-
### Suggestion Description
hi, rocm expert, I am trying to build a rocm lib as an python package with setuptools. I tried a few ways, neither works, can someone give a hint ?
```py
ext_module…
-
Hi here, i tryed to test CogVideoX with an RTX 3090. But when i launch i got this error message :
`RuntimeError: torch._scaled_mm is only supported on CUDA devices with compute capability >= 9.0 or…
-
### Problem Description
hi, rocm expert, when building a rocm /torch project with following link flags:
```py
torch_libs = ["torch", "torch_hip", "torch_python", "c10", "c10_hip"]
torch_link_l…
-
## Bug Report
Build error in a Cuda UVM build:
https://sems-cdash-son.sandia.gov/cdash/viewBuildError.php?buildid=235465
```
/builds/muelu/nightly-testing/Trilinos/packages/tpetra/core/src/Tpetr…
-
This is an umbrella issue for allowing fp8 type(s) in shark, spanning all the required layers of the stack: Turbine, IREE, MLIR, LLVM, including backends of interest like ROCm.
Some initial researc…
kuhar updated
9 months ago
-
In a recent PR, the TopK e2e test fails in CI: https://github.com/iree-org/iree/actions/runs/11107992173/job/30867743807?pr=18634
The following test is what fails:
```
func.func @topk_2d_dim1_inv…
-
### Idea
Use int4 as the compression technique to fit larger models onto Navi machines or possibly MI series machines. Weights would be compressed using encoding scheme that would pack two 4 bits n…
-
hipBLASLt error: heuristic fetch failed.
-
### Problem Description
Running CTest fails with GPU Hang videoDecodeBatch sample.
Running Perf sample with t > 1 also fails with GPU Hang
dmesg log shows:
```
amdgpu 0000:65:00.0: amdgpu: ri…
-
An increasing number of genomics tasks require the presence of additional resource types such as GPU, FPGA and the option of using `aarch64` architecture. If there is not already a way to add these, a…