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

Add vcn and jpeg activity

Browse Source 
Changes:
    - Add new engine field vcn_activity (from 1.4/1.5
      gpu_metrics
    - Updated log output to enhance view of gpu_metric
      data as json pretty print
    - Added new fields provided in 1.5
    - Added unit overview in python API, CLI is WIP

Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Change-Id: I7d9f29e7ecc35dcd0697814c222cdd02b0d5518e


[ROCm/amdsmi commit: 8f3861e1d9]
This commit is contained in:
Charis Poag
2023-12-06 03:25:38 -06:00
parent 030a971ce4
commit 4f502e5dab
8 changed files with 229 additions and 116 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/amdsmi/src/amd_smi/amd_smi.cc
+26 -2
View File
@@ -68,6 +68,7 @@
#include "amd_smi/impl/amdgpu_drm.h"
#include "amd_smi/impl/amd_smi_utils.h"
#include "amd_smi/impl/amd_smi_processor.h"
#include "rocm_smi/rocm_smi_logger.h"
#ifdef ENABLE_ESMI_LIB
#include "amd_smi/impl/amd_smi_cpu_socket.h"
#include "amd_smi/impl/amd_smi_cpu_core.h"
@@ -1154,9 +1155,32 @@ amdsmi_status_t amdsmi_get_gpu_metrics_info(
amdsmi_gpu_metrics_t *pgpu_metrics) {
AMDSMI_CHECK_INIT();
// nullptr api supported
return rsmi_wrapper(rsmi_dev_gpu_metrics_info_get, processor_handle,
amdsmi_status_t ret =
rsmi_wrapper(rsmi_dev_gpu_metrics_info_get, processor_handle,
reinterpret_cast<rsmi_gpu_metrics_t*>(pgpu_metrics));
// WARNING: TEMPORARY - awaiting 1.5 update from amdgpu driver/firmware
// intended to be removed later
// START: REMOVE WHATS BELOW ME
uint8_t content_ver = pgpu_metrics->common_header.content_revision;
int8_t format_ver = pgpu_metrics->common_header.format_revision;
const uint8_t expected_format_ver = 1;
const uint8_t expected_content_ver = 4;
if (ret == AMDSMI_STATUS_SUCCESS &&
(format_ver == expected_format_ver &&
content_ver <= expected_content_ver)) {
std::ostringstream ss;
ss << __PRETTY_FUNCTION__ << " | SET JPEG_ACTIVITY to MAX_UINT16, "
<< "detected content version: " << std::dec << +content_ver
<< "; format version: " << std::dec << +format_ver
<< "; awaiting 1.5 metrics remove once released";
LOG_ALWAYS(ss);
std::fill_n(&pgpu_metrics->jpeg_activity[0],
(sizeof(pgpu_metrics->jpeg_activity) /
sizeof(pgpu_metrics->jpeg_activity[0])),
std::numeric_limits<uint16_t>::max());
}
// END: REMOVE WHATS ABOVE ME
return ret;
}
projects/amdsmi/src/amd_smi/amd_smi_utils.cc
+17 -2
View File
@@ -43,6 +43,7 @@
#include "amd_smi/impl/amd_smi_utils.h"
#include "shared_mutex.h" // NOLINT
#include "rocm_smi/rocm_smi_logger.h"
static const uint32_t kAmdGpuId = 0x1002;
@@ -153,6 +154,20 @@ amdsmi_status_t smi_amdgpu_get_board_info(amd::smi::AMDSmiGPUDevice* device, amd
fgets(info->product_name, sizeof(info->product_name), fp);
fclose(fp);
}
std::ostringstream ss;
ss << __PRETTY_FUNCTION__
<< "Returning status = AMDSMI_STATUS_SUCCESS"
<< " | model_number_path = " << model_number_path
<< "; info->model_number: " << info->model_number
<< "\n product_serial_path = " << product_serial_path
<< "; info->product_serial: " << info->product_serial
<< "\n fru_id_path = " << fru_id_path
<< "; info->fru_id: " << info->fru_id
<< "\n manufacturer_name_path = " << manufacturer_name_path
<< "; info->manufacturer_name: " << info->manufacturer_name
<< "\n product_name_path = " << product_name_path
<< "; info->product_name: " << info->product_name;
LOG_INFO(ss);
return AMDSMI_STATUS_SUCCESS;
}
@@ -236,8 +251,8 @@ amdsmi_status_t smi_amdgpu_get_ranges(amd::smi::AMDSmiGPUDevice* device, amdsmi_
unsigned int dpm_level, freq;
char firstChar = line[0];
if (firstChar == 'S'){
if (sscanf(line.c_str(), "%c: %d%s", &single_char, &sleep_freq, str) <= 2){
if (firstChar == 'S') {
if (sscanf(line.c_str(), "%c: %d%s", &single_char, &sleep_freq, str) <= 2) {
ranges.close();
return AMDSMI_STATUS_NO_DATA;
}