Thanks for info.

But now I'm confused and I get the question, why published MW@h a new nVIDIA app (OpenCL) and the old nVIDIA app (CUDA23) crunch happy the currently WUs?

The new nVIDIA OpenCL app use the whole Fermi chip (GTX4xx/5xx)?

Maybe an admin of MW@h could answer if it's allowed to use still the old nVIDIA CUDA23 app (which would be the only way for me to crunch MW@h)?
Thanks.

Maybe it's possible to send out the old CUDA23 app to hosts with < 197.x driver and the new OpenCL app to hosts with >/= 197.x driver?

Please read this http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1505&nowrap=true#46230
Cuda should still work.

BTW.
IIRC, this isn't the stock 0.24_cuda23 app of MW@h.
It's Crunch3r's Fermi mod which have arkayn on his DL site.

But I guess the stock 0.24_cuda23 app of MW@h should work then on old </= GTX2xx GPUs.

But it would be nice to know from an admin, if it's still allowed to use the stock/mod CUDA23 app.

---

Other than the few WU that were too large and errored out, I seem to be running OpenCL WU on my 64 bit Linux 0.54 app successfully. Just looked at two WU, just reported and they validated. Maybe Matt needs to look at the build of the 0.52 Windows app and see if he missed something like the obvious goof he made on the Linux 0.52 app.

No, the Linux one was a huge mistake which would never work for anyone. This Windows problem is more subtle, and I haven't been able to reproduce it. It seems like it crashes very early on fairly consistently for some people, but works for some people. It has been working fine on my systems.

Please read this http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1505&nowrap=true#46230
Cuda should still work.

BTW.
IIRC, this isn't the stock 0.24_cuda23 app of MW@h.
It's Crunch3r's Fermi mod which have arkayn on his DL site.

But I guess the stock 0.24_cuda23 app of MW@h should work then on old
But it would be nice to know from an admin, if it's still allowed to use the stock/mod CUDA23 app.

It should work for now. At some point after I get the new ATI application ready things will start breaking for old stuff.

Matt, there seems to be a lot of OpenCL tasks today that error out almost immediately on the Linux 0.54 app. Yet, some tasks in the same group of 24 downloads process correctly. I don't know if there is a mix of bad tasks from the oversized ones still being distributed but this might tie into the supposed computation errors on the Windows 0.52 app.

Keith

---

All my linux hosts seem to be failing all tasks with the new 0.52 OpenCL app.

I made a really dumb mistake in the Linux build. Should be fixed now (0.54).

Many thanks, it seems to be working now

At some point after I get the new ATI application ready things will start breaking for old stuff.

Is it already known if the new ATI application will have the same requirements as the old one i.e. will it run on all ATI GPUs listed in the GPU Requirements Thread?

---

At some point after I get the new ATI application ready things will start breaking for old stuff.

Is it already known if the new ATI application will have the same requirements as the old one i.e. will it run on all ATI GPUs listed in the GPU Requirements Thread?

Yes, it should.

At some point after I get the new ATI application ready things will start breaking for old stuff.

Is it already known if the new ATI application will have the same requirements as the old one i.e. will it run on all ATI GPUs listed in the GPU Requirements Thread?

Yes, it should.

except for the fact that OpenCL runs only on HD 4xxx and higher (so no more 3xxx)...
In addition, if you use any OpenCL 1.1 instruction, then only HD 5xxx and above will work...

At some point after I get the new ATI application ready things will start breaking for old stuff.

Is it already known if the new ATI application will have the same requirements as the old one i.e. will it run on all ATI GPUs listed in the GPU Requirements Thread?

Yes, it should.

except for the fact that OpenCL runs only on HD 4xxx and higher (so no more 3xxx)...
In addition, if you use any OpenCL 1.1 instruction, then only HD 5xxx and above will work...

Well, according to http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1477&nowrap=true#46691, he's not using opencl, which is a very good thing!!

Instead, CAL++ (or calpp) is being used,which supports 38xx cards as well ;)

---

Join Support science! Joinc Team BOINC United now!