It knows there are 3 GPU's but doesn't know the exact type. WUProp isn't differentiating between the 7870, 79. I think zombie67's results highlight the problem very well: Message 1417 - Posted:, 8:32:29 UTC - in response to Message 1415. GPU name: AMD Radeon HD 7870/7950/7970 series (Tahiti) But for several GPU projects, they run multiple tasks per GPU. These two machines have three physical GPUs. Unfortunately, I had a couple of machines with both a discrete GPU as well as an Intel HD 4000. I assume this is intended for only windows machines, and then only those with more than one GPU. Can you also point out if detection is correct. Can you post file content together with BOINC version. If you have a similar machine, can you please run this application? I'm searching how I can improve gpu detection for machines which have more than one gpu. Message 1415 - Posted:, 5:29:22 UTC - in response to Message 1402. " and then click on the "post reply" button. " post the comment then go back and EDIT my post and ADD the "markup flags for url=. I had to initially REMOVE ALL "markup flags for url=. I had the same problem earlier today when I was posting to another forum. ON THE ELECTRODYNAMICS OF MOVING BODIES By A. ) In (unsuccessful) attempt to overcome this 'anti-spam system' I even tried to add to my post (which is now in ) big parts of: I think this 'Akismet anti-spam system' have to be disabled for users with amount of Credit/RAC (e.g. which was flagged as "spam" by 'Akismet anti-spam system'? If you want to read my perfectly normal post: I have to use strange methods to post here. Message 1414 - Posted:, 5:01:34 UTC - in response to Message 1413. I can swap in/out up to 3 discrete GPU's to test if you want. "GPU compute units: 5" correct for the GTX660, but the GTX660Ti has 7. 21:08:28 | | OpenCL: Intel GPU 0: Intel(R) HD Graphics 4000 (driver version 9., device version OpenCL 1.2, 1496MB, 1496MB available, 45 GFLOPS peak)īoinc lists GPU's as CUDA and then OpenCL, so it will list the NVidia GPU's twice. 21:08:28 | | OpenCL: NVIDIA GPU 1: GeForce GTX 660 (driver version 314.22, device version OpenCL 1.1 CUDA, 2048MB, 1968MB available, 1982 GFLOPS peak) 21:08:28 | | OpenCL: NVIDIA GPU 0: GeForce GTX 660 Ti (driver version 314.22, device version OpenCL 1.1 CUDA, 2048MB, 1959MB available, 2985 GFLOPS peak) 21:08:28 | | CUDA: NVIDIA GPU 1: GeForce GTX 660 (driver version 314.22, CUDA version 5.0, compute capability 3.0, 2048MB, 1968MB available, 1982 GFLOPS peak) 21:08:28 | | CUDA: NVIDIA GPU 0: GeForce GTX 660 Ti (driver version 314.22, CUDA version 5.0, compute capability 3.0, 2048MB, 1959MB available, 2985 GFLOPS peak) However, Boinc reads them correctly (Boinc event log): WRT reporting 2 GeForce GTX 660 GPU's, that on par with GPUGrid and Einstein:Ĭoprocessors NVIDIA GeForce GTX 660 (2047MB) driver: 314.22 Message 1410 - Posted:, 20:26:09 UTC - in response to Message 1407.
0 Comments
Leave a Reply. |