Yu-Hsiang M. Tsai

@qir-at-tum thanks for your report. I guess it is something from bfloat16. could you turn off it by `-DGINKGO_ENABLE_BFLOAT16=OFF`

no, if you do not need bfloat16, there will no negative impact. Maybe cuda 11.6 does not fully support all ptx for bfloat16 yet. We will investigate it further.

Hi @VictorEijkhout Thanks for reporting. Could you give more detail of your compiler setup and CMake Version? The followings may help - `ldd lib/libginkgo_omp.so` check whether it can find the...

Hi @VictorEijkhout I guess these issues are related to CMake 4.1.0 with `CLICOLOR_FORCE=1` Could you check whether you have set this variable? Related Issue in CMake https://gitlab.kitware.com/cmake/cmake/-/issues/27137, https://gitlab.kitware.com/cmake/cmake/-/issues/27155. It is...

we also test the benchmark output. you need to modify the `*solver*.stdout` in benchmark/test/reference. it should be mostly `"apply": { "components": { * } }` to `"apply": { "components": {...

Hi @nbeams During the discussion in #1939 , we realize the diagonal matrix must be sorted in CSR because it only has one entry per row. Do you mean some...

Hi @emre-safa , do you have some time to work on this?