Hi,
when I register a new client Boinc does a CPU benchmark.
But there is no benchmark for my GPU.

Why not?

GPUs do the work more efficiently and are more powerfull than CPUs.
But without a GPU benchmark there is no classification of GPU power possible.

regards
Veit

The relative speed of a GPU is calculated from the card geometry when the device is detected and queried by BOINC at startup. That "GFLOPS peak" value is used instead of a benchmark.

Hi,
where or how can I see this value?

Veit

In BOINC's Event Log at startup. This machine says

28-Mar-2016 09:23:02 [---] CUDA: NVIDIA GPU 0: GeForce GTX 970 (driver version 350.12, CUDA version 7.0, compute capability 5.2, 4096MB, 3903MB available, 4087 GFLOPS peak)
28-Mar-2016 09:23:02 [---] CUDA: NVIDIA GPU 1: GeForce GTX 750 Ti (driver version 350.12, CUDA version 7.0, compute capability 5.0, 2048MB, 1967MB available, 1639 GFLOPS peak)
28-Mar-2016 09:23:02 [---] OpenCL: NVIDIA GPU 0: GeForce GTX 970 (driver version 350.12, device version OpenCL 1.2 CUDA, 4096MB, 3903MB available, 4087 GFLOPS peak)
28-Mar-2016 09:23:02 [---] OpenCL: NVIDIA GPU 1: GeForce GTX 750 Ti (driver version 350.12, device version OpenCL 1.2 CUDA, 2048MB, 1967MB available, 1639 GFLOPS peak)
28-Mar-2016 09:23:02 [---] OpenCL: Intel GPU 0: Intel(R) HD Graphics 4600 (driver version 10.18.10.3621, device version OpenCL 1.2, 1298MB, 1298MB available, 192 GFLOPS peak)

Note that I usually refer to these as theoretical or 'Marketing' FLOPS - they are never achieved in practice by any practical scientific computing program.

Hi,
OK I had found it.
Is this the SP or the DP peak?


How can I repeat this after an upgrade of my graphic card?


veit

I think it's SP. Now, if you can upgrade you graphic card without restarting BOINC, I'd like to know how. Otherwise, look at the Event Log after the next restart.