mfakto: an OpenCL program for Mersenne prefactoring - Page 47

Bdot · 2012-07-29, 21:53

I just released the Windows-version of mfakto 0.12 on
http://mersenneforum.org/mfakto/mfakto-0.12/

These are the changes:

IMPORTANT: bugfix for missing factors for certain exponent ranges. Thanks to dabaichi and Axelsson for bug report and help on http://mersenneforum.org/showthread....77&page=16#383
handling of <worktodo>.add (worktodo file name as configured, with suffix replaced): will be appended to <worktodo> between 5 and ~10 minutes after the add file's creation, or when moving to the next assignment
new mfakto.ini key SieveCPUMask to set the siever thread's CPU affinity
auto-detection of the GPU-type, along with new GPUType key in mfakto.ini, removed PreferKernel key
optimization for GCN (Graphics Core Next: HD77xx, HD78xx, HD79xx)
improved estimation of the number of compute units
bugfix: occasional abort on very high SievePrimes (> 450,000)
progress line output format: percent complete now %5.1f (was %6.2f), exponent now %-10u (was %d)
bugfix: if a kernel file cannot be found, tell it's name instead of KERNEL_FILE (thanks to Axelsson)
approx. 30,000 new test cases (factors) for pre-release testing (thanks to James Heinrich)

The Linux version will follow tomorrow (probably).

@kracker: I just found my first new factor with the release candidate, so it is still capable of finding factors ...

sdbardwick · 2012-07-29, 22:19

Cool! Thanks!
Quick question regarding SieveCPUMask:
Value of 3 (0011) would allow running on CPU 0 and 1, while 12 (1100) would designate CPU 2 and 3?
(Obviously not a programmer here...)

kracker · 2012-07-29, 22:57

Quote:

Originally Posted by Bdot

I just released the Windows-version of mfakto 0.12 on
http://mersenneforum.org/mfakto/mfakto-0.12/

These are the changes:

IMPORTANT: bugfix for missing factors for certain exponent ranges. Thanks to dabaichi and Axelsson for bug report and help on http://mersenneforum.org/showthread....77&page=16#383
handling of <worktodo>.add (worktodo file name as configured, with suffix replaced): will be appended to <worktodo> between 5 and ~10 minutes after the add file's creation, or when moving to the next assignment
new mfakto.ini key SieveCPUMask to set the siever thread's CPU affinity
auto-detection of the GPU-type, along with new GPUType key in mfakto.ini, removed PreferKernel key
optimization for GCN (Graphics Core Next: HD77xx, HD78xx, HD79xx)
improved estimation of the number of compute units
bugfix: occasional abort on very high SievePrimes (> 450,000)
progress line output format: percent complete now %5.1f (was %6.2f), exponent now %-10u (was %d)
bugfix: if a kernel file cannot be found, tell it's name instead of KERNEL_FILE (thanks to Axelsson)
approx. 30,000 new test cases (factors) for pre-release testing (thanks to James Heinrich)

The Linux version will follow tomorrow (probably).

@kracker: I just found my first new factor with the release candidate, so it is still capable of finding factors ...

Thanks! is this the same as pre2... ?

Bdot · 2012-07-29, 23:02

Quote:

Originally Posted by kracker

Thanks! is this the same as pre2... ?

yes, except for
progress line output format: percent complete now %5.1f (was %6.2f), exponent now %-10u (was %d)

This is the late change that probably breaks something

Bdot · 2012-07-29, 23:05

Quote:

Originally Posted by sdbardwick

Cool! Thanks!
Quick question regarding SieveCPUMask:
Value of 3 (0011) would allow running on CPU 0 and 1, while 12 (1100) would designate CPU 2 and 3?
(Obviously not a programmer here...)

Yet you got it right. Maybe I need to hire someone expressing my techno-speak more user-friendly ...

NormanRKN · 2012-07-29, 23:38

good work !

NormanRKN · 2012-07-30, 00:19

could it be possible to integrate p-1 testing in the gpu-app ?

Dubslow · 2012-07-30, 00:53

Quote:

Originally Posted by NormanRKN

could it be possible to integrate p-1 testing in the gpu-app ?

That might be possible, if we extend CUDALucas. mfakt* use completely different routines from LL testing and P-1 (which are both relatively similar).

kracker · 2012-07-30, 01:34

Quote:

Originally Posted by Dubslow

That might be possible, if we extend CUDALucas. mfakt* use completely different routines from LL testing and P-1 (which are both relatively similar).

+1
P-1 for gpu would rock...

axn · 2012-07-30, 03:54

Quote:

Originally Posted by Dubslow

That might be possible, if we extend CUDALucas.

Even there, stage 2 is not a trivial thing to extend because of memory issues.

Dubslow · 2012-07-30, 04:12

Quote:

Originally Posted by axn

Even there, stage 2 is not a trivial thing to extend because of memory issues.

[OT]True, but it should theoretically be possible to write S1 save files in Prime95 format to do a proper stage 2.[/OT]

2012-07-29, 21:53	#507
Bdot Nov 2010 Germany 3·199 Posts	mfakto 0.12 available I just released the Windows-version of mfakto 0.12 on http://mersenneforum.org/mfakto/mfakto-0.12/ These are the changes: IMPORTANT: bugfix for missing factors for certain exponent ranges. Thanks to dabaichi and Axelsson for bug report and help on http://mersenneforum.org/showthread....77&page=16#383 handling of <worktodo>.add (worktodo file name as configured, with suffix replaced): will be appended to <worktodo> between 5 and ~10 minutes after the add file's creation, or when moving to the next assignment new mfakto.ini key SieveCPUMask to set the siever thread's CPU affinity auto-detection of the GPU-type, along with new GPUType key in mfakto.ini, removed PreferKernel key optimization for GCN (Graphics Core Next: HD77xx, HD78xx, HD79xx) improved estimation of the number of compute units bugfix: occasional abort on very high SievePrimes (> 450,000) progress line output format: percent complete now %5.1f (was %6.2f), exponent now %-10u (was %d) bugfix: if a kernel file cannot be found, tell it's name instead of KERNEL_FILE (thanks to Axelsson) approx. 30,000 new test cases (factors) for pre-release testing (thanks to James Heinrich) The Linux version will follow tomorrow (probably). @kracker: I just found my first new factor with the release candidate, so it is still capable of finding factors ...

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
gpuOwL: an OpenCL program for Mersenne primality testing	preda	GpuOwl	2718	2021-07-06 18:30
mfaktc: a CUDA program for Mersenne prefactoring	TheJudger	GPU Computing	3497	2021-06-05 12:27
LL with OpenCL	msft	GPU Computing	433	2019-06-23 21:11
OpenCL for FPGAs	TObject	GPU Computing	2	2013-10-12 21:09
Program to TF Mersenne numbers with more than 1 sextillion digits?	Stargate38	Factoring	24	2011-11-03 00:34

2012-07-29, 22:19	#508
sdbardwick Aug 2002 North San Diego County 5·137 Posts	Cool! Thanks! Quick question regarding SieveCPUMask: Value of 3 (0011) would allow running on CPU 0 and 1, while 12 (1100) would designate CPU 2 and 3? (Obviously not a programmer here...)

2012-07-29, 23:38	#512
NormanRKN Jul 2012 Saarland / Germany 2²·17 Posts	good work !

2012-07-30, 00:19	#513
NormanRKN Jul 2012 Saarland / Germany 2²×17 Posts	could it be possible to integrate p-1 testing in the gpu-app ?