01:FS01:0x22:        PID: 5943
17:04:33:WU01:FS01:0x22:        CWD: /var/lib/fahclient/work
17:04:33:WU01:FS01:0x22:************************************ OpenMM ************************************
17:04:33:WU01:FS01:0x22:    Version: 7.6.0
17:04:33:WU01:FS01:0x22:********************************************************************************
17:04:33:WU01:FS01:0x22:Project: 17806 (Run 49, Clone 13, Gen 354)
17:04:33:WU01:FS01:0x22:Unit: 0x00000000000000000000000000000000
17:04:33:WU01:FS01:0x22:Reading tar file core.xml
17:04:33:WU01:FS01:0x22:Reading tar file integrator.xml.bz2
17:04:33:WU01:FS01:0x22:Reading tar file state.xml.bz2
17:04:33:WU01:FS01:0x22:Reading tar file system.xml.bz2
17:04:33:WU01:FS01:0x22:Digital signatures verified
17:04:33:WU01:FS01:0x22:Folding@home GPU Core22 Folding@home Core
17:04:33:WU01:FS01:0x22:Version 0.0.18
17:04:33:WU01:FS01:0x22:  Checkpoint write interval: 250000 steps (5%) [20 total]
17:04:33:WU01:FS01:0x22:  JSON viewer frame write interval: 50000 steps (1%) [100 total]
17:04:33:WU01:FS01:0x22:  XTC frame write interval: 25000 steps (0.5%) [200 total]
17:04:33:WU01:FS01:0x22:  Global context and integrator variables write interval: disabled
17:04:33:WU01:FS01:0x22:There are 4 platforms available.
17:04:33:WU01:FS01:0x22:Platform 0: Reference
17:04:33:WU01:FS01:0x22:Platform 1: CPU
17:04:33:WU01:FS01:0x22:Platform 2: OpenCL
17:04:33:WU01:FS01:0x22:  opencl-device 0 specified
17:04:33:WU01:FS01:0x22:Platform 3: CUDA
17:04:33:WU01:FS01:0x22:  cuda-device 0 specified
17:04:36:WU01:FS01:0x22:Attempting to create CUDA context:
17:04:36:WU01:FS01:0x22:  Configuring platform CUDA
17:04:36:WU01:FS01:0x22:Failed to create CUDA context:
17:04:36:WU01:FS01:0x22:Error compiling program: nvrtc: error: invalid value for --gpu-architecture (-arch)
17:04:36:WU01:FS01:0x22:Attempting to create OpenCL context:
17:04:36:WU01:FS01:0x22:  Configuring platform OpenCL
17:04:37:WU01:FS01:0x22:Failed to create OpenCL context:
17:04:37:WU01:FS01:0x22:Error compiling kernel: <kernel>:635:16: error: call to 'native_rsqrt' is ambiguous
17:04:37:WU01:FS01:0x22:        eAB *= RSQRT(eAB.x*eAB.x + eAB.y*eAB.y + eAB.z*eAB.z);
17:04:37:WU01:FS01:0x22:               ^~~~~
17:04:37:WU01:FS01:0x22:<kernel>:18:15: note: expanded from macro 'RSQRT'
17:04:37:WU01:FS01:0x22:#define RSQRT native_rsqrt
17:04:37:WU01:FS01:0x22:              ^~~~~~~~~~~~
17:04:37:WU01:FS01:0x22:cl_kernel.h:994:24: note: candidate function
17:04:37:WU01:FS01:0x22:float __OVERLOADABLE__ native_rsqrt(float);
17:04:37:WU01:FS01:0x22:                       ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:995:25: note: candidate function
17:04:37:WU01:FS01:0x22:float2 __OVERLOADABLE__ native_rsqrt(float2); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:997:25: note: candidate function
17:04:37:WU01:FS01:0x22:float3 __OVERLOADABLE__ native_rsqrt(float3); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:999:25: note: candidate function
17:04:37:WU01:FS01:0x22:float4 __OVERLOADABLE__ native_rsqrt(float4); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1000:25: note: candidate function
17:04:37:WU01:FS01:0x22:float8 __OVERLOADABLE__ native_rsqrt(float8); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1001:26: note: candidate function
17:04:37:WU01:FS01:0x22:float16 __OVERLOADABLE__ native_rsqrt(float16); 
17:04:37:WU01:FS01:0x22:                         ^
17:04:37:WU01:FS01:0x22:<kernel>:636:16: error: call to 'native_rsqrt' is ambiguous
17:04:37:WU01:FS01:0x22:        eBC *= RSQRT(eBC.x*eBC.x + eBC.y*eBC.y + eBC.z*eBC.z);
17:04:37:WU01:FS01:0x22:               ^~~~~
17:04:37:WU01:FS01:0x22:<kernel>:18:15: note: expanded from macro 'RSQRT'
17:04:37:WU01:FS01:0x22:#define RSQRT native_rsqrt
17:04:37:WU01:FS01:0x22:              ^~~~~~~~~~~~
17:04:37:WU01:FS01:0x22:cl_kernel.h:994:24: note: candidate function
17:04:37:WU01:FS01:0x22:float __OVERLOADABLE__ native_rsqrt(float);
17:04:37:WU01:FS01:0x22:                       ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:995:25: note: candidate function
17:04:37:WU01:FS01:0x22:float2 __OVERLOADABLE__ native_rsqrt(float2); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:997:25: note: candidate function
17:04:37:WU01:FS01:0x22:float3 __OVERLOADABLE__ native_rsqrt(float3); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:999:25: note: candidate function
17:04:37:WU01:FS01:0x22:float4 __OVERLOADABLE__ native_rsqrt(float4); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1000:25: note: candidate function
17:04:37:WU01:FS01:0x22:float8 __OVERLOADABLE__ native_rsqrt(float8); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1001:26: note: candidate function
17:04:37:WU01:FS01:0x22:float16 __OVERLOADABLE__ native_rsqrt(float16); 
17:04:37:WU01:FS01:0x22:                         ^
17:04:37:WU01:FS01:0x22:<kernel>:637:16: error: call to 'native_rsqrt' is ambiguous
17:04:37:WU01:FS01:0x22:        eCA *= RSQRT(eCA.x*eCA.x + eCA.y*eCA.y + eCA.z*eCA.z);
17:04:37:WU01:FS01:0x22:               ^~~~~
17:04:37:WU01:FS01:0x22:<kernel>:18:15: note: expanded from macro 'RSQRT'
17:04:37:WU01:FS01:0x22:#define RSQRT native_rsqrt
17:04:37:WU01:FS01:0x22:              ^~~~~~~~~~~~
17:04:37:WU01:FS01:0x22:cl_kernel.h:994:24: note: candidate function
17:04:37:WU01:FS01:0x22:float __OVERLOADABLE__ native_rsqrt(float);
17:04:37:WU01:FS01:0x22:                       ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:995:25: note: candidate function
17:04:37:WU01:FS01:0x22:float2 __OVERLOADABLE__ native_rsqrt(float2); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:997:25: note: candidate function
17:04:37:WU01:FS01:0x22:float3 __OVERLOADABLE__ native_rsqrt(float3); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:999:25: note: candidate function
17:04:37:WU01:FS01:0x22:float4 __OVERLOADABLE__ native_rsqrt(float4); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1000:25: note: candidate function
17:04:37:WU01:FS01:0x22:float8 __OVERLOADABLE__ native_rsqrt(float8); 
17:04:37:WU01:FS01:0x22:                        ^
17:04:37:WU01:FS01:0x22:cl_kernel.h:1001:26: note: candidate function
17:04:37:WU01:FS01:0x22:float16 __OVERLOADABLE__ native_rsqrt(float16); 
17:04:37:WU01:FS01:0x22:                         ^
17:04:37:WU01:FS01:0x22:ERROR:125: Failed to create a GPU-enabled OpenMM Context.
17:04:37:WU01:FS01:0x22:Saving result file ../logfile_01.txt
17:04:37:WU01:FS01:0x22:Saving result file science.log
17:04:37:WU01:FS01:0x22:Folding@home Core Shutdown: BAD_WORK_UNIT
[93m17:04:37:WARNING:WU01:FS01:FahCore returned: BAD_WORK_UNIT (114 = 0x72)[0m
17:04:37:WU01:FS01:Sending unit results: id:01 state:SEND error:FAULTY project:17806 run:49 clone:13 gen:354 core:0x22 unit:0x0000000d000001620000458e00000031
17:04:37:WU01:FS01:Uploading 18.00KiB to 207.53.233.146
17:04:37:WU01:FS01:Connecting to 207.53.233.146:8080
17:04:37:WU02:FS01:Connecting to assign1.foldingathome.org:80
17:04:38:WU01:FS01:Upload complete
17:04:38:WU01:FS01:Server responded WORK_ACK (400)
17:04:38:WU01:FS01:Cleaning up
17:04:38:WU02:FS01:Assigned to work server 207.53.233.146
17:04:38:WU02:FS01:Requesting new work unit for slot 01: gpu:1:0 GF106 [GeForce GTS 450] from 207.53.233.146
17:04:38:WU02:FS01:Connecting to 207.53.233.146:8080
17:04:39:WU02:FS01:Downloading 5.49MiB
17:04:45:WU02:FS01:Download 45.52%
17:04:51:WU02:FS01:Download 62.59%
17:04:57:WU02:FS01:Download 73.98%
17:05:03:WU02:FS01:Download 92.18%
17:05:04:WU02:FS01:Download complete
17:05:04:WU02:FS01:Received Unit: id:02 state:DOWNLOAD error:NO_ERROR project:17804 run:20 clone:63 gen:644 core:0x22 unit:0x0000003f000002840000458c00000014
17:05:05:WU02:FS01:Starting
17:05:05:WU02:FS01:Running FahCore: /usr/bin/FAHCoreWrapper /var/lib/fahclient/cores/cores.foldingathome.org/lin/64bit/22-0.0.18/Core_22.fah/FahCore_22 -dir 02 -suffix 01 -version 706 -lifeline 5469 -checkpoint 15 -opencl-platform 0 -opencl-device 0 -cuda-device 0 -gpu-vendor nvidia -gpu 0 -gpu-usage 100
17:05:05:WU02:FS01:Started FahCore on PID 5970
17:05:05:WU02:FS01:Core PID:5974
17:05:05:WU02:FS01:FahCore 0x22 started
JimboPalmer
Posts: 2573
Joined: Mon Feb 16, 2009 4:12 am
Location: Greenwood MS USA

Re: opencl no longer works on gts 450 (fedora)

Post by JimboPalmer »

viewtopic.php?f=24&t=37545
a newer Nvidia driver may solve this in the short run.
Tsar of all the Rushers
I tried to remain childlike, all I achieved was childish.
A friend to those who want no friends
Joe_H
Site Admin
Posts: 7856
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: opencl no longer works on gts 450 (fedora)

Post by Joe_H »

In the long run the GTS 450 is going t be unsupported at all. The Fermi chips used ti make them do not support OpenCL 1.2. If you have the variant made with a GF116 chip instead of a GF106 it supports a bit more of the OpenCL commands introduced between 1.1 and 1.2 and will fail on projects less often.

But the current requirements are support for both OpenCL 1.2 and double precision (FP64).
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
JimboPalmer
Posts: 2573
Joined: Mon Feb 16, 2009 4:12 am
Location: Greenwood MS USA

Re: opencl no longer works on gts 450 (fedora)

Post by JimboPalmer »

In early 2020 there were still some Core_21 Work Units, Which were lax about needing OpenCL 1.2, but with the new COVID work, they were all Core_22, which needs OpenCL 1.2. While I have seen no official announcement that Core_21 is retired, it has been almost 2 years since I saw one.
Tsar of all the Rushers
I tried to remain childlike, all I achieved was childish.
A friend to those who want no friends
Joe_H
Site Admin
Posts: 7856
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: opencl no longer works on gts 450 (fedora)

Post by Joe_H »

Early versions of Core_22 would sometimes process on Fermi based GPUs as well. But WUs from some projects that created calls to missing OpenCL functions would fail, not always predictably. WUs might work on one model GPU using one type of Fermi chip and fail on a different one. There are also two generations of Fermi chips, the GF11n chips worked a bit better than the first GF10n.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
toTOW
Site Moderator
Posts: 6296
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:

Re: opencl no longer works on gts 450 (fedora)

Post by toTOW »

It's a Fermi GPU that has been unsupported by nVidia for a long time now ... Kepler GPU support has also been dropped recently but the drivers are still new enough to use them.

I guess the time has come to completely turn off Fermi support on FAH with the latest version of FAHCore 22 ... :(
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
Post Reply