Block or Report
Block or report pfultz2
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePopular repositories
1,096 contributions in the last year
Less
More
Contribution activity
September 2022
Created 23 commits in 3 repositories
Created a pull request in ROCmSoftwarePlatform/AMDMIGraphX that received 5 comments
Opened 22 other pull requests in 3 repositories
danmar/cppcheck
12
merged
1
open
- Fix 11298: FP danglingTemporaryLifetime when constructing from const char*
- Fix 11308: FP accessMoved with virtual function
- Fix 11306: FP knownConditionTrueFalse with strlen()
- Fix 9755: false negative: access of moved variable in conditional code
- Set valueflow for container elements
- Partial fix 11154: FN: knownConditionTrueFalse
- Fix 11194: FN: knownConditionTrueFalse
- Fix 11285: FN: zerodiv
- Fix 6370: ValueFlow: array element with known value
- Use return instead of condition for alwaysTrueFalse check
- Fix 11296: FN: knownConditionTrueFalse ( a !=0 && b != 0 && a == 0)
- Fix 11295: Regression FP: knownConditionTrueFalse
- Refactor knownConditionTrueFalse check and isUsedAsBool function
ROCmSoftwarePlatform/AMDMIGraphX
8
open
- Remove unused device functions
- Add nhwc layout to gpu backend
- Fix invalid program in debug mode from find_splits
- Use larger vector size instead of preloading for broadcasted inputs
- Concat pointwise fusions
- Refactor memory coloring pass
- Add gpu debug builds
- Add MIGRAPHX_VALIDATE_MATCHES env variable to validate each matcher
ROCmSoftwarePlatform/MIOpen
1
open
Reviewed 19 pull requests in 3 repositories
ROCmSoftwarePlatform/AMDMIGraphX
15 pull requests
- Refactor memory coloring pass
- Add output_alias and runs_on_offload_target flags for the custom ops
- Parameterize epsilon for layernorm kernel
- Multibroadcast find_mul_conv
- Fix typo for add_sigmoid
- Use find_2.0 API for the convolution
- Update deprecated Pybind constructor
-
[mlir] Replaced
find_librarywithfind_packageto locate MLIR static library - Improve layernorm and reductions performance
- Reduce problem size of unbatched_gemm tests
- Fix split_reshape for slice len of 1
- expose underlying migraphx::argument data pointer in pybind
- Stream sync Changset
- Implement concat using jit compilation
- Use rocblas_gemm_ex for batched gemms with broadcasted B






