Linux TF WUs on ATI 6800 cards

Author	Message
rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9832 - Posted: 23 Mar 2024, 11:21:26 UTC - in response to Message 9831.
Nope, try to download mfakto-linux64-v9.zip It contains the older mfakto file, replace it with yours. Also overwrite all .cl files. 1) I unzipped this downloaded zip creating a new test folder (mfakto-linux64-v9). 2) Replaced the mfakto-64 in this new folder with mfakto-64 from my 'wrapper=test' folder. 3) Replaced the .cl files in this new folder with wrapper-test's 7 .cl files. 5) Executed this -- ES7C95:~/mfakto/mfakto-linux64-v9$ ./mfakto-x64 6) Got the following but feel like I regressed... mfakto 0.15pre8 (64-bit build) Runtime options INI file mfakto.ini Verbosity 1 SieveOnGPU yes MoreClasses yes GPUSievePrimes 82486 GPUSieveProcessSize 24 Kib GPUSieveSize 96 Mib FlushInterval 8 WorkFile worktodo.txt ResultsFile results.txt Checkpoints enabled CheckpointDelay 60 s Stages enabled StopAfterFactor class PrintMode compact V5UserID none ComputerID none TimeStampInResults no VectorSize 2 GPUType AUTO SmallExp no UseBinfile mfakto_Kernels.elf Compile-time options Select device - GPU not found, fallback to CPU. Error -1 (Device not found): clCreateContextFromType(CPU) ERROR: init_CL(3, 0) failed anyhow mfakto 0.15pre8 is still be used not pre7
ID: 9832 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9833 - Posted: 23 Mar 2024, 14:24:35 UTC - in response to Message 9832.
IF I understand this last test was back to stand-alone w/o the wrapper and it now did not find the GPU. Whereas previously the GPU wasn't found only when we ran inside the wrapper. Is this correct?
ID: 9833 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9834 - Posted: 23 Mar 2024, 14:31:57 UTC - in response to Message 9833. Last modified: 23 Mar 2024, 14:58:39 UTC
IF I understand this last test was back to stand-alone w/o the wrapper and it now did not find the GPU. Whereas previously the GPU wasn't found only when we ran inside the wrapper. Is this correct? To be sure to find the GPU and you have all the files except the wrapper file run mfakto only. Edit: If you still have issues with GPU detection try to change in mfakto.ini GPUType=AUTO to GPUType=GCN
ID: 9834 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9835 - Posted: 23 Mar 2024, 15:07:22 UTC - in response to Message 9832.
I went back and re-ran the original stand-alone test we did. root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5/mfakto-test/mfakto-linux64-v10# ./mfakto-x64 -st > test.log As you can see this one was run in a test folder (mfakto-linux64-v10) under the srbase projects directory of BOINC (BOINC was running) and again it finds the GPU and completes. This is the bottom of test.log: ######### test case 34071/34071 (M112404491[91-92]) ######### Starting trial factoring M112404491 from 2^91 to 2^92 (4461450.54 GHz-days) Date Time \| class Pct \| time ETA \| GHz-d/day Sieve Wait Mar 23 09:40 \| 1848 0.1% \| 0.009 n.a. \| n.a. 82486 0.00% M112404491 has a factor: 3941616367695054034124905537 (91.670846 bits, 2992945.937358 GHz-d) found 1 factor for M112404491 from 2^91 to 2^92 [mfakto 0.15pre8 cl_barrett32_92_gs_2] self-test for M112404491 passed (cl_barrett32_92_gs)! tf(): total time spent: 0.009s Self-test statistics number of tests 34026 successful tests 34026 self-test PASSED! The 'wrapper' test and all since where it doesn't find the GPU was run in test folders under my home directory. Here's a list of what's in srbase.my-firewall.org_sr5/mfakto-test/mfakto-linux64-v10: barrett15.cl barrett.cl Changelog-mfakto.txt common.cl COPYING datatypes.h gpusieve.cl mfakto.ini mfakto_Kernels.cl mfakto_Kernels.elf mfakto-x64 montgomery.cl mul24.cl project_dir README-SpecialVersions.txt README.txt test.log test.log.old tf_debug.h todo.txt I don't know what's inherited from /var/lib/boinc-client or projects/srbase.my-firewall.org_sr5 that allows the GPU to be found when running from here. I will re-try the wrapper test using sudo just to make sure it's not the root privileges that's allowing it to be found. Skip
ID: 9835 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9836 - Posted: 23 Mar 2024, 15:20:54 UTC - in response to Message 9834. Last modified: 23 Mar 2024, 15:23:54 UTC
I re-ran the 'wrapper' test as sudo and it finds the GPU! After a couple minutes it gets some sort of error with worktodo.txt but it was running on the GPU as I could see the load go up in corectrl. I forgot to kill BOINC before running this so it ran while 2 E@H tasks were also processing on the RX6800. Let me know what you want me to do next boss. sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64 [sudo] password for sharon: 2024-03-23 10:16:32 (8297): wrapper (7.24.26018): starting mfakto 0.15pre8 (64-bit build) Runtime options INI file mfakto.ini Verbosity 1 SieveOnGPU yes MoreClasses yes GPUSievePrimes 82486 GPUSieveProcessSize 24 Kib GPUSieveSize 96 Mib FlushInterval 8 WorkFile worktodo.txt ResultsFile results.txt Checkpoints enabled CheckpointDelay 60 s Stages enabled StopAfterFactor class PrintMode compact V5UserID none ComputerID none TimeStampInResults no VectorSize 2 GPUType AUTO SmallExp no UseBinfile mfakto_Kernels.elf Compile-time options Select device - Get device info: OpenCL device info name gfx1030 (Advanced Micro Devices, Inc.) device (driver) version OpenCL 2.0 (3513.0 (HSA1.1,LC)) maximum threads per block 1024 maximum threads per grid 1073741824 number of multiprocessors 30 (1920 compute elements) clock rate 2475 MHz Automatic parameters threads per grid 0 optimizing kernels for RDNA Compiling kernels. GPUSievePrimes (adjusted) 82486 GPUsieve minimum exponent 1055144 Started a simple self-test ... Self-test statistics number of tests 30 successful tests 30 self-test PASSED! got assignment: exp=758604829 bit_min=73 bit_max=74 (2.52 GHz-days) Starting trial factoring M758604829 from 2^73 to 2^74 (2.52 GHz-days) Using GPU kernel "cl_barrett15_74_gs_2" Date Time \| class Pct \| time ETA \| GHz-d/day Sieve Wait Mar 23 10:19 \| 4619 100.0% \| 0.162 0m00s \| 1400.98 82486 0.00% no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2] tf(): total time spent: 2m 33.955s (1415.22 GHz-days / day) ERROR: get_next_assignment(): no valid assignment found in "worktodo.txt"
ID: 9836 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9837 - Posted: 23 Mar 2024, 15:23:51 UTC - in response to Message 9836. Last modified: 23 Mar 2024, 15:24:02 UTC
Looks fine. The test was successfully. You should have a results.txt file in there.
ID: 9837 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9838 - Posted: 23 Mar 2024, 15:25:13 UTC - in response to Message 9837. Last modified: 23 Mar 2024, 15:31:04 UTC
Yes. It has one line in it: no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2] Releasing a suspended TF task in BOINC still gets this: https://srbase.my-firewall.org/sr5/result.php?resultid=146304037
ID: 9838 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9839 - Posted: 23 Mar 2024, 15:27:16 UTC - in response to Message 9838. Last modified: 23 Mar 2024, 15:28:16 UTC
Yes. It has one line in it: no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2] Then you did the test but why you cant run in BOINC? Do you still have -d 0 in job.xml?
ID: 9839 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9840 - Posted: 23 Mar 2024, 15:32:12 UTC - in response to Message 9839. Last modified: 23 Mar 2024, 15:38:00 UTC
<command_line>-d 0</command_line> I'll assume this in the BOINC world comes from the file "job_TF_l64m1_00029.xml" which has <job_desc> <task> <application>./mfakto-x64</application> <command_line>-d $GPU_DEVICE_NUM</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc>
ID: 9840 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9841 - Posted: 23 Mar 2024, 15:32:56 UTC - in response to Message 9840.
-d 0 and that was running?
ID: 9841 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9842 - Posted: 23 Mar 2024, 15:44:08 UTC - in response to Message 9841. Last modified: 23 Mar 2024, 15:55:04 UTC
In terminal I ran: sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64 mfakto/mfakto-wrappertest folder has this in job.xml: <job_desc> <task> <application>./mfakto-x64</application> <command_line>-d 0</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc> sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ echo $GPU_DEVICE_NUM Doesn't display anything visible... I was hoping to see a # Do you know where this gets set?
ID: 9842 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9843 - Posted: 23 Mar 2024, 15:48:38 UTC - in response to Message 9842.
In terminal I ran: sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64 mfakto/mfakto-wrappertest folder has this in job.xml: <job_desc> <task> <application>./mfakto-x64</application> <command_line>-d 0</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc> Where do you have the files? Must be a permission issue somewhere.
ID: 9843 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9844 - Posted: 23 Mar 2024, 16:00:39 UTC - in response to Message 9843.
What files? R we talking wrapper-test or BOINC world?
ID: 9844 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9845 - Posted: 23 Mar 2024, 16:01:36 UTC - in response to Message 9844.
What files? R we talking wrapper-test or BOINC world? wrapper test
ID: 9845 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9846 - Posted: 23 Mar 2024, 16:06:32 UTC - in response to Message 9845.
/home/sharon/mfacto/mfakto-wrappertest/ has these files in it: barrett15.cl barrett.cl boinc_finish_called common.cl COPYING datatypes.h gpusieve.cl job.xml mfakto.ini mfakto_Kernels.cl mfakto_Kernels.elf mfakto-x64 montgomery.cl mul24.cl README.txt results.txt stderr.txt tf_debug.h todo.txt worktodo.txt wrapper_26018_linux-u18_x86-64 wrapper_checkpoint.txt wrapperdir
ID: 9846 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9847 - Posted: 23 Mar 2024, 16:10:03 UTC - in response to Message 9846.
/home/sharon/mfacto/mfakto-wrappertest/ has these files in it: barrett15.cl barrett.cl boinc_finish_called common.cl COPYING datatypes.h gpusieve.cl job.xml mfakto.ini mfakto_Kernels.cl mfakto_Kernels.elf mfakto-x64 montgomery.cl mul24.cl README.txt results.txt stderr.txt tf_debug.h todo.txt worktodo.txt wrapper_26018_linux-u18_x86-64 wrapper_checkpoint.txt wrapperdir ok, then sudo is ok, now we need to know why its not running in BOINC, all looks fine.
ID: 9847 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9848 - Posted: 23 Mar 2024, 16:22:46 UTC - in response to Message 9847.
Trying to show what that variable is set to (that s/b = 0) by going into the boinc folder where srbase runs from .... as root.. gives me nothing back so I assume it's set to null. root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5# echo $GPU_DEVICE_NUM
ID: 9848 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9849 - Posted: 23 Mar 2024, 16:27:29 UTC - in response to Message 9848. Last modified: 23 Mar 2024, 16:30:05 UTC
Trying to show what that variable is set to (that s/b = 0) by going into the boinc folder where srbase runs from .... as root.. gives me nothing back so I assume it's set to null. root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5# echo $GPU_DEVICE_NUM Yes, every card is counting from 0, if you have AMD and nvidia both starting from 0, if you have 2 AMDs then 0, 1
ID: 9849 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7255 Credit: 42,729,227 RAC: 2	Message 9850 - Posted: 23 Mar 2024, 16:32:38 UTC Last modified: 23 Mar 2024, 16:33:41 UTC
In summary of both tests on vega20 and 6000er: - wrapper standalone test was successful without errors Now we need to find out why they failed in BOINC
ID: 9850 · Rating: 0 · rate: / Reply Quote

Skip Da Shu Send message Joined: 27 Mar 23 Posts: 24 Credit: 39,719,772 RAC: 186,770	Message 9851 - Posted: 23 Mar 2024, 17:15:41 UTC - in response to Message 9849. Last modified: 23 Mar 2024, 17:18:03 UTC
Yes, counted from 0... back when I had a 5700G APU/CPU the discrete card was 0 & the iGPU was 1. I excluded the iGPU in cc_config.xml. I'm still curious how $GPU_DEVICE_NUM is set.
ID: 9851 · Rating: 0 · rate: / Reply Quote

Author

Message

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
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

Nope, try to download mfakto-linux64-v9.zip It contains the older mfakto file, replace it with yours. Also overwrite all .cl files.

1) I unzipped this downloaded zip creating a new test folder (mfakto-linux64-v9).
2) Replaced the mfakto-64 in this new folder with mfakto-64 from my 'wrapper=test' folder.
3) Replaced the .cl files in this new folder with wrapper-test's 7 .cl files.
5) Executed this --
ES7C95:~/mfakto/mfakto-linux64-v9$ ./mfakto-x64

6) Got the following but feel like I regressed...

mfakto 0.15pre8 (64-bit build)

Runtime options INI file mfakto.ini Verbosity 1 SieveOnGPU yes MoreClasses yes GPUSievePrimes 82486 GPUSieveProcessSize 24 Kib GPUSieveSize 96 Mib FlushInterval 8 WorkFile worktodo.txt ResultsFile results.txt Checkpoints enabled CheckpointDelay 60 s Stages enabled StopAfterFactor class PrintMode compact V5UserID none ComputerID none TimeStampInResults no VectorSize 2 GPUType AUTO SmallExp no UseBinfile mfakto_Kernels.elf Compile-time options Select device - GPU not found, fallback to CPU. Error -1 (Device not found): clCreateContextFromType(CPU) ERROR: init_CL(3, 0) failed

anyhow mfakto 0.15pre8 is still be used not pre7

ID: 9832 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

10k in Sierpinski / Riesel Base - short credit

2M in Sierpinski / Riesel Base - long credit

IF I understand this last test was back to stand-alone w/o the wrapper and it now did not find the GPU. Whereas previously the GPU wasn't found only when we ran inside the wrapper. Is this correct?

ID: 9833 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

IF I understand this last test was back to stand-alone w/o the wrapper and it now did not find the GPU. Whereas previously the GPU wasn't found only when we ran inside the wrapper. Is this correct?

To be sure to find the GPU and you have all the files except the wrapper file run mfakto only.

Edit:
If you still have issues with GPU detection try to change in mfakto.ini

GPUType=AUTO to GPUType=GCN

ID: 9834 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

I went back and re-ran the original stand-alone test we did.

root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5/mfakto-test/mfakto-linux64-v10# ./mfakto-x64 -st > test.log

As you can see this one was run in a test folder (mfakto-linux64-v10) under the srbase projects directory of BOINC (BOINC was running) and again it finds the GPU and completes. This is the bottom of test.log:

######### test case 34071/34071 (M112404491[91-92]) ######### Starting trial factoring M112404491 from 2^91 to 2^92 (4461450.54 GHz-days) Date Time | class Pct | time ETA | GHz-d/day Sieve Wait Mar 23 09:40 | 1848 0.1% | 0.009 n.a. | n.a. 82486 0.00% M112404491 has a factor: 3941616367695054034124905537 (91.670846 bits, 2992945.937358 GHz-d) found 1 factor for M112404491 from 2^91 to 2^92 [mfakto 0.15pre8 cl_barrett32_92_gs_2] self-test for M112404491 passed (cl_barrett32_92_gs)! tf(): total time spent: 0.009s Self-test statistics number of tests 34026 successful tests 34026 self-test PASSED!

The 'wrapper' test and all since where it doesn't find the GPU was run in test folders under my home directory.

Here's a list of what's in srbase.my-firewall.org_sr5/mfakto-test/mfakto-linux64-v10:

barrett15.cl
barrett.cl
Changelog-mfakto.txt
common.cl
COPYING
datatypes.h
gpusieve.cl
mfakto.ini
mfakto_Kernels.cl
mfakto_Kernels.elf
mfakto-x64
montgomery.cl
mul24.cl
project_dir
README-SpecialVersions.txt
README.txt
test.log
test.log.old
tf_debug.h
todo.txt

I don't know what's inherited from /var/lib/boinc-client or projects/srbase.my-firewall.org_sr5 that allows the GPU to be found when running from here.

I will re-try the wrapper test using sudo just to make sure it's not the root privileges that's allowing it to be found.

Skip

ID: 9835 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

I re-ran the 'wrapper' test as sudo and it finds the GPU! After a couple minutes it gets some sort of error with worktodo.txt but it was running on the GPU as I could see the load go up in corectrl.

I forgot to kill BOINC before running this so it ran while 2 E@H tasks were also processing on the RX6800.

Let me know what you want me to do next boss.

sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64 [sudo] password for sharon: 2024-03-23 10:16:32 (8297): wrapper (7.24.26018): starting mfakto 0.15pre8 (64-bit build) Runtime options INI file mfakto.ini Verbosity 1 SieveOnGPU yes MoreClasses yes GPUSievePrimes 82486 GPUSieveProcessSize 24 Kib GPUSieveSize 96 Mib FlushInterval 8 WorkFile worktodo.txt ResultsFile results.txt Checkpoints enabled CheckpointDelay 60 s Stages enabled StopAfterFactor class PrintMode compact V5UserID none ComputerID none TimeStampInResults no VectorSize 2 GPUType AUTO SmallExp no UseBinfile mfakto_Kernels.elf Compile-time options Select device - Get device info: OpenCL device info name gfx1030 (Advanced Micro Devices, Inc.) device (driver) version OpenCL 2.0 (3513.0 (HSA1.1,LC)) maximum threads per block 1024 maximum threads per grid 1073741824 number of multiprocessors 30 (1920 compute elements) clock rate 2475 MHz Automatic parameters threads per grid 0 optimizing kernels for RDNA Compiling kernels. GPUSievePrimes (adjusted) 82486 GPUsieve minimum exponent 1055144 Started a simple self-test ... Self-test statistics number of tests 30 successful tests 30 self-test PASSED! got assignment: exp=758604829 bit_min=73 bit_max=74 (2.52 GHz-days) Starting trial factoring M758604829 from 2^73 to 2^74 (2.52 GHz-days) Using GPU kernel "cl_barrett15_74_gs_2" Date Time | class Pct | time ETA | GHz-d/day Sieve Wait Mar 23 10:19 | 4619 100.0% | 0.162 0m00s | 1400.98 82486 0.00% no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2] tf(): total time spent: 2m 33.955s (1415.22 GHz-days / day) ERROR: get_next_assignment(): no valid assignment found in "worktodo.txt"

ID: 9836 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

Looks fine. The test was successfully. You should have a results.txt file in there.

ID: 9837 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

Yes.

It has one line in it:

no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2]

Releasing a suspended TF task in BOINC still gets this:

https://srbase.my-firewall.org/sr5/result.php?resultid=146304037

ID: 9838 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

Yes.

It has one line in it:

no factor for M758604829 from 2^73 to 2^74 [mfakto 0.15pre8 cl_barrett15_74_gs_2]

Then you did the test but why you cant run in BOINC? Do you still have -d 0 in job.xml?

ID: 9839 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

<command_line>-d 0</command_line>

I'll assume this in the BOINC world comes from the file "job_TF_l64m1_00029.xml" which has

<job_desc> <task> <application>./mfakto-x64</application> <command_line>-d $GPU_DEVICE_NUM</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc>

ID: 9840 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

-d 0

and that was running?

ID: 9841 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

In terminal I ran:

sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64

mfakto/mfakto-wrappertest folder has this in job.xml:

<job_desc> <task> <application>./mfakto-x64</application> <command_line>-d 0</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc>

sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ echo $GPU_DEVICE_NUM
Doesn't display anything visible... I was hoping to see a #

Do you know where this gets set?

ID: 9842 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

In terminal I ran:

sharon@sharon-ES7C95:~/mfakto/mfakto-wrappertest$ sudo ./wrapper_26018_linux-u18_x86-64

mfakto/mfakto-wrappertest folder has this in job.xml:

<job_desc> <task> <application>./mfakto-x64</application> <command_line>-d 0</command_line> </task> <unzip_input> <zipfilename>mfakto-linux64-v10.zip</zipfilename> </unzip_input> </job_desc>

Where do you have the files? Must be a permission issue somewhere.

ID: 9843 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

What files? R we talking wrapper-test or BOINC world?

ID: 9844 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

What files? R we talking wrapper-test or BOINC world?

wrapper test

ID: 9845 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

/home/sharon/mfacto/mfakto-wrappertest/

has these files in it:

barrett15.cl barrett.cl boinc_finish_called common.cl COPYING datatypes.h gpusieve.cl job.xml mfakto.ini mfakto_Kernels.cl mfakto_Kernels.elf mfakto-x64 montgomery.cl mul24.cl README.txt results.txt stderr.txt tf_debug.h todo.txt worktodo.txt wrapper_26018_linux-u18_x86-64 wrapper_checkpoint.txt wrapperdir

ID: 9846 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

/home/sharon/mfacto/mfakto-wrappertest/

has these files in it:

barrett15.cl barrett.cl boinc_finish_called common.cl COPYING datatypes.h gpusieve.cl job.xml mfakto.ini mfakto_Kernels.cl mfakto_Kernels.elf mfakto-x64 montgomery.cl mul24.cl README.txt results.txt stderr.txt tf_debug.h todo.txt worktodo.txt wrapper_26018_linux-u18_x86-64 wrapper_checkpoint.txt wrapperdir

ok, then sudo is ok, now we need to know why its not running in BOINC, all looks fine.

ID: 9847 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

Trying to show what that variable is set to (that s/b = 0) by going into the boinc folder where srbase runs from .... as root.. gives me nothing back so I assume it's set to null.

root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5# echo $GPU_DEVICE_NUM

ID: 9848 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

Trying to show what that variable is set to (that s/b = 0) by going into the boinc folder where srbase runs from .... as root.. gives me nothing back so I assume it's set to null.

root@sharon-ES7C95:/var/lib/boinc-client/projects/srbase.my-firewall.org_sr5# echo $GPU_DEVICE_NUM

Yes, every card is counting from 0, if you have AMD and nvidia both starting from 0, if you have 2 AMDs then 0, 1

ID: 9849 · Rating: 0 · rate:

Reply Quote

rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 2 Jan 13
Posts: 7255
Credit: 42,729,227
RAC: 2
2M in Sierpinski / Riesel Base credit

In summary of both tests on vega20 and 6000er:

- wrapper standalone test was successful without errors

Now we need to find out why they failed in BOINC

ID: 9850 · Rating: 0 · rate:

Reply Quote

Skip Da Shu
Send message
Joined: 27 Mar 23
Posts: 24
Credit: 39,719,772
RAC: 186,770
100k in Sierpinski / Riesel Base credit

Yes, counted from 0... back when I had a 5700G APU/CPU the discrete card was 0 & the iGPU was 1. I excluded the iGPU in cc_config.xml.

I'm still curious how $GPU_DEVICE_NUM is set.

ID: 9851 · Rating: 0 · rate:

Reply Quote