AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Updated 6.4.2 Changelog

Browse Source 
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Change-Id: I975f5db0bde9ebccec3756415cb1e7dc47e78988
This commit is contained in:
Maisam Arif
2025-06-12 10:26:38 -05:00
committed by Arif, Maisam
parent 6da33b8ded
commit 772b572913
1 changed files with 15 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+15 -18
View File
@@ -95,8 +95,6 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
### Changed
- **Padded `asic_serial` in `amdsmi_get_asic_info` with 0s**.
- **Updated `amdsmi_get_clock_info` in `amdsmi_interface.py`**.
- The `clk_deep_sleep` field now returns the sleep integer value.
@@ -294,6 +292,7 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
- **Added Compute Unit Occupancy information per process**
Measuring compute units are the best way currently to determine gfx usage on a per process basis
- Added `cu_occupancy` field to `amdsmi_proc_info_t` structure in C & Python APIs, in minor version update
- Added `CU_OCCUPANCY` to `amd-smi process` output.
- Added `CU%` to `amd-smi monitor -q`
@@ -310,25 +309,19 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
- **Added new firmware PLDM_BUNDLE**.
- `amd-smi firmware` can now show the PLDM Bundle on supported systems.
- **Added `amd-smi ras --afid --cper-file <file_path>` to decode CPER records**
- Python and C have added the `amdsmi_get_afids_from_cper()` to decode
### Changed
- N/A
- **Padded `asic_serial` in `amdsmi_get_asic_info` with 0s**.
### Removed
- **Renamed `amd-smi --partition`'s field `COMPUTE_PARTITION` to `ACCELERATOR_PARTITION`**.
We are changing the field named `COMPUTE_PARTITION` to `ACCELERATOR_PARTITION`. Refer to example outputs provided below.
API and associated struct naming will remain the same:
- `amdsmi_status_t amdsmi_get_gpu_compute_partition(amdsmi_processor_handle processor_handle, char *compute_partition, uint32_t len)`
- `amdsmi_status_t amdsmi_set_gpu_compute_partition(amdsmi_processor_handle processor_handle, amdsmi_compute_partition_type_t compute_partition)`
- `amdsmi_status_t amdsmi_get_gpu_accelerator_partition_profile_config(amdsmi_processor_handle processor_handle, amdsmi_accelerator_partition_profile_config_t *profile_config)`
- `amdsmi_status_t amdsmi_get_gpu_accelerator_partition_profile(amdsmi_processor_handle processor_handle, amdsmi_accelerator_partition_profile_t *profile, uint32_t *partition_id)`
- `amdsmi_status_t amdsmi_set_gpu_accelerator_partition_profile(amdsmi_processor_handle processor_handle, uint32_t profile_index)`
Reason(s) for this change:
- Align with host AMD SMI's `static --partition` field naming
- Align with naming seen in `amd-smi partition`
- **Renamed fields `COMPUTE_PARTITION` to `ACCELERATOR_PARTITION` in CLI call `amd-smi --partition`**.
- We are changing the field named `COMPUTE_PARTITION` to `ACCELERATOR_PARTITION`.
- API and associated struct naming will remain the same
- Reason(s) for this change:
- Align with host AMD SMI's `static --partition` field naming
- Align with naming seen in `amd-smi partition`
*Previous Output:*
@@ -352,6 +345,10 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
PARTITION_ID: 0
```
### Removed
- N/A
### Optimized
- N/A