AMGX
AMGX copied to clipboard
NVIDIA
Distributed multigrid linear solver library on GPU
Can you please let us know how cusparse API are getting leveraged by AMGx. Specifically what configurations (condition/criteria) needs to be met in the code flow for AMGx to use...
Hello, I tried to install AMGX latest version with the flag `WITH_NO_MPI=True`, but it kept on searching for MPI and showing an error after not finding it. In v.2.3.0 this...
AMGx currently supports Nvidia GPUs through CUDA. We want to contribute a new backend through Intel's oneAPI [SYCL](https://www.intel.com/content/www/us/en/develop/documentation/oneapi-programming-guide/top/oneapi-programming-model.html). This would enable support for Nvidia, AMD and Intel devices with a...
Nobody would ever spot it, but I have luck: it creates a strange compile error on nvcc with -ccbin=clang: ``` /usr/lib/llvm-14/lib/clang/14.0.0/include/omp.h(496): error: function omp_is_initial_device has already been defined ``` Anyway,...
By some reason, SM90-related constructs in `cuda_fp16.hpp` make the compilation to fail for me: ``` /usr/local/cuda/bin/../targets/x86_64-linux/include/cuda_fp16.hpp(1337): error: type name is not allowed unsigned int tmp; asm("cvt.rzi.s8.f16 %0, %1;" : "=r"(tmp)...
**Describe the issue** Dear AMGX's developers, I'm trying to build AMGX with gcc in windows, but got error in cmake. The follows are my environment information. **Environment information:** - OS:...
Re-instated the eigensolver functionality as first class. Moved the configuration files to the examples directory.
Problem: -------- ```c include/amgx_config.h:128:42: error: arithmetic between different enumeration types ‘AMGX_MemorySpace’ and ‘AMGX_VecPrecision’ is deprecated [-Werror=deprecated-enum-enum-conversion] 128 | AMGX_modeRange = AMGX_memorySpaceNum * AMGX_vecPrecisionNum * AMGX_matPrecisionNum * AMGX_indPrecisionNum, | ~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~ ```...
Hello, I'm working on solving a matrix equation that seems to be ill-conditioned, and I'm encountering convergence issues with all the configuration files provided in AMGX. I've attached a JPEG...
