mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU Computing (https://www.mersenneforum.org/forumdisplay.php?f=92)
-   -   mfakto: an OpenCL program for Mersenne prefactoring (https://www.mersenneforum.org/showthread.php?t=15646)

ixfd64 2019-11-22 22:54

What platform are you compiling for?

Uncwilly 2019-11-23 01:54

[QUOTE=ixfd64;531284]What platform are you compiling for?[/QUOTE]Aye, there's the rub. I don't have a compiler. It is Windows 10, 64-bit.

Uncwilly 2019-12-10 16:02

[QUOTE=Uncwilly;531265]I was handed a new machine. It has an Intel integrated processor. Specifically an Intel UHD Graphics 630..[/QUOTE]
[QUOTE=Uncwilly;531283]I tried to run it in "PowerShell" and get the following at the output (I trimmed the good.)[CODE]Compiling kernels.

BUILD OUTPUT

Unrecognized build options: -O3
END OF BUILD OUTPUT
Error -11 (Build program failure): clBuildProgram
ERROR: load_kernels(0) failed[/CODE][/QUOTE]
[QUOTE=ixfd64;531284]What platform are you compiling for?[/QUOTE]
I am up and running. Here is what what I did that eventually got me going.:[LIST][*]Copied the mfakto folder from my laptop to the new machine.[*]Set GPUType to INTEL[*]Renamed the existing mafakto_Kernels.elf[*]Ran mfakto[/LIST]The second and third steps seemed to be what worked.

Currently ~20.5 GHz-D/D

kriesel 2019-12-10 19:15

[QUOTE=Uncwilly;532554]I am up and running. Here is what what I did that eventually got me going.:[LIST][*]Copied the mfakto folder from my laptop to the new machine.[*]Set GPUType to INTEL[*]Renamed the existing mafakto_Kernels.elf[*]Ran mfakto[/LIST]The second and third steps seemed to be what worked.

Currently ~20.5 GHz-D/D[/QUOTE]Excellent. Running prime95 also, on what hardware? I saw on my UHD630, 20+ w/o prime95, 18 with, and prime95 was about half speed with both going, on an i8750H system & Win 10.

Uncwilly 2019-12-13 23:04

2 Attachment(s)
[QUOTE=kriesel;532562]Excellent. Running prime95 also, on what hardware? I saw on my UHD630, 20+ w/o prime95, 18 with, and prime95 was about half speed with both going, on an i8750H system & Win 10.[/QUOTE]Yes, running Prime95 too on an OptiPlex 7060. The specifics are 6 core i5-8500 running at 3.00GHz with 16 GB running Windows 10 Pro for Workstations (64 bit).
Prime95 reports
[CODE]3.465.05 MHZ
Prefetchw, SSE, SSE2, SSE4, AVX, AVX2, FMA
L1 size 6x32 KB, L2 size 6x256 KB, L3 size 9MB[/CODE]


Currently I have P95 set to run 1 DC on 4 cores and 1 P-1 on the other 2 (with 5GB day and 10GB at night)

kriesel 2019-12-13 23:57

[QUOTE=Uncwilly;532858]
Currently I have P95 set to run 1 DC on 4 cores and 1 P-1 on the other 2 (with 5GB day and 10GB at night)[/QUOTE]I try to make memory constant for P-1 whenever possible. It will restart the task when memory available changes. I think that's documented behavior of prime95/mprime.

axn 2019-12-14 02:00

[QUOTE=kriesel;532865]I try to make memory constant for P-1 whenever possible. It will restart the task when memory available changes. I think that's documented behavior of prime95/mprime.[/QUOTE]

Yes; and in restarting, it will wipeout any efficiency improvements you get from moving to 5GB->10GB. Constant 5GB is definitely better.

Uncwilly 2019-12-14 02:07

[QUOTE=axn;532881]Yes; and in restarting, it will wipeout any efficiency improvements you get from moving to 5GB->10GB. Constant 5GB is definitely better.[/QUOTE]I will fix that next time I am in front of the machine.

kotenok2000 2019-12-28 13:12

[QUOTE=Bdot;263174]This is an early announcement that I have ported parts of Olivers (aka TheJudger) mfaktc to OpenCL.

Currently, I have only the Win64 binary, running an adapted version of Olivers 71-bit-mul24 kernel. Not yet optimized, not yet making use of the vectors available in OpenCL. A very simple (and slow) 95-bit kernel is there as well so that the complete selftest finished successfully on my box.

On my HD5750 it runs about 60M/s in the 50M exponent range - certainly a lot of headroom :smile:

As I have only this one ATI GPU I wanted to see if anyone would be willing to help testing on different hardware.

Current requirements: OpenCL 1.1 (i.e. only ATI GPUs), Windows 64-bit.

There's still a lot of work until I may eventually release this to the public, but I'm optimistic for the summer.

Next steps (unordered):
[LIST][*]Linux port (Is Windows 32-bit needed too?)[*]check, if [URL]http://mersenneforum.org/showpost.php?p=258140&postcount=7[/URL] can be used (looks like it's way faster)[*]fast 92/95-bit kernels (barrett)[*]use of vector data types[*]various other performance/optimization tests&enhancements[*]of course, bug fixes:boxer:[*]docs and licensing stuff :yucky:[*]clarify if/how this new kid may contribute to primenet[/LIST]Bdot[/QUOTE]
Why is this link missing?
[url]https://www.mersenneforum.org/mfakto/[/url]

kriesel 2019-12-28 14:01

[QUOTE=kotenok2000;533690]Why is this link missing?
[URL]https://www.mersenneforum.org/mfakto/[/URL][/QUOTE]
Probably because the repository did not exist until years after Bdot wrote the post you quoted. But see [url]https://www.mersenneforum.org/showthread.php?t=24607[/url] for the current compilation of reference info, including the tabulation of useful software for computation type and hardware combination [url]http://www.mersenneforum.org/showpost.php?p=488291&postcount=2[/url]

ixfd64 2019-12-29 19:13

1 Attachment(s)
I've updated my script for launching multiple mfakto instances. It is now more compact as it uses fewer variables and symbolic links.


All times are UTC. The time now is 13:00.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.