mfakto: an OpenCL program for Mersenne prefactoring

[Bdot]

Originally Posted by apsen

I actually tried to install the whole bunch. Every other component got installed but that one. After that I tried to install just the APP but it failed too.

The Catalyst drivers are a separate package - they're not included in the AMD APP SDK. They are the Graphics drivers for your card.

[apsen]

Originally Posted by Bdot

The Catalyst drivers are a separate package - they're not included in the AMD APP SDK. They are the Graphics drivers for your card.

But drivers come with SDK if you get the big package.

[KingKurly]

Bdot,

Thank you for your work on mfakto; I enjoy it and find it useful.

When do you think the program will be ready for us to start using it to submit "no factor found" results? I have found one factor (which I already mentioned to you) but only "no factor" results since then. Until I get a go-ahead, I will continue to NOT submit those results: only factors.

-Greg

[Bdot]

Originally Posted by KingKurly

Bdot,

Thank you for your work on mfakto; I enjoy it and find it useful.

When do you think the program will be ready for us to start using it to submit "no factor found" results? I have found one factor (which I already mentioned to you) but only "no factor" results since then. Until I get a go-ahead, I will continue to NOT submit those results: only factors.

-Greg

I did fixes for the issues reported for 0.07. Currently there are some tests underway because of confusing performance results with 0.07: It seems on HD6xxx, version 0.05 was faster than 0.07 (on HD5xxx, 0.07 is about 10% faster than 0.05 was). If that is true, then I will need one more change before releasing the next version - should come within the next few days.

[AldoA]

Originally Posted by Bdot

This teaches us that we really need a (fairly) up-to-date Catalyst driver. The clinfo output does not even mention your GPU.

Please try again to download and install 11.8. If your antivirus still complains, try to get an updated virus definition, or temporarily disable checking. If all fails, AMD provides a link to older versions - you can try 11.7.

Sorry, but the new driver give me too many problem. However I installed mfakto in my Notebook that it has an ATI Mobility Radeon HD5470 with the 11.7 version, but also with this it doesn't found the GPU. I tried to copy and post the clinfo, but it closed too fast. Have you any advide to what to do? Thanks

[Bdot]

Originally Posted by AldoA

Sorry, but the new driver give me too many problem. However I installed mfakto in my Notebook that it has an ATI Mobility Radeon HD5470 with the 11.7 version, but also with this it doesn't found the GPU. I tried to copy and post the clinfo, but it closed too fast. Have you any advide to what to do? Thanks

To get the output of command-line tools it is usually best to open a command prompt (Run -> cmd, or find the black "Command Prompt" icon). In there, run the tool you want, in this case clinfo. You may need to add the full path to it, or cd (change directory) to where the tool sits.

Redirecting the output to a file is quite useful, such as
clinfo >clinfo.txt or
mfakto --CLtest >mfakto-test.txt

For the installation of the Catalyst drivers and AMD APP SDK I cannot suggest much more ... if both installations succeeded yet the GPU cannot be found, I don't know. You can try some of the sample applications that come with the APP SDK, and see if they find the GPU, but if clinfo does not list it, that's not likely.

[DigiK-oz]

Originally Posted by Bdot

Did anyone else give mfakto a try? Any experiences to share (anything strange happening, suggestions you'd like to get included or excluded for the next versions, performance figures for other GPUs, ...)?

I'm running this version on a SuSE 11.4 box with AMD APP SDK 2.4, and when multiple instances are running I occasionally see one instance hanging. It will completely occupy one CPU core but no GPU resources. It is looping inside some kernel code, being immune to kill, kill -9 or attempts to attach a debugger or gcore. So far, reboot was the only way I know to get rid of it. How can I find out where that hang occurs? And what else could I try to kick such a process without a reboot?

Just set up 0.07 on my WIN7 X64 system with AMD 5770 (driver version 11.8). Having the same problem as reported earlier, selftest fails on 1-5 and 9-11.

I do not have the entire SDK installed, just the drivers including OpenCL. Should I install the entire SDK? Anything else I can do to get started? (I know the source can be changed slightly, but no compiler/SDK handy).

Looking forward to the "official" version we can actually report "No factors" with!

[MrHappy]

I just updated catalyst drivers from 11.6 to 11.8 and.... got errors in the test and mfakto just shuts down...

[Bdot]

Here it is, mfakto version 0.08

This one is now compatible with Catalyst 11.8, and has a few other changes (from the changelog):

• added --help and parameter checking
• exclude single-vectored 72-bit-mul24-kernel (not working with AMD APP SDK 2.5)
• removed THREADS_PER_BLOCK (no longer needed, will be selected automatically based on the GPU capabilities)
• removed slow 95-bit kernel (not usable anyway)
• added to mfakto.ini a config setting
  PreferKernel=mfakto_cl_barrett79|mfakto_cl_71
  as HD5xxx and HD6xxx have their top-speed with different kernels
• tuned the settings for SievePrimesAdjust - should now be usable

Quite important is the new config setting PreferKernel at the end of mfakto.ini. The two possible settings, mfakto_cl_barrett79 and mfakto_cl_71 decide, which of these two kernels will be used for factors between 2⁶⁴ and 2⁷¹. On HD5xxx cards, mfakto_cl_barrett79 seems to be faster while HD6xxx may run better with mfakto_cl_71 (which is the MUL24 kernel). Try it out and let me know .

And feel free to submit your results to primenet with this version.

[Bdot]

Windows 32-bit and 64-bit.

[Bdot]

and the source