AI/rocm-systems
2
0
Fork 0
程式碼 問題 合併請求 Actions 套件 專案 版本發布 Wiki 動態

[SWDEV-484382] Added fclk and socclk to amd-smi metric -c

瀏覽代碼 
Signed-off-by: gabrpham <Gabriel.Pham@amd.com>
Change-Id: Ie7e19c757b05455693c0d26eeb5e8b6c1e238375
此提交包含在：
gabrpham
2024-12-04 13:54:20 -06:00
提交者 Gabriel Pham
父節點 5f9c2db6f3
當前提交 fe290a2056
共有 4 個檔案被更改，包括 101 行新增 和 1 行删除
下載 Patch 檔 下載差異檔 展開所有檔案 折疊所有檔案
CHANGELOG.md
+21 -1
查看文件
@@ -7,8 +7,28 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
### Added
- **Added fclk and socclk info to `amd-smi metric -c/--clock`**.
fclk and socclk information such as min and max clock have been added to the metric command, in line with all the other clocks.
```shell
amd-smi metric -c -g 1
...
FCLK_0:
CLK: 2301 MHz
MIN_CLK: 601 MHz
MAX_CLK: 2301 MHz
CLK_LOCKED: N/A
DEEP_SLEEP: DISABLED
SOCCLK_0:
CLK: 1500 MHz
MIN_CLK: 500 MHz
MAX_CLK: 1500 MHz
CLK_LOCKED: N/A
DEEP_SLEEP: DISABLED
```
- **Added new command `amd-smi set -c/--clock-level`**.
This new command sets the performance level of the selected clock on the desired GPUs. The command can accept a range of acceptable levels, but will not set the level when a level is beyond the number of frequency levels as show in `amd-smi static -C/--clock`
This new command sets the performance level of the selected clock on the desired GPUs. The command can accept a range of acceptable levels, but will not set the level when a level is beyond the number of frequency levels as show in `amd-smi static -C/--clock`.
```shell
sudo amd-smi set -c sclk 5 6
amdsmi_cli/amdsmi_commands.py
+67
查看文件
@@ -1682,6 +1682,18 @@ class AMDSMICommands():
"clk_locked" : "N/A",
"deep_sleep" : "N/A"}
clocks["fclk_0"] = {"clk" : "N/A",
"min_clk" : "N/A",
"max_clk" : "N/A",
"clk_locked" : "N/A",
"deep_sleep" : "N/A"}
clocks["socclk_0"] = {"clk" : "N/A",
"min_clk" : "N/A",
"max_clk" : "N/A",
"clk_locked" : "N/A",
"deep_sleep" : "N/A"}
clock_unit = "MHz"
# TODO make the deepsleep threshold correspond to the * in sysfs for current deep sleep status
deep_sleep_threshold = 140
@@ -1759,6 +1771,31 @@ class AMDSMICommands():
clocks[dclk_index]["deep_sleep"] = "ENABLED"
else:
clocks[dclk_index]["deep_sleep"] = "DISABLED"
# Populate FCLK clock value; fclk not present in gpu_metrics so use amdsmi_get_clk_freq
frequency_dict = amdsmi_interface.amdsmi_get_clk_freq(args.gpu, amdsmi_interface.AmdSmiClkType.DF)
current_fclk_clock = frequency_dict['frequency'][frequency_dict['current']]
current_fclk_clock = self.helpers.convert_SI_unit(current_fclk_clock, self.helpers.SI_Unit.MICRO)
clocks["fclk_0"]["clk"] = self.helpers.unit_format(self.logger,
current_fclk_clock,
clock_unit)
if int(current_fclk_clock) <= deep_sleep_threshold:
clocks["fclk_0"]["deep_sleep"] = "ENABLED"
else:
clocks["fclk_0"]["deep_sleep"] = "DISABLED"
# Populate SOCCLK clock value
current_socclk_clock = gpu_metric["current_socclk"]
clocks["socclk_0"]["clk"] = self.helpers.unit_format(self.logger,
current_socclk_clock,
clock_unit)
if int(current_socclk_clock) <= deep_sleep_threshold:
clocks["socclk_0"]["deep_sleep"] = "ENABLED"
else:
clocks["socclk_0"]["deep_sleep"] = "DISABLED"
except Exception as e:
logging.debug("Failed to get gpu_metrics_info for gpu %s | %s", gpu_id, e)
@@ -1831,6 +1868,36 @@ class AMDSMICommands():
except amdsmi_exception.AmdSmiLibraryException as e:
logging.debug("Failed to get vclk and/or dclk clock info for gpu %s | %s", gpu_id, e.get_error_info())
# FCLK min and max clocks
try:
fclk_clk_info_dict = amdsmi_interface.amdsmi_get_clock_info(args.gpu,
amdsmi_interface.AmdSmiClkType.DF)
# if the current clock is N/A then we shouldn't populate the max and min values
if clocks["fclk_0"]["clk"] != "N/A":
clocks["fclk_0"]["min_clk"] = self.helpers.unit_format(self.logger,
fclk_clk_info_dict["min_clk"],
clock_unit)
clocks["fclk_0"]["max_clk"] = self.helpers.unit_format(self.logger,
fclk_clk_info_dict["max_clk"],
clock_unit)
except amdsmi_exception.AmdSmiLibraryException as e:
logging.debug("Failed to get fclk info for gpu %s | %s", gpu_id, e.get_error_info())
# SOCCLK min and max clocks
try:
socclk_clk_info_dict = amdsmi_interface.amdsmi_get_clock_info(args.gpu,
amdsmi_interface.AmdSmiClkType.SOC)
# if the current clock is N/A then we shouldn't populate the max and min values
if clocks["socclk_0"]["clk"] != "N/A":
clocks["socclk_0"]["min_clk"] = self.helpers.unit_format(self.logger,
socclk_clk_info_dict["min_clk"],
clock_unit)
clocks["socclk_0"]["max_clk"] = self.helpers.unit_format(self.logger,
socclk_clk_info_dict["max_clk"],
clock_unit)
except amdsmi_exception.AmdSmiLibraryException as e:
logging.debug("Failed to get socclk info for gpu %s | %s", gpu_id, e.get_error_info())
values_dict['clock'] = clocks
if "temperature" in current_platform_args:
if args.temperature:
src/amd_smi/amd_smi.cc
+7
查看文件
@@ -2217,6 +2217,13 @@ amdsmi_get_clock_info(amdsmi_processor_handle processor_handle, amdsmi_clk_type_
case AMDSMI_CLK_TYPE_DCLK1:
info->clk = metrics.current_dclk1;
break;
case AMDSMI_CLK_TYPE_SOC:
info->clk = metrics.current_socclk;
break;
// fclk/df not supported by gpu metrics so providing default value which cannot be contrued to be valid
case AMDSMI_CLK_TYPE_DF:
info->clk = UINT32_MAX;
break;
default:
return AMDSMI_STATUS_INVAL;
}
src/amd_smi/amd_smi_utils.cc
+6
查看文件
@@ -275,6 +275,12 @@ amdsmi_status_t smi_amdgpu_get_ranges(amd::smi::AMDSmiGPUDevice* device, amdsmi_
case AMDSMI_CLK_TYPE_DCLK1:
fullpath += "/pp_dpm_dclk1";
break;
case AMDSMI_CLK_TYPE_SOC:
fullpath += "/pp_dpm_socclk";
break;
case AMDSMI_CLK_TYPE_DF:
fullpath += "/pp_dpm_fclk";
break;
default:
return AMDSMI_STATUS_INVAL;
}