Funny thing is that the system with the A10-5700 has three HD6670 cards, and it own IGP, and crunches four of the GPU72 WUs at the same time.

I get the idea that in Radeon-coding the option '--device 0' just doesn't work, not for Linux and not for Windows.

-d 0

That is a code/compiler thing, not something I can change

For all with this error ERROR: unknown option '--device'
try in standalone with
./mfakto -d 0 -st (mfakto must be renamed from zipfile)

If its running I can add the -d option to the template but also need to test on windows. Normally the program taking always the first GPU by default.

I have updated the FAQ, mfakto has no multi-GPU support but mfaktc has it (all GPUs running on one WU) so I cant use the -d option in template. If you have success to run it in standalone we need another solution.

Makes no sense about no multi-GPU support. 1 WU per GPU shouldn't be this hard. Every other project that uses GPU can do it. So what's the problem here?

Well, sorry but that's just lame.

It makes the other GPU not usable by other projects because BOINC thinks it's in use even thought it isn't.

Well, sorry but that's just lame.

It makes the other GPU not usable by other projects because BOINC thinks it's in use even thought it isn't.

Are you relating to mfakto or mfaktc? I can report this to the devs if there are some issues.

Both, does it on both AMD and NVIDIA systems with 2 GPUs in each. Shows 2 tasks running but only 1 GPU with utilization.

On Windows machines:
http://srbase.my-firewall.org/sr5/hosts_user.php?userid=2080

Lovely, So I have to tell the app how to run my machine properly instead of the app telling my machine how to run properly.

You have an app_config.xml template for this I could use?

Yes.

Yeah, that's not going to work. I basically had that in my app_config and cc_config anyway. Nothing there tells the app to use all GPUs on 1 WU. Only to run 1 WU at a time per GPU and 1 WU max per project. The app is not doing it's job properly by not using all GPUs towards 1 WU on NVIDIA.

I'm also not even sure it knows which device to even use as in one system (NVIDIA) its using device 1, not device 0. And in another system its using device 0, not device 1. You would think it would use the 1st device 0 by default. Both have identical 2x GPUs per system.

Nope, no errors. Both HWiNFO and Afterburner show the 1st GPU deivce 0 being ignored/not used and device 1 being used in my 2x 1660ti setup. Only device 1 shows any utilization in those 2 programs.