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

Removed backwards compatibility for jpeg_activity/vcn_activity (#357)

Browse Source 
Updated:
- Removed backwards compatibility for jpeg_activity/vcn_activity
- On supported ASICs users can use XCP (partition) stat values:
  jpeg_busy and vcn_busy

Signed-off-by: Charis Poag <Charis.Poag@amd.com>
This commit is contained in:
Poag, Charis
2025-05-29 11:58:06 -05:00
committed by GitHub
parent 8724658c14
commit f89a8c895c
2 changed files with 85 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+85
View File
@@ -78,6 +78,91 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
### Removed ### Removed
- **Removed backwards compatibility `amdsmi_get_gpu_metrics_info()`'s `jpeg_activity` or `vcn_activity` fields: use `xcp_stats.jpeg_busy` or `xcp_stats.vcn_busy`**
- Backwards compatibility is removed for `jpeg_activity` and `vcn_activity` fields, if the `jpeg_busy` or `vcn_busy` field is available.
- <i>Reasons for this change</i>:
- Providing both `vcn_activity`/`jpeg_activity` and XCP (partition) stats `vcn_busy`/`jpeg_busy` caused confusion for users about which field to use. By removing backward compatibility, it is easier to identify the relevant field.
- The `jpeg_busy` field increased in size (for supported ASICs), making backward compatibility unable to fully copy the structure into `jpeg_activity`.
See below for comparison of updated CLI outputs:
Original output:
```shell
$ amd-smi metric --usage
GPU: 0
USAGE:
GFX_ACTIVITY: 0 %
UMC_ACTIVITY: 0 %
MM_ACTIVITY: N/A
VCN_ACTIVITY: [0 %, N/A, N/A, N/A]
JPEG_ACTIVITY: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
GFX_BUSY_INST:
XCP_0: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_1: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_2: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_3: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_4: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_5: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_6: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_7: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
JPEG_BUSY:
XCP_0: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_1: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_2: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_3: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_4: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_5: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_6: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_7: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
VCN_BUSY:
XCP_0: [0 %, N/A, N/A, N/A]
XCP_1: [0 %, N/A, N/A, N/A]
XCP_2: [0 %, N/A, N/A, N/A]
XCP_3: [0 %, N/A, N/A, N/A]
XCP_4: [0 %, N/A, N/A, N/A]
XCP_5: [0 %, N/A, N/A, N/A]
XCP_6: [0 %, N/A, N/A, N/A]
XCP_7: [0 %, N/A, N/A, N/A]
```
New output:
```shell
$ amd-smi metric --usage
GPU: 0
USAGE:
GFX_ACTIVITY: 0 %
UMC_ACTIVITY: 0 %
MM_ACTIVITY: N/A
VCN_ACTIVITY: [N/A, N/A, N/A, N/A]
JPEG_ACTIVITY: [N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
GFX_BUSY_INST:
XCP_0: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_1: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_2: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_3: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_4: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_5: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_6: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_7: [0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
JPEG_BUSY:
XCP_0: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_1: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_2: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_3: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_4: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_5: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_6: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
XCP_7: [0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, 0 %, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A, N/A]
VCN_BUSY:
XCP_0: [0 %, N/A, N/A, N/A]
XCP_1: [0 %, N/A, N/A, N/A]
XCP_2: [0 %, N/A, N/A, N/A]
XCP_3: [0 %, N/A, N/A, N/A]
XCP_4: [0 %, N/A, N/A, N/A]
XCP_5: [0 %, N/A, N/A, N/A]
XCP_6: [0 %, N/A, N/A, N/A]
XCP_7: [0 %, N/A, N/A, N/A]
```
### Optimized ### Optimized
### Resolved issues ### Resolved issues
rocm_smi/src/rocm_smi_gpu_metrics.cc
-42
View File
@@ -2040,20 +2040,6 @@ AMGpuMetricsPublicLatestTupl_t GpuMetricsBase_v18_t::copy_internal_to_external_m
metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1]; metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1];
// separate by XCP
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::begin(metrics_public_init.vcn_activity));
}
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::begin(metrics_public_init.jpeg_activity));
}
return metrics_public_init; return metrics_public_init;
}(); }();
@@ -2258,20 +2244,6 @@ AMGpuMetricsPublicLatestTupl_t GpuMetricsBase_v17_t::copy_internal_to_external_m
metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1]; metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1];
// separate by XCP
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::begin(metrics_public_init.vcn_activity));
}
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::begin(metrics_public_init.jpeg_activity));
}
return metrics_public_init; return metrics_public_init;
}(); }();
@@ -2464,20 +2436,6 @@ AMGpuMetricsPublicLatestTupl_t GpuMetricsBase_v16_t::copy_internal_to_external_m
metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1]; metrics_public_init.current_dclk1 = metrics_public_init.current_dclk0s[1];
// separate by XCP
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].vcn_busy),
std::begin(metrics_public_init.vcn_activity));
}
if (this->m_partition_id < kRSMI_MAX_NUM_XCP
&& m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy[0] != UINT16_MAX) {
std::copy(std::begin(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::end(m_gpu_metrics_tbl.m_xcp_stats[this->m_partition_id].jpeg_busy),
std::begin(metrics_public_init.jpeg_activity));
}
return metrics_public_init; return metrics_public_init;
}(); }();