Computation Error

[danhtruong]

Hi all,

I use my HD 6850 and HD 6870 in CrossfireX, both underclocked to 700MHz to keep temps in check and have done this for about 3 months now.
However, occassionally, I get a computation error for Collatz Conjecture.

When i check the task ID it gives me: 1 Stream Allocation : Failed to create Buffer

I might have approximately 40 in Collatz workunits in queue and when i get a computation error, i then get it for all 40 of them until i run out of units to crunch.
Normally a restart and an update fixes the problem, but has anyone come across this error before and know why we get it?

I thought it was the GPU's overheating, but it runs at a nice 70 degrees C, so i doubt that's the problem.

Any ideas?

[kashi]

Yes, seen this many times before.

One possible cause is not enough available system memory to support GPU processing on two GPUs. This is possible even when Task Manager shows there is still free memory available.

Two GPUs with 1GB of video memory causing "1 Stream Allocation : Failed to create Buffer" errors when paired with 8 core computer with 3.3GB of available system memory.

Possible solutions:

1. Install additional memory/64-bit OS in computer.

2. On CPU, switch to a WCG sub project or another project that uses less memory per core.

3. (For 32-bit Windows OS and Catalyst driver 9.2 and above)
Empty your cache of Collatz GPU tasks either by processing them or aborting them. Download and install (follow instructions in readme file) "Win32_2.09_ATI.zip (Win32 ATI with ATICALxx.DLL files)" from Power (Optimized) Applications page. Set <cmdline>r25</cmdline> in app_info.xml file. Test and see if amount of GPU memory allocated to step array has been reduced from 256MB to 128MB (see this in stderr out section of task details). It may be necessary to reduce r parameter further to reduce GPU memory allocated to step array to 64MB. If there is still not enough memory available after that to create the buffer there may be no other option but to use a 64-bit operating system that is able to use more than 3GB memory and possibly add some more memory as well.

Option 3 has worked for the majority of people who have tried it.

These are just suggestions, hopefully one of them will work for you or the cause may be something else.