Issue with second GPU

Author	Message
OffDutyTaoist Send message Joined: 30 Jun 24 Posts: 3 Credit: 21,239,350 RAC: 148,334	Message 10144 - Posted: 1 Oct 2024, 6:28:50 UTC Last modified: 1 Oct 2024, 6:30:35 UTC
I recently started running a second GPU (Primary GTX 1060 Secondary GTX 960). SRBase will run fine with a second project task (ie Einstein, Amicable, etc) but when another SRBase task starts the task on the 960 fails out with a commutation error. Do I need to do something in the XML files? Included is one of the errors: <core_client_version>8.0.2</core_client_version> <![CDATA[ <message> The operating system cannot run (null). (0xc3) - exit code 195 (0xc3)</message> <stderr_txt> 2024-10-01 01:13:05 (15324): wrapper (7.24.26018): starting 2024-10-01 01:13:05 (15324): wrapper: running mfaktc-win-64.exe (-d 1) 2024-10-01 01:13:05 (15324): wrapper: created child process 10536 mfaktc v0.21 (64bit built) Compiletime options THREADS_PER_BLOCK 256 SIEVE_SIZE_LIMIT 32kiB SIEVE_SIZE 193154bits SIEVE_SPLIT 250 MORE_CLASSES enabled Runtime options SievePrimes 25000 SievePrimesAdjust 1 SievePrimesMin 5000 SievePrimesMax 100000 NumStreams 3 CPUStreams 3 GridSize 3 GPU Sieving enabled GPUSievePrimes 82486 GPUSieveSize 2047Mi bits GPUSieveProcessSize 32Ki bits Checkpoints enabled CheckpointDelay 60s WorkFileAddDelay disabled Stages enabled StopAfterFactor class PrintMode full V5UserID (none) ComputerID (none) AllowSleep no TimeStampInResults no CUDA version info binary compiled for CUDA 12.0 CUDA runtime version 12.0 CUDA driver version 12.60 CUDA device info name NVIDIA GeForce GTX 960 compute capability 5.2 max threads per block 1024 max shared memory per MP 98304 byte number of multiprocessors 8 CUDA cores per MP 128 CUDA cores - total 1024 clock rate (CUDA cores) 1367MHz memory clock rate: 3505MHz memory bus width: 128 bit Automatic parameters threads per grid 1048576 GPUSievePrimes (adjusted) 82486 GPUsieve minimum exponent 1055144 running a simple selftest... ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device 2024-10-01 01:13:06 (15324): mfaktc-win-64.exe exited; CPU time 0.000000 2024-10-01 01:13:06 (15324): app exit status: 0x1 2024-10-01 01:13:06 (15324): called boinc_finish(195) </stderr_txt> ]]>
ID: 10144 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,659,667 RAC: 40,856	Message 10146 - Posted: 1 Oct 2024, 7:53:29 UTC - in response to Message 10144.
ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device Looks like the GTX960 is too old and was not compiled in mfaktc with older cc.
ID: 10146 · Rating: 0 · rate: / Reply Quote

zlodeck Send message Joined: 3 Mar 15 Posts: 4 Credit: 138,500,843 RAC: 115,194	Message 10148 - Posted: 1 Oct 2024, 10:05:55 UTC - in response to Message 10146. Last modified: 1 Oct 2024, 10:06:17 UTC
ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device Looks like the GTX960 is too old and was not compiled in mfaktc with older cc. No, it looks like CC 5.2 is enough to run mfaktc with "barrett76_mul32_gs" kernel. Probably large GPUSieveSize (2047Mi bits vs 64Mi bits below) may be an issue for GTX960 with 2Gb memory onboard. 2024-10-01 12:16:10 (9532): wrapper (7.24.26018): starting 2024-10-01 12:16:10 (9532): wrapper: running mfaktc-win-64.exe (-d 0) 2024-10-01 12:16:10 (9532): wrapper: created child process 8352 mfaktc v0.21 (64bit built) Compiletime options THREADS_PER_BLOCK 256 SIEVE_SIZE_LIMIT 32kiB SIEVE_SIZE 193154bits SIEVE_SPLIT 250 MORE_CLASSES enabled Runtime options SievePrimes 25000 SievePrimesAdjust 1 SievePrimesMin 5000 SievePrimesMax 100000 NumStreams 3 CPUStreams 3 GridSize 3 GPU Sieving enabled GPUSievePrimes 82486 GPUSieveSize 64Mi bits GPUSieveProcessSize 16Ki bits Checkpoints enabled CheckpointDelay 60s WorkFileAddDelay disabled Stages enabled StopAfterFactor class PrintMode full V5UserID (none) ComputerID (none) AllowSleep no TimeStampInResults no CUDA version info binary compiled for CUDA 10.0 CUDA runtime version 10.0 CUDA driver version 11.30 CUDA device info name NVIDIA GeForce GTX 980 compute capability 5.2 max threads per block 1024 max shared memory per MP 98304 byte number of multiprocessors 16 CUDA cores per MP 128 CUDA cores - total 2048 clock rate (CUDA cores) 1329MHz memory clock rate: 3505MHz memory bus width: 256 bit Automatic parameters threads per grid 1048576 GPUSievePrimes (adjusted) 82486 GPUsieve minimum exponent 1055144 running a simple selftest... Selftest statistics number of tests 107 successfull tests 107 selftest PASSED! got assignment: exp=178691207 bit_min=75 bit_max=76 (42.82 GHz-days) Starting trial factoring M178691207 from 2^75 to 2^76 (42.82 GHz-days) k_min = 105710103187920 k_max = 211420206384062 Using GPU kernel "barrett76_mul32_gs" Date Time \| class Pct \| time ETA \| GHz-d/day Sieve Wait Oct 01 12:16 \| 0 0.1% \| 6.301 1h40m \| 611.66 82485 n.a.% Oct 01 12:16 \| 4 0.2% \| 6.298 1h40m \| 611.95 82485 n.a.% Oct 01 12:16 \| 9 0.3% \| 6.304 1h40m \| 611.37 82485 n.a.%
ID: 10148 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,659,667 RAC: 40,856	Message 10150 - Posted: 1 Oct 2024, 13:33:55 UTC - in response to Message 10149.
You can only exclude the GTX960 and use it for another project.
ID: 10150 · Rating: 0 · rate: / Reply Quote

OffDutyTaoist Send message Joined: 30 Jun 24 Posts: 3 Credit: 21,239,350 RAC: 148,334	Message 10151 - Posted: 1 Oct 2024, 17:30:16 UTC - in response to Message 10150.
So my CC file should look something like this: <cc_config> <options> <use_all_gpus>1</use_all_gpus> <exclude_gpu> <url>https://srbase.my-firewall.org/sr5/</url> [<device_num>1</device_num>] [<type>NVIDIA_gpu</type>] [<app>SRBase</app>] </exclude_gpu> </options> </cc_config> Correct? My 960 being the second card.
ID: 10151 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,659,667 RAC: 40,856	Message 10152 - Posted: 1 Oct 2024, 17:40:58 UTC - in response to Message 10151. Last modified: 1 Oct 2024, 17:45:09 UTC
Correct? My 960 being the second card. <cc_config> <options> <use_all_gpus>1</use_all_gpus> <exclude_gpu> <url>https://srbase.my-firewall.org/sr5/</url> [<device_num>1</device_num>] [<type>NVIDIA</type>] [<app>TF</app>] </exclude_gpu> </options> </cc_config>
ID: 10152 · Rating: 0 · rate: / Reply Quote

OffDutyTaoist Send message Joined: 30 Jun 24 Posts: 3 Credit: 21,239,350 RAC: 148,334	Message 10159 - Posted: 3 Oct 2024, 3:05:01 UTC
If anyone else has this issue in the future, and needs a solution, I ended up grouping my projects in the cc_config like this: <cc_config> <options> <use_all_gpus>1</use_all_gpus> <exclude_gpu> <url>https://srbase.my-firewall.org/sr5/</url> <device_num>1</device_num> <type>NVIDIA</type> </exclude_gpu> <exclude_gpu> <url>https://asteroidsathome.net/boinc/</url> <device_num>1</device_num> <type>NVIDIA</type> </exclude_gpu> <exclude_gpu> <url>https://www.gpugrid.net/</url> <device_num>1</device_num> <type>NVIDIA</type> </exclude_gpu> <exclude_gpu> <url>https://einstein.phys.uwm.edu/</url> <device_num>0</device_num> <type>NVIDIA</type> </exclude_gpu> <exclude_gpu> <url>https://sech.me/boinc/Amicable/</url> <device_num>0</device_num> <type>NVIDIA</type> </exclude_gpu> <exclude_gpu> <url>https://numberfields.asu.edu/NumberFields/</url> <device_num>0</device_num> <type>NVIDIA</type> </exclude_gpu> </options> </cc_config> The projects I'm working on that get me the most mag with GridCoin I assigned to my 1090 (excluded them from the 960, device 1), and assigned the secondary projects to my 960 (excluded them from the 1060, device 0). It seems to be working out pretty well so far.
ID: 10159 · Rating: 0 · rate: / Reply Quote

Author

Message

OffDutyTaoist
Send message
Joined: 30 Jun 24
Posts: 3
Credit: 21,239,350
RAC: 148,334
20M in project total credit

10k in Sierpinski / Riesel Base - average credit

I recently started running a second GPU (Primary GTX 1060 Secondary GTX 960). SRBase will run fine with a second project task (ie Einstein, Amicable, etc) but when another SRBase task starts the task on the 960 fails out with a commutation error. Do I need to do something in the XML files? Included is one of the errors:

<core_client_version>8.0.2</core_client_version>
<![CDATA[
<message>
The operating system cannot run (null).
(0xc3) - exit code 195 (0xc3)</message>
<stderr_txt>
2024-10-01 01:13:05 (15324): wrapper (7.24.26018): starting
2024-10-01 01:13:05 (15324): wrapper: running mfaktc-win-64.exe (-d 1)
2024-10-01 01:13:05 (15324): wrapper: created child process 10536
mfaktc v0.21 (64bit built)

Compiletime options
THREADS_PER_BLOCK 256
SIEVE_SIZE_LIMIT 32kiB
SIEVE_SIZE 193154bits
SIEVE_SPLIT 250
MORE_CLASSES enabled

Runtime options
SievePrimes 25000
SievePrimesAdjust 1
SievePrimesMin 5000
SievePrimesMax 100000
NumStreams 3
CPUStreams 3
GridSize 3
GPU Sieving enabled
GPUSievePrimes 82486
GPUSieveSize 2047Mi bits
GPUSieveProcessSize 32Ki bits
Checkpoints enabled
CheckpointDelay 60s
WorkFileAddDelay disabled
Stages enabled
StopAfterFactor class
PrintMode full
V5UserID (none)
ComputerID (none)
AllowSleep no
TimeStampInResults no

CUDA version info
binary compiled for CUDA 12.0
CUDA runtime version 12.0
CUDA driver version 12.60

CUDA device info
name NVIDIA GeForce GTX 960
compute capability 5.2
max threads per block 1024
max shared memory per MP 98304 byte
number of multiprocessors 8
CUDA cores per MP 128
CUDA cores - total 1024
clock rate (CUDA cores) 1367MHz
memory clock rate: 3505MHz
memory bus width: 128 bit

Automatic parameters
threads per grid 1048576
GPUSievePrimes (adjusted) 82486
GPUsieve minimum exponent 1055144

running a simple selftest...
ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device
2024-10-01 01:13:06 (15324): mfaktc-win-64.exe exited; CPU time 0.000000
2024-10-01 01:13:06 (15324): app exit status: 0x1
2024-10-01 01:13:06 (15324): called boinc_finish(195)

</stderr_txt>
]]>

ID: 10144 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7479
Credit: 43,659,667
RAC: 40,856
2M in Sierpinski / Riesel Base credit

5M in Sierpinski / Riesel Base - short credit

1M in Sierpinski / Riesel Base - long credit

500k in Sierpinski / Riesel Base - average credit

ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device

Looks like the GTX960 is too old and was not compiled in mfaktc with older cc.

ID: 10146 · Rating: 0 · rate:

Reply Quote

zlodeck
Send message
Joined: 3 Mar 15
Posts: 4
Credit: 138,500,843
RAC: 115,194
5M in Sierpinski / Riesel Base credit

2M in Sierpinski / Riesel Base - short credit

20M in Sierpinski / Riesel Base - long credit

20M in Sierpinski / Riesel Base - average credit

ERROR: cudaGetLastError() returned 209: no kernel image is available for execution on the device

Looks like the GTX960 is too old and was not compiled in mfaktc with older cc.

No, it looks like CC 5.2 is enough to run mfaktc with "barrett76_mul32_gs" kernel. Probably large GPUSieveSize (2047Mi bits vs 64Mi bits below) may be an issue for GTX960 with 2Gb memory onboard.

2024-10-01 12:16:10 (9532): wrapper (7.24.26018): starting
2024-10-01 12:16:10 (9532): wrapper: running mfaktc-win-64.exe (-d 0)
2024-10-01 12:16:10 (9532): wrapper: created child process 8352
mfaktc v0.21 (64bit built)

Compiletime options
THREADS_PER_BLOCK 256
SIEVE_SIZE_LIMIT 32kiB
SIEVE_SIZE 193154bits
SIEVE_SPLIT 250
MORE_CLASSES enabled

Runtime options
SievePrimes 25000
SievePrimesAdjust 1
SievePrimesMin 5000
SievePrimesMax 100000
NumStreams 3
CPUStreams 3
GridSize 3
GPU Sieving enabled
GPUSievePrimes 82486
GPUSieveSize 64Mi bits
GPUSieveProcessSize 16Ki bits
Checkpoints enabled
CheckpointDelay 60s
WorkFileAddDelay disabled
Stages enabled
StopAfterFactor class
PrintMode full
V5UserID (none)
ComputerID (none)
AllowSleep no
TimeStampInResults no

CUDA version info
binary compiled for CUDA 10.0
CUDA runtime version 10.0
CUDA driver version 11.30

CUDA device info
name NVIDIA GeForce GTX 980
compute capability 5.2
max threads per block 1024
max shared memory per MP 98304 byte
number of multiprocessors 16
CUDA cores per MP 128
CUDA cores - total 2048
clock rate (CUDA cores) 1329MHz
memory clock rate: 3505MHz
memory bus width: 256 bit

Automatic parameters
threads per grid 1048576
GPUSievePrimes (adjusted) 82486
GPUsieve minimum exponent 1055144

running a simple selftest...
Selftest statistics
number of tests 107
successfull tests 107

selftest PASSED!

got assignment: exp=178691207 bit_min=75 bit_max=76 (42.82 GHz-days)
Starting trial factoring M178691207 from 2^75 to 2^76 (42.82 GHz-days)
k_min = 105710103187920
k_max = 211420206384062
Using GPU kernel "barrett76_mul32_gs"
Date Time | class Pct | time ETA | GHz-d/day Sieve Wait
Oct 01 12:16 | 0 0.1% | 6.301 1h40m | 611.66 82485 n.a.%
Oct 01 12:16 | 4 0.2% | 6.298 1h40m | 611.95 82485 n.a.%
Oct 01 12:16 | 9 0.3% | 6.304 1h40m | 611.37 82485 n.a.%

ID: 10148 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7479
Credit: 43,659,667
RAC: 40,856
2M in Sierpinski / Riesel Base credit

You can only exclude the GTX960 and use it for another project.

ID: 10150 · Rating: 0 · rate:

Reply Quote

OffDutyTaoist
Send message
Joined: 30 Jun 24
Posts: 3
Credit: 21,239,350
RAC: 148,334
20M in project total credit

So my CC file should look something like this:

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
<exclude_gpu>
<url>https://srbase.my-firewall.org/sr5/</url>
[<device_num>1</device_num>]
[<type>NVIDIA_gpu</type>]
[<app>SRBase</app>]
</exclude_gpu>
</options>
</cc_config>

Correct? My 960 being the second card.

ID: 10151 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7479
Credit: 43,659,667
RAC: 40,856
2M in Sierpinski / Riesel Base credit

Correct? My 960 being the second card.

<cc_config> <options> <use_all_gpus>1</use_all_gpus> <exclude_gpu> <url>https://srbase.my-firewall.org/sr5/</url> [<device_num>1</device_num>] [<type>NVIDIA</type>] [<app>TF</app>] </exclude_gpu> </options> </cc_config>

ID: 10152 · Rating: 0 · rate:

Reply Quote

OffDutyTaoist
Send message
Joined: 30 Jun 24
Posts: 3
Credit: 21,239,350
RAC: 148,334
20M in project total credit

If anyone else has this issue in the future, and needs a solution, I ended up grouping my projects in the cc_config like this:

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
<exclude_gpu>
<url>https://srbase.my-firewall.org/sr5/</url>
<device_num>1</device_num>
<type>NVIDIA</type>
</exclude_gpu>
<exclude_gpu>
<url>https://asteroidsathome.net/boinc/</url>
<device_num>1</device_num>
<type>NVIDIA</type>
</exclude_gpu>
<exclude_gpu>
<url>https://www.gpugrid.net/</url>
<device_num>1</device_num>
<type>NVIDIA</type>
</exclude_gpu>
<exclude_gpu>
<url>https://einstein.phys.uwm.edu/</url>
<device_num>0</device_num>
<type>NVIDIA</type>
</exclude_gpu>
<exclude_gpu>
<url>https://sech.me/boinc/Amicable/</url>
<device_num>0</device_num>
<type>NVIDIA</type>
</exclude_gpu>
<exclude_gpu>
<url>https://numberfields.asu.edu/NumberFields/</url>
<device_num>0</device_num>
<type>NVIDIA</type>
</exclude_gpu>
</options>
</cc_config>

The projects I'm working on that get me the most mag with GridCoin I assigned to my 1090 (excluded them from the 960, device 1), and assigned the secondary projects to my 960 (excluded them from the 1060, device 0). It seems to be working out pretty well so far.

ID: 10159 · Rating: 0 · rate:

Reply Quote