- Need to check the availability of `__has_attribute` builtin macro
instead of compiler versions. That's more reliable and portable among
various compilers.
- Provides a very basic support of vectors for unknown compilers.
[ROCm/clr commit: fc223dc584]
- Need to check the availability of `__has_attribute` builtin macro
instead of compiler versions. That's more reliable and portable among
various compilers.
- Provides a very basic support of vectors for unknown compilers.
- Need to check the availability of `__has_attribute` builtin macro
instead of compiler versions. That's more reliable and portable among
various compilers.
- Provides a very basic support of vectors for unknown compilers.
[ROCm/hip commit: 774035d869]
- Need to check the availability of `__has_attribute` builtin macro
instead of compiler versions. That's more reliable and portable among
various compilers.
- Provides a very basic support of vectors for unknown compilers.
Current NCCL code does not abort for failed Flush operations by
underlying network. This may compromise data integrity.
Signed-off-by: Rashika Kheria <rashika@amazon.com>
[ROCm/rccl commit: 6c61492eba]
Current NCCL code does not abort for failed Flush operations by
underlying network. This may compromise data integrity.
Signed-off-by: Rashika Kheria <rashika@amazon.com>
For gfx908, support simple detection of ring topology.
Call ReduceOrCopyMulti directly from kernel.
Also simplify code by removing kernel start synchronization option
which has no effect on throughput measurements.
[ROCm/rccl commit: ebc823e603]
For gfx908, support simple detection of ring topology.
Call ReduceOrCopyMulti directly from kernel.
Also simplify code by removing kernel start synchronization option
which has no effect on throughput measurements.
This solution takes into account that some hwmons use
label files to map sensor types. The previous solution
did not take this into account.
Change-Id: I1d6204573cefa8197b2cfe0ffb412b545df3d80a
[ROCm/rocm_smi_lib commit: 324c0ca0e5]
This solution takes into account that some hwmons use
label files to map sensor types. The previous solution
did not take this into account.
Change-Id: I1d6204573cefa8197b2cfe0ffb412b545df3d80a
This solution takes into account that some hwmons use
label files to map sensor types. The previous solution
did not take this into account.
Change-Id: I1d6204573cefa8197b2cfe0ffb412b545df3d80a
[ROCm/amdsmi commit: 324c0ca0e5]
There are now two implementations of printf in HIP:
1. The implemenation for HCC is controlled by the HC_FEATURE_PRINTF
macro, and it works only with the HCC compiler used in combination
with the HCC runtime.
2. The implementation for hip-clang requires the VDI runtime, and is
always enabled with that combination.
Change-Id: Ibaeda7900ffe2ce602ca0094aafed0f1147ac2b6
[ROCm/clr commit: d48738856c]
There are now two implementations of printf in HIP:
1. The implemenation for HCC is controlled by the HC_FEATURE_PRINTF
macro, and it works only with the HCC compiler used in combination
with the HCC runtime.
2. The implementation for hip-clang requires the VDI runtime, and is
always enabled with that combination.
Change-Id: Ibaeda7900ffe2ce602ca0094aafed0f1147ac2b6
There are now two implementations of printf in HIP:
1. The implemenation for HCC is controlled by the HC_FEATURE_PRINTF
macro, and it works only with the HCC compiler used in combination
with the HCC runtime.
2. The implementation for hip-clang requires the VDI runtime, and is
always enabled with that combination.
Change-Id: Ibaeda7900ffe2ce602ca0094aafed0f1147ac2b6
[ROCm/hip commit: 64cd527335]
There are now two implementations of printf in HIP:
1. The implemenation for HCC is controlled by the HC_FEATURE_PRINTF
macro, and it works only with the HCC compiler used in combination
with the HCC runtime.
2. The implementation for hip-clang requires the VDI runtime, and is
always enabled with that combination.
Change-Id: Ibaeda7900ffe2ce602ca0094aafed0f1147ac2b6
Test window will quit or show error message when run OpenCL API.
Add a workaround for the race condition with the first page
during pinning.
Change-Id: I9a27b4e173cf94c84aefcb94e255f11169453d94
[ROCm/clr commit: ab85674f8a]
Test window will quit or show error message when run OpenCL API.
Add a workaround for the race condition with the first page
during pinning.
Change-Id: I9a27b4e173cf94c84aefcb94e255f11169453d94
Also fix potential issue with evaluating functionality of
functions with multiple sub-variants.
Change-Id: I9a09e52f3d3f3181e72578ed1f3bfd0d85516aa3
[ROCm/rocm_smi_lib commit: 1d8e16bff2]
Also fix potential issue with evaluating functionality of
functions with multiple sub-variants.
Change-Id: I9a09e52f3d3f3181e72578ed1f3bfd0d85516aa3
[ROCm/amdsmi commit: 1d8e16bff2]
mhbliao
Jatin
Rashika Kheria
Wenkai Du
Chris Freehill
Sameer Sahasrabuddhe
Vladislav Sytchenko
Evgeny
rkebichi
Chauncey Hui
Jason Tang
Saleel Kudchadker
German Andryeyev
Jiabao Xie