Very educational Zydor

Hello everybody !!

Thank you very much for your post, Zydor !!
The new range of settings in CCC is created.
Yesterday, I tried to crunch new WU, whith these new settings of Memory Clock.

Nothing changes, I stopped it because the crunching time is always as long as before.

It's really very strange !!

I think that I will stop this project...

Thank you for your help !!

Regards.

Hello everybody !!

Thank you very much for your post, Zydor !!
The new range of settings in CCC is created.
Yesterday, I tried to crunch new WU, whith these new settings of Memory Clock.

Nothing changes, I stopped it because the crunching time is always as long as before.

It's really very strange !!

I think that I will stop this project...

Thank you for your help !!

Regards.

Would you consider to take away one gpu and try with a singel card? Just to find out if the process of the WU has anything to do with multi gpu. I know this is a strange request, but maybe worth trying. Maybe I´m stupid! Is it two GPU´s on a single card or two cards?

---

Hello Bernt !

I have two cards.
That's a good idea to try with only one card to see what's happening.
I'll try as soon as possible and post result.

Thanks.

Regards.

I use Catalyst 11.9 suite with my 5850 and The biggest problem I see Is GPU temperature...

Setting is 850 and memory 1100 with a max temperature of 74 degrees Celsius.

I have read a few things about ATI and drivers and other problems.

They might be worth trying...

1. disable SLI-capability
2. drivers 11.2, 11.3 maybe 11.4 and driver 11.9 works for me....uninstall old drivers first, then reboot, then install new drivers.
3. Use 2monitors, 1 on each cards RGB connector, or a hdmi to rgb adapter on one card and plug it into a TV-set or monitor that is on. I don't think The TV needs to be in HDMI mode, just keep the TV on so that HDMI can communicate.

4. download gpu-z from www.download.com or msi afterburner mentioned in the thread above.

---

[quote].... Moo is an OpenCL application. Collatz and Milkyway are written in the old ATI native language or NVidia CUDA, and inherently faster
...
Moo runs inside the BOINC Wrapper. The Wrapper is an application written for BOINC that allows a core application to run using ANY language. For Moo, the Core application is made by DNet (not sure what it uses, probably C++ at a guess). However the Wrapper is written to interface with Moo using OpenCL to talk to the Wrapper, and its that the BOINC Client "sees". Its the fact that Moo uses OpenCL in talking to the Wrapper which makes Moo vulnerable to the OpenCL bug.

Yes, we are using (a modified) wrapper but that part has no card related functions so it's neither OpenCL nor native application on itself. BOINC Client has some detection code related (existence, features and memory size, etc) to cards (native and lately OpenCL as well) but otherwise it doesn't have any card related functions.

All the card related functions are in the core D.Net Client application, which is written in C (perhaps mixed with some C++). To my knowledge that's not an OpenCL application but rather is written natively either for ATI Stream or nVidia CUDA platforms*. I'm not sure why we are affected by the OpenCL bug in the drivers. :( I have a feeling it's more complicated bug or there's more problems in certain versions of drivers (especially in mixed or multiple card configurations).

-w

*Source code for D.Net Client is available at http://www.distributed.net/Source and it's Stream and CUDA parts look similar to C++ which I believe is a sign of native application and not OpenCL.