Trial Factoring
log in

Advanced search

Message boards : Number crunching : Trial Factoring

Previous · 1 . . . 20 · 21 · 22 · 23
Author Message
magic_sam
Send message
Joined: 12 Aug 23
Posts: 9
Credit: 1,979,288
RAC: 0
Message 9128 - Posted: 9 Oct 2023, 17:09:19 UTC - in response to Message 9127.

Hello rebirther,

root@kymera:~/mfakto# ./mfakto-x64 -st [...] ######### testcase 34071/34071 (M112404491[91-92]) ######### Starting trial factoring M112404491 from 2^91 to 2^92 (4461450.54GHz-days) Date Time | class Pct | time ETA | GHz-d/day Sieve Wait Oct 09 17:05 | 1848 0.1% | 0.005 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 0.15pre7 cl_barrett32_92_gs_2] selftest for M112404491 passed (cl_barrett32_92_gs)! tf(): total time spent: 0.005s Selftest statistics number of tests 34026 successful tests 34026 selftest PASSED!


Thank you for your help.

Best regards, Samuel

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9129 - Posted: 9 Oct 2023, 17:38:33 UTC - in response to Message 9128.

looks good, try this next

./mfakto-64 --perftest

magic_sam
Send message
Joined: 12 Aug 23
Posts: 9
Credit: 1,979,288
RAC: 0
Message 9130 - Posted: 9 Oct 2023, 17:57:22 UTC - in response to Message 9129.

This one is looking good too (IMHO):

root@kymera:~/mfakto# ./mfakto-x64 --perftest [...] Best GPUSieveSize for SievePrimes: 54 310 1078 1078 1846 2614 3382 5686 8502 13622 19766 31030 47414 74038 113206 175670 270902 419382 647734 1075766 at MiB: 105 48 127 24 36 96 102 101 128 16 120 103 104 101 102 20 24 104 125 124 max M/s: 1199743.7 1078251.6 866078.2 747935.3 543832.0 524124.2 433830.7 464281.5 367334.0 418859.2 445633.7 297175.1 401461.2 229553.6 188893.7 149561.5 102728.1 80825.1 53634.5 36243.5 Survivors: 48.30% 35.49% 29.96% 29.96% 28.09% 27.01% 26.26% 24.87% 23.90% 22.86% 22.11% 21.26% 20.52% 19.80% 19.16% 18.53% 17.95% 17.41% 16.90% 16.36% removal rate average: 620327.4 695610.0 606612.3 523861.5 391064.5 382577.3 319915.2 348799.9 279529.3 323090.5 347108.6 234002.2 319085.1 184102.5 152707.9 121846.2 84287.3 66756.4 44569.5 30315.6 incremental: n/a 1363749.0 243329.1 -14.2 37227.9 156860.3 18844.4 -91603.0 17063.3 -31027.3 -52648.6 7592.9 -8451.2 3856.0 6857.7 4493.7 1902.8 2065.0 804.9 610.1 Runtime options Inifile 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 120 s Stages enabled StopAfterFactor class PrintMode compact V5UserID none ComputerID none TimeStampInResults no VectorSize 2 GPUType AUTO SmallExp no UseBinfile mfakto_Kernels.elf Could not read TestExponents from mfakto.ini - not testing TF kernels

Best regards, Samuel

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9131 - Posted: 9 Oct 2023, 18:11:34 UTC - in response to Message 9130.

Iam missing some infos, try on a sample,

create a worktodo.txt in the same folder with a line

Factor=DC8E3C41903B44953A316B21D5709403,312745357,73,74

and run ./mfakto-64

you will get a result.txt with percentage status lines and finished if all is going well.

magic_sam
Send message
Joined: 12 Aug 23
Posts: 9
Credit: 1,979,288
RAC: 0
Message 9132 - Posted: 9 Oct 2023, 18:38:37 UTC - in response to Message 9131.
Last modified: 9 Oct 2023, 18:39:13 UTC

Select device - Get device info: WARNING: Unknown GPU name, assuming GCN. Please post the device name "gfx1100 (Advanced Micro Devices, Inc.)" to http://www.mersenneforum.org/showthread.php?t=15646 to have it added to mfakto. Set GPUType in mfakto.ini to select a GPU type yourself to avoid this warning. OpenCL device info name gfx1100 (Advanced Micro Devices, Inc.) device (driver) version OpenCL 2.0 (3590.0 (HSA1.1,LC)) maximum threads per block 1024 maximum threads per grid 1073741824 number of multiprocessors 48 (3072 compute elements) clock rate 2371 MHz Automatic parameters threads per grid 0 optimizing kernels for GCN Loading binary kernel file mfakto_Kernels.elf Compiling kernels. GPUSievePrimes (adjusted) 82486 GPUsieve minimum exponent 1055144 Started a simple selftest ... Selftest statistics number of tests 30 successful tests 30 selftest PASSED! got assignment: exp=312745357 bit_min=73 bit_max=74 (6.12 GHz-days) Starting trial factoring M312745357 from 2^73 to 2^74 (6.12GHz-days) Using GPU kernel "cl_barrett15_74_gs_2" Date Time | class Pct | time ETA | GHz-d/day Sieve Wait Oct 09 18:29 | 4619 100.0% | 0.232 0m00s | 2372.92 82486 0.00% no factor for M312745357 from 2^73 to 2^74 [mfakto 0.15pre7 cl_barrett15_74_gs_2] tf(): total time spent: 3m 41.448s (2386.55 GHz-days / day)


And results.txt

root@kymera:~/mfakto# more results.txt no factor for M312745357 from 2^73 to 2^74 [mfakto 0.15pre7 cl_barrett15_74_gs_2 ]


Best regards, Samuel

P.S: now to bed, waking up early in the morning :)

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9133 - Posted: 9 Oct 2023, 18:50:44 UTC - in response to Message 9132.

perfect, now we need to know why the wrapper is causing issues. Recreate the worktodo.txt with the same content and run only the wrapper file, should have the same result.

magic_sam
Send message
Joined: 12 Aug 23
Posts: 9
Credit: 1,979,288
RAC: 0
Message 9139 - Posted: 11 Oct 2023, 17:04:07 UTC - in response to Message 9133.

Hello,

Silly question, but which wrapper file ?

As far as I can tell there are two of them (wrapper_26016-lt6_x86_64-pc-linux-gnu), one in the slot directory, and one in /var/lib/boinc/projects/srbase.my-firewall.org_sr5. The latter is executable and much bigger than the former.

Best regards, Samuel

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9141 - Posted: 14 Oct 2023, 14:20:23 UTC - in response to Message 9139.

Hello,

Silly question, but which wrapper file ?

As far as I can tell there are two of them (wrapper_26016-lt6_x86_64-pc-linux-gnu), one in the slot directory, and one in /var/lib/boinc/projects/srbase.my-firewall.org_sr5. The latter is executable and much bigger than the former.

Best regards, Samuel


the one from linux

Luminescence
Send message
Joined: 12 Sep 21
Posts: 4
Credit: 11,327,616
RAC: 1
Message 9370 - Posted: 10 Dec 2023, 7:46:03 UTC

I've looked into the "device 0" issue.

Under https://boinc.berkeley.edu/trac/wiki/WrapperApp it is stated:

command_line
command-line arguments to be passed to the worker program. This string is macro-substituted as follows:
$NTHREADS is replaced with the number of CPUs the client is allocating for this job.
$GPU_DEVICE_NUM is replaced with the device number of the GPU allocated to this job.


The current jobfile only contains

<task> <application>mfaktc-win-64.exe</application> <append_cmdline_args/> </task>


mfaktc ignores any command-line arguments it doesn't know, so it ignores the appended
--device N
passed to the wrapper.

Could the following change to the jobfile work?

<task> <application>mfaktc-win-64.exe</application> <command_line>-d $GPU_DEVICE_NUM</command_line> </task>


The wrapper should replace $GPU_DEVICE_NUM with the device number it is passed from the BOINC client and mfaktc recognizes the argument -d.

Otherwise it would be trivial to get mfaktc to recognize --device.

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9371 - Posted: 10 Dec 2023, 9:17:16 UTC - in response to Message 9370.

You can ignore that:

The current jobfile only contains <task> <application>mfaktc-win-64.exe</application> <append_cmdline_args/> //this config is empty </task>


You should add this to your app_config.xml

<cmdline>-d $GPU_DEVICE_NUM</cmdline>


Is this only for mfaktc or mfakto too? Can it work with multi-GPU?

Luminescence
Send message
Joined: 12 Sep 21
Posts: 4
Credit: 11,327,616
RAC: 1
Message 9372 - Posted: 10 Dec 2023, 13:30:52 UTC - in response to Message 9371.

You can ignore that:

The current jobfile only contains <task> <application>mfaktc-win-64.exe</application> <append_cmdline_args/> //this config is empty </task>


You should add this to your app_config.xml

<cmdline>-d $GPU_DEVICE_NUM</cmdline>


Is this only for mfaktc or mfakto too? Can it work with multi-GPU?


Looking at mfakto's source, it should work as well. Takes the same -d argument.

Unfortunately I can't test this stuff, as I don't have a multi-gpu system. Not too sure if it works with the app_config.xml, someone else would have to try.

Speedy51
Send message
Joined: 7 Feb 18
Posts: 70
Credit: 260,372,797
RAC: 505,306
Message 9377 - Posted: 13 Dec 2023, 7:21:27 UTC
Last modified: 13 Dec 2023, 7:28:21 UTC

Since I have upgraded to a RTX 4080, where am I best to focus my resource? I can complete a task (full length) here in 1 minute 30 seconds or should I help at GPU72 where I complete LL numbers in about 30 minutes each? (full length) using mfaktc-win-64

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9378 - Posted: 13 Dec 2023, 9:40:42 UTC - in response to Message 9377.

Since I have upgraded to a RTX 4080, where am I best to focus my resource? I can complete a task (full length) here in 1 minute 30 seconds or should I help at GPU72 where I complete LL numbers in about 30 minutes each? (full length) using mfaktc-win-64


Its your choice.

Speedy51
Send message
Joined: 7 Feb 18
Posts: 70
Credit: 260,372,797
RAC: 505,306
Message 9379 - Posted: 13 Dec 2023, 20:13:01 UTC - in response to Message 9378.

Since I have upgraded to a RTX 4080, where am I best to focus my resource? I can complete a task (full length) here in 1 minute 30 seconds or should I help at GPU72 where I complete LL numbers in about 30 minutes each? (full length) using mfaktc-win-64


Its your choice.

Thanks Reb. I am asking for project input as I would like to put my resource in the best possible place for the project

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7476
Credit: 43,522,725
RAC: 31,022
Message 9581 - Posted: 27 Jan 2024, 22:17:26 UTC

We could have a solution for multi-gpu. Just waiting for updated wrapper files. More infos coming soon in the news.

Previous · 1 . . . 20 · 21 · 22 · 23
Post to thread

Message boards : Number crunching : Trial Factoring


Main page · Your account · Message boards


Copyright © 2014-2024 BOINC Confederation / rebirther