# Cuda backend parameters **URL:** \ **Category:** General\ **Created:** [6 January 2020 17:34 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262 "2020-01-06T17:34:40Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![Junting\_Chen](https://avatars.discourse-cdn.com/v4/letter/j/ac8455/32.png) [@Junting\_Chen](https://pyfr.discourse.group/u/Junting_Chen)\ **Post date:** [6 January 2020 17:34 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/1 "2020-01-06T17:34:40Z")
Hello, I am wondering if someone can provide a bit more descriptions on these parameters to optimize performance. As far as I know, when using multiple GPUs, I had to select local-rank for device-id and cuda-aware for mpi-type. When exactly should i be using round-robin and local-rank? And when should i be using standard or cuda-aware? How would you select GiMMiK cutoff? How does it affect accuracy / performance? I believe block-1d and block-2d are determined by GPU’s specification. I am not very familiar with Cuda. Please someone can elaborate a bit. For example I am running pyfr with two Tesla k80s in parallel, what’s the block size for 1d and 2d pointswise kernels? Parameterises the CUDA backend with 1. `device-id` — method for selecting which device(s) to run on: 2. `gimmik-max-nnz` — cutoff for GiMMiK in terms of the number of non-zero entires in a constant matrix: 3. `mpi-type` — type of MPI library that is being used: 4. `block-1d` — block size for one dimensional pointwise kernels: 5. `block-2d` — block size for two dimensional pointwise kernels: Thanks a lot! Junting Chen ---
**Author:** ![fdw](https://avatars.discourse-cdn.com/v4/letter/f/a587f6/32.png) [@fdw](https://pyfr.discourse.group/u/fdw)\ **Post date:** [6 January 2020 22:18 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/2 "2020-01-06T22:18:24Z")
Hi Junting, > As far as I know, when using multiple GPUs, I had to select local-rank > for device-id and cuda-aware for mpi-type. When exactly should i be > using round-robin and local-rank? And when should i be using standard or > cuda-aware? If the GPUs in your system are in compute exclusive mode then round-robin is probably what you want. Otherwise, opt for local-rank. So long as each rank gets its own GPU there should be no impact on performance. In terms of the mpi-type this depends heavily on the hardware you're running on and the MPI library you're using. If your MPI library is CUDA aware then setting mpi-type = cuda-aware can improve performance. > How would you select GiMMiK cutoff? How does it affect accuracy / > performance? Some experimentation is needed here as the optimal value depends on the element types you're using, if anti-aliasing is enabled, and the CPU that you are running on. > I believe block-1d and block-2d are determined by GPU's specification. I > am not very familiar with Cuda. Please someone can elaborate a bit. For > example I am running pyfr with two Tesla k80s in parallel, what's the > block size for 1d and 2d pointswise kernels? You should seldom need to modify either of these two values. On some pathological meshes reducing block-1d can improve performance, but not by a lot. Regards, Freddie. ---
**Author:** ![Junting\_Chen](https://avatars.discourse-cdn.com/v4/letter/j/ac8455/32.png) [@Junting\_Chen](https://pyfr.discourse.group/u/Junting_Chen)\ **Post date:** [7 January 2020 17:40 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/3 "2020-01-07T17:40:47Z")
Thanks Freddie, So when starting a run, do you usually play with the GiMMiK cutoff a bit to find the most optimized value (does it influence the performance significantly / worth the effort of finding the optimized value)? What’s the range of this value? Is a power of 2 (example uses 512) somewhat beneficial? Junting ---
**Author:** ![fdw](https://avatars.discourse-cdn.com/v4/letter/f/a587f6/32.png) [@fdw](https://pyfr.discourse.group/u/fdw)\ **Post date:** [8 January 2020 14:32 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/4 "2020-01-08T14:32:01Z")
Hi Junting, The values I would try are 0 (disables GiMMiK), 512 (the default), and 8192. There is nothing special about the number being a power of two. Regards, Freddie. ---
**Author:** ![WillT](https://avatars.discourse-cdn.com/v4/letter/w/bc79bd/32.png) [@WillT](https://pyfr.discourse.group/u/WillT)\ **Post date:** [23 May 2021 21:50 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/5 "2021-05-23T21:50:04Z")
6 posts were split to a new topic: [Cuda backend error ‘CudaOutOfMemory’ ‘CudaInvaildDevice’](https://pyfr.discourse.group/t/cuda-backend-error-cudaoutofmemory-cudainvailddevice/401) ---
**Author:** ![WillT](https://avatars.discourse-cdn.com/v4/letter/w/bc79bd/32.png) [@WillT](https://pyfr.discourse.group/u/WillT)\ **Post date:** [23 May 2021 21:41 UTC](https://pyfr.discourse.group/t/cuda-backend-parameters/262/11 "2021-05-23T21:41:45Z")