Merge amd-dev into amd-master 20241022
Change-Id: I76f797f267679882d02106a4177359bd31e0b0d5
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
[ROCm/amdsmi commit: 148b015cab]
Этот коммит содержится в:
Charis Poag
@@ -568,6 +568,10 @@ GPU: 0
|
||||
|
||||
### Removals
|
||||
|
||||
- **Removed `amd-smi reset --compute-partition` and `... --memory-partition` and associated APIs**.
|
||||
- This change is part of the partition redesign. Reset functionality will be reintroduced in a later update.
|
||||
- associated APIs include `amdsmi_reset_gpu_compute_partition()` and `amdsmi_reset_gpu_memory_partition()`
|
||||
|
||||
- **Removed usage of _validate_positive in Parser and replaced with _positive_int and _not_negative_int as appropriate**.
|
||||
- This will allow 0 to be a valid input for several options in setting CPUs where appropriate (for example, as a mode or NBIOID)
|
||||
|
||||
@@ -649,6 +653,33 @@ GPU POWER GPU_TEMP MEM_TEMP VRAM_USED VRAM_TOTAL
|
||||
|
||||
- **Python API for `amdsmi_get_energy_count()` will deprecate the `power` field in ROCm 6.4 and use `energy_accumulator` field instead**.
|
||||
|
||||
- **New memory and compute partition APIs incoming for ROCm 6.4**.
|
||||
- These APIs will be updated to fully populate the CLI and allowing compute (accelerator) partitions to be set by profile ID.
|
||||
- One API will be provided, to reset both memory and compute (accelerator).
|
||||
- There are dependencies regarding available compute partitions when in other memory modes.
|
||||
- Driver will be providing these default modes
|
||||
- Memory partition resets (for BM) require driver reloads - this will allow us to notify users before taking this action, then change to the default compute partition modes.
|
||||
- The following APIs will remain:
|
||||
|
||||
```C
|
||||
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_compute_partition(amdsmi_processor_handle processor_handle,
|
||||
char *compute_partition, uint32_t len);
|
||||
amdsmi_status_t
|
||||
amdsmi_get_gpu_memory_partition(amdsmi_processor_handle processor_handle,
|
||||
|
||||
char *memory_partition, uint32_t len);
|
||||
amdsmi_status_t
|
||||
amdsmi_set_gpu_memory_partition(amdsmi_processor_handle processor_handle,
|
||||
amdsmi_memory_partition_type_t memory_partition);
|
||||
```
|
||||
|
||||
- **amd-smi set --compute-partition "SPX/DPX/CPX..." will no longer be supported for ROCm 6.4**.
|
||||
- This is due to aligning with Host setups and providing more robust partition information through the APIs outlined above. Furthermore, new APIs which will be available on both BM/Host can set by profile ID. (functionality coming soon!)
|
||||
|
||||
- **Added preliminary `amd-smi partition` command**.
|
||||
- The new partition command can be used to display GPU information, including memory and accelerator partition information.
|
||||
- The command will be at full functionality once additional partition information from `amdsmi_get_gpu_accelerator_partition_profile()` has been implemented.
|
||||
|
||||
@@ -134,10 +134,10 @@ do_install_amdsmi_python_lib() {
|
||||
export PIP_BREAK_SYSTEM_PACKAGES=1
|
||||
|
||||
# Remove old python library
|
||||
local pip_list_output
|
||||
pip_list_output=$(python3 -m pip list --format=columns --disable-pip-version-check)
|
||||
local amdsmi_pip_list_output
|
||||
amdsmi_pip_list_output=$(python3 -m pip list --format=columns --disable-pip-version-check)
|
||||
# check pip list output for amdsmi
|
||||
if [[ $pip_list_output == *"amdsmi"* ]]; then
|
||||
if [[ $amdsmi_pip_list_output == *"amdsmi"* ]]; then
|
||||
echo "Detected old AMD-SMI python library (amdsmi)..."
|
||||
python3 -m pip uninstall amdsmi --yes --quiet --disable-pip-version-check
|
||||
echo "Removed old AMD-SMI python library (amdsmi)..."
|
||||
@@ -148,9 +148,32 @@ do_install_amdsmi_python_lib() {
|
||||
return
|
||||
fi
|
||||
|
||||
# install python library at @CPACK_PACKAGING_INSTALL_PREFIX@/@SHARE_INSTALL_PREFIX@/amdsmi
|
||||
local python_lib_path=@CPACK_PACKAGING_INSTALL_PREFIX@/@SHARE_INSTALL_PREFIX@
|
||||
python3 -m pip install "$python_lib_path" --quiet --disable-pip-version-check --no-build-isolation --no-index
|
||||
check_and_install_amdsmi() {
|
||||
local setuptools_version
|
||||
setuptools_version=$(python3 -c 'import setuptools; print(setuptools.__version__)')
|
||||
if [ $? -ne 0 ]; then
|
||||
echo "[WARNING] Could not determine setuptools version. "\
|
||||
"AMD-SMI python library will not be installed."
|
||||
return
|
||||
fi
|
||||
|
||||
local amdsmi_python_lib_path="/opt/rocm/share/amd_smi"
|
||||
local amdsmi_setup_py_path="/opt/rocm/share/amd_smi/setup.py"
|
||||
|
||||
# Decide installation method based on setuptools version
|
||||
if [[ "$(printf '%s\n' "$setuptools_version" "28.5" | sort -V | head -n1)" == "$setuptools_version" ]]; then
|
||||
echo "[WARNING] Setuptools version is less than 28.5. AMD-SMI will not be installed."
|
||||
elif [[ "$(printf '%s\n' "$setuptools_version" "41.0.1" | sort -V | head -n1)" != "41.0.1" ]]; then
|
||||
echo "Using setup.py for installation due to setuptools version $setuptools_version"
|
||||
python3 "$amdsmi_setup_py_path" install
|
||||
else
|
||||
echo "Using pyproject.toml for installation due to setuptools version $setuptools_version"
|
||||
python3 -m pip install "$amdsmi_python_lib_path" --quiet --disable-pip-version-check --no-build-isolation --no-index
|
||||
fi
|
||||
}
|
||||
|
||||
# Call the function
|
||||
check_and_install_amdsmi
|
||||
|
||||
export PIP_ROOT_USER_ACTION="$PREVIOUS_PIP_ROOT_USER_ACTION"
|
||||
export PIP_BREAK_SYSTEM_PACKAGES="$PREVIOUS_PIP_BREAK_SYSTEM_PACKAGES"
|
||||
|
||||
@@ -134,10 +134,10 @@ do_install_amdsmi_python_lib() {
|
||||
|
||||
|
||||
# Remove old python library
|
||||
local pip_list_output
|
||||
pip_list_output=$(python3 -m pip list --format=columns --disable-pip-version-check)
|
||||
local amdsmi_pip_list_output
|
||||
amdsmi_pip_list_output=$(python3 -m pip list --format=columns --disable-pip-version-check)
|
||||
# check pip list output for amdsmi
|
||||
if [[ $pip_list_output == *"amdsmi"* ]]; then
|
||||
if [[ $amdsmi_pip_list_output == *"amdsmi"* ]]; then
|
||||
echo "Detected old AMD-SMI python library (amdsmi)..."
|
||||
python3 -m pip uninstall amdsmi --yes --quiet --disable-pip-version-check
|
||||
echo "Removed old AMD-SMI python library (amdsmi)..."
|
||||
@@ -148,9 +148,33 @@ do_install_amdsmi_python_lib() {
|
||||
return
|
||||
fi
|
||||
|
||||
# install python library at @CPACK_PACKAGING_INSTALL_PREFIX@/@SHARE_INSTALL_PREFIX@/amdsmi
|
||||
local python_lib_path=@CPACK_PACKAGING_INSTALL_PREFIX@/@SHARE_INSTALL_PREFIX@
|
||||
python3 -m pip install "$python_lib_path" --quiet --disable-pip-version-check --no-build-isolation --no-index
|
||||
check_and_install_amdsmi() {
|
||||
local setuptools_version
|
||||
setuptools_version=$(python3 -c 'import setuptools; print(setuptools.__version__)')
|
||||
if [ $? -ne 0 ]; then
|
||||
echo "[WARNING] Could not determine setuptools version. "\
|
||||
"AMD-SMI python library will not be installed."
|
||||
return
|
||||
fi
|
||||
|
||||
local amdsmi_python_lib_path="/opt/rocm/share/amd_smi"
|
||||
local amdsmi_setup_py_path="/opt/rocm/share/amd_smi/setup.py"
|
||||
|
||||
# Decide installation method based on setuptools version
|
||||
if [[ "$(printf '%s\n' "$setuptools_version" "28.5" | sort -V | head -n1)" == "$setuptools_version" ]]; then
|
||||
echo "[WARNING] Setuptools version is less than 28.5. AMD-SMI will not be installed."
|
||||
elif [[ "$(printf '%s\n' "$setuptools_version" "41.0.1" | sort -V | head -n1)" != "41.0.1" ]]; then
|
||||
echo "Using setup.py for installation due to setuptools version $setuptools_version"
|
||||
python3 "$amdsmi_setup_py_path" install
|
||||
else
|
||||
echo "Using pyproject.toml for installation due to setuptools version $setuptools_version"
|
||||
python3 -m pip install "$amdsmi_python_lib_path" --quiet --disable-pip-version-check --no-build-isolation --no-index
|
||||
fi
|
||||
}
|
||||
|
||||
# Call the function
|
||||
check_and_install_amdsmi
|
||||
|
||||
export PIP_ROOT_USER_ACTION="$PREVIOUS_PIP_ROOT_USER_ACTION"
|
||||
export PIP_BREAK_SYSTEM_PACKAGES="$PREVIOUS_PIP_BREAK_SYSTEM_PACKAGES"
|
||||
|
||||
|
||||
@@ -4210,7 +4210,7 @@ class AMDSMICommands():
|
||||
|
||||
def reset(self, args, multiple_devices=False, gpu=None, gpureset=None,
|
||||
clocks=None, fans=None, profile=None, xgmierr=None, perf_determinism=None,
|
||||
compute_partition=None, memory_partition=None, power_cap=None, clean_local_data=None):
|
||||
power_cap=None, clean_local_data=None):
|
||||
"""Issue reset commands to target gpu(s)
|
||||
|
||||
Args:
|
||||
@@ -4223,8 +4223,6 @@ class AMDSMICommands():
|
||||
profile (bool, optional): Value override for args.profile. Defaults to None.
|
||||
xgmierr (bool, optional): Value override for args.xgmierr. Defaults to None.
|
||||
perf_determinism (bool, optional): Value override for args.perf_determinism. Defaults to None.
|
||||
compute_partition (bool, optional): Value override for args.compute_partition. Defaults to None.
|
||||
memory_partition (bool, optional): Value override for args.memory_partition. Defaults to None.
|
||||
power_cap (bool, optional): Value override for args.power_cap. Defaults to None.
|
||||
clean_local_data (bool, optional): Value override for args.run_cleaner_shader. Defaults to None.
|
||||
|
||||
@@ -4250,10 +4248,6 @@ class AMDSMICommands():
|
||||
args.xgmierr = xgmierr
|
||||
if perf_determinism:
|
||||
args.perf_determinism = perf_determinism
|
||||
if compute_partition:
|
||||
args.compute_partition = compute_partition
|
||||
if memory_partition:
|
||||
args.memory_partition = memory_partition
|
||||
if power_cap:
|
||||
args.power_cap = power_cap
|
||||
if clean_local_data:
|
||||
@@ -4355,67 +4349,48 @@ class AMDSMICommands():
|
||||
reset_profile_results['power_profile'] = "N/A"
|
||||
logging.debug("Failed to reset power profile on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
|
||||
try:
|
||||
level_auto = amdsmi_interface.AmdSmiDevPerfLevel.AUTO
|
||||
amdsmi_interface.amdsmi_set_gpu_perf_level(args.gpu, level_auto)
|
||||
reset_profile_results['performance_level'] = 'Successfully reset Performance Level'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
reset_profile_results['performance_level'] = "N/A"
|
||||
logging.debug("Failed to reset perf level on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_profile', reset_profile_results)
|
||||
if args.xgmierr:
|
||||
try:
|
||||
amdsmi_interface.amdsmi_reset_gpu_xgmi_error(args.gpu)
|
||||
result = 'Successfully reset XGMI Error count'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to reset xgmi error count on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_xgmi_err', result)
|
||||
if args.perf_determinism:
|
||||
try:
|
||||
level_auto = amdsmi_interface.AmdSmiDevPerfLevel.AUTO
|
||||
amdsmi_interface.amdsmi_set_gpu_perf_level(args.gpu, level_auto)
|
||||
result = 'Successfully disabled performance determinism'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to set perf level on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_perf_determinism', result)
|
||||
if args.compute_partition:
|
||||
try:
|
||||
amdsmi_interface.amdsmi_reset_gpu_compute_partition(args.gpu)
|
||||
result = 'Successfully reset compute partition'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to reset compute partition on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_compute_partition', result)
|
||||
if args.memory_partition:
|
||||
try:
|
||||
amdsmi_interface.amdsmi_reset_gpu_memory_partition(args.gpu)
|
||||
result = 'Successfully reset memory partition'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to reset memory partition on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_memory_partition', result)
|
||||
if args.power_cap:
|
||||
try:
|
||||
power_cap_info = amdsmi_interface.amdsmi_get_power_cap_info(args.gpu)
|
||||
logging.debug(f"Power cap info for gpu {gpu_id} | {power_cap_info}")
|
||||
default_power_cap_in_w = power_cap_info["default_power_cap"]
|
||||
default_power_cap_in_w = self.helpers.convert_SI_unit(default_power_cap_in_w, AMDSMIHelpers.SI_Unit.MICRO)
|
||||
current_power_cap_in_w = power_cap_info["power_cap"]
|
||||
current_power_cap_in_w = self.helpers.convert_SI_unit(current_power_cap_in_w, AMDSMIHelpers.SI_Unit.MICRO)
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
raise ValueError(f"Unable to get power cap info from {gpu_id}") from e
|
||||
try:
|
||||
level_auto = amdsmi_interface.AmdSmiDevPerfLevel.AUTO
|
||||
amdsmi_interface.amdsmi_set_gpu_perf_level(args.gpu, level_auto)
|
||||
reset_profile_results['performance_level'] = 'Successfully reset Performance Level'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
reset_profile_results['performance_level'] = "N/A"
|
||||
logging.debug("Failed to reset perf level on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
|
||||
self.logger.store_output(args.gpu, 'reset_profile', reset_profile_results)
|
||||
if args.xgmierr:
|
||||
try:
|
||||
amdsmi_interface.amdsmi_reset_gpu_xgmi_error(args.gpu)
|
||||
result = 'Successfully reset XGMI Error count'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to reset xgmi error count on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_xgmi_err', result)
|
||||
if args.perf_determinism:
|
||||
try:
|
||||
level_auto = amdsmi_interface.AmdSmiDevPerfLevel.AUTO
|
||||
amdsmi_interface.amdsmi_set_gpu_perf_level(args.gpu, level_auto)
|
||||
result = 'Successfully disabled performance determinism'
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
|
||||
raise PermissionError('Command requires elevation') from e
|
||||
result = "N/A"
|
||||
logging.debug("Failed to set perf level on gpu %s | %s", gpu_id, e.get_error_info())
|
||||
self.logger.store_output(args.gpu, 'reset_perf_determinism', result)
|
||||
if args.power_cap:
|
||||
try:
|
||||
power_cap_info = amdsmi_interface.amdsmi_get_power_cap_info(args.gpu)
|
||||
logging.debug(f"Power cap info for gpu {gpu_id} | {power_cap_info}")
|
||||
default_power_cap_in_w = power_cap_info["default_power_cap"]
|
||||
default_power_cap_in_w = self.helpers.convert_SI_unit(default_power_cap_in_w, AMDSMIHelpers.SI_Unit.MICRO)
|
||||
current_power_cap_in_w = power_cap_info["power_cap"]
|
||||
current_power_cap_in_w = self.helpers.convert_SI_unit(current_power_cap_in_w, AMDSMIHelpers.SI_Unit.MICRO)
|
||||
except amdsmi_exception.AmdSmiLibraryException as e:
|
||||
raise ValueError(f"Unable to get power cap info from {gpu_id}") from e
|
||||
|
||||
if current_power_cap_in_w == default_power_cap_in_w:
|
||||
self.logger.store_output(args.gpu, 'powercap', f"Power cap is already set to {default_power_cap_in_w}")
|
||||
|
||||
@@ -1128,8 +1128,6 @@ class AMDSMIParser(argparse.ArgumentParser):
|
||||
reset_profile_help = "Reset power profile back to default"
|
||||
reset_xgmierr_help = "Reset XGMI error counts"
|
||||
reset_perf_det_help = "Disable performance determinism"
|
||||
reset_compute_help = "Reset compute partitions on the specified GPU"
|
||||
reset_memory_help = "Reset memory partitions on the specified GPU"
|
||||
reset_power_cap_help = "Reset power capacity limit to max capable"
|
||||
reset_gpu_clean_local_data_help = "Clean up local data in LDS/GPRs on a per partition basis"
|
||||
|
||||
@@ -1152,8 +1150,6 @@ class AMDSMIParser(argparse.ArgumentParser):
|
||||
reset_parser.add_argument('-p', '--profile', action='store_true', required=False, help=reset_profile_help)
|
||||
reset_parser.add_argument('-x', '--xgmierr', action='store_true', required=False, help=reset_xgmierr_help)
|
||||
reset_parser.add_argument('-d', '--perf-determinism', action='store_true', required=False, help=reset_perf_det_help)
|
||||
reset_parser.add_argument('-C', '--compute-partition', action='store_true', required=False, help=reset_compute_help)
|
||||
reset_parser.add_argument('-M', '--memory-partition', action='store_true', required=False, help=reset_memory_help)
|
||||
reset_parser.add_argument('-o', '--power-cap', action='store_true', required=False, help=reset_power_cap_help)
|
||||
|
||||
# Add Baremetal and Virtual OS reset arguments
|
||||
|
||||
@@ -3703,35 +3703,6 @@ except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_reset_gpu_compute_partition
|
||||
|
||||
Description: Reset the compute partitioning on the given GPU
|
||||
|
||||
Input parameters:
|
||||
|
||||
* `processor_handle` the device handle
|
||||
|
||||
Output: String of the partition type
|
||||
|
||||
Exceptions that can be thrown by `amdsmi_reset_gpu_compute_partition` function:
|
||||
|
||||
* `AmdSmiLibraryException`
|
||||
* `AmdSmiRetryException`
|
||||
* `AmdSmiParameterException`
|
||||
|
||||
Example:
|
||||
|
||||
```python
|
||||
try:
|
||||
devices = amdsmi_get_processor_handles()
|
||||
if len(devices) == 0:
|
||||
print("No GPUs on machine")
|
||||
else:
|
||||
for device in devices:
|
||||
amdsmi_reset_gpu_compute_partition(device)
|
||||
except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_get_gpu_memory_partition
|
||||
|
||||
@@ -3796,36 +3767,6 @@ except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_reset_gpu_memory_partition
|
||||
|
||||
Description: Reset the memory partitioning on the given GPU
|
||||
|
||||
Input parameters:
|
||||
|
||||
* `processor_handle` the device handle
|
||||
|
||||
Output: String of the partition type
|
||||
|
||||
Exceptions that can be thrown by `amdsmi_reset_gpu_memory_partition` function:
|
||||
|
||||
* `AmdSmiLibraryException`
|
||||
* `AmdSmiRetryException`
|
||||
* `AmdSmiParameterException`
|
||||
|
||||
Example:
|
||||
|
||||
```python
|
||||
try:
|
||||
devices = amdsmi_get_processor_handles()
|
||||
if len(devices) == 0:
|
||||
print("No GPUs on machine")
|
||||
else:
|
||||
for device in devices:
|
||||
amdsmi_reset_gpu_memory_partition(device)
|
||||
except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_get_xgmi_info
|
||||
|
||||
Description: Returns XGMI information for the GPU.
|
||||
|
||||
@@ -4578,25 +4578,6 @@ amdsmi_status_t
|
||||
amdsmi_set_gpu_compute_partition(amdsmi_processor_handle processor_handle,
|
||||
amdsmi_compute_partition_type_t compute_partition);
|
||||
|
||||
/**
|
||||
* @brief Reverts a selected device's compute partition setting back to its
|
||||
* boot state.
|
||||
*
|
||||
* @platform{gpu_bm_linux}
|
||||
*
|
||||
* @details Given a processor handle @p processor_handle, this function will attempt to
|
||||
* revert its compute partition setting back to its boot state.
|
||||
*
|
||||
* @param[in] processor_handle Device which to query
|
||||
*
|
||||
* @retval ::AMDSMI_STATUS_SUCCESS call was successful
|
||||
* @retval ::AMDSMI_STATUS_PERMISSION function requires root access
|
||||
* @retval ::AMDSMI_STATUS_NOT_SUPPORTED installed software or hardware does not
|
||||
* support this function
|
||||
*
|
||||
*/
|
||||
amdsmi_status_t amdsmi_reset_gpu_compute_partition(amdsmi_processor_handle processor_handle);
|
||||
|
||||
/** @} */ // end of compute_partition
|
||||
|
||||
/*****************************************************************************/
|
||||
@@ -4667,27 +4648,6 @@ amdsmi_status_t
|
||||
amdsmi_set_gpu_memory_partition(amdsmi_processor_handle processor_handle,
|
||||
amdsmi_memory_partition_type_t memory_partition);
|
||||
|
||||
/**
|
||||
* @brief Reverts a selected device's memory partition setting back to its
|
||||
* boot state.
|
||||
*
|
||||
* @platform{gpu_bm_linux}
|
||||
*
|
||||
* @details Given a processor handle @p processor_handle, this function will attempt to
|
||||
* revert its current memory partition setting back to its boot state.
|
||||
*
|
||||
* @param[in] processor_handle Device which to query
|
||||
*
|
||||
* @retval ::AMDSMI_STATUS_SUCCESS call was successful
|
||||
* @retval ::AMDSMI_STATUS_PERMISSION function requires root access
|
||||
* @retval ::AMDSMI_STATUS_NOT_SUPPORTED installed software or hardware does not
|
||||
* support this function
|
||||
* @retval ::AMDSMI_STATUS_AMDGPU_RESTART_ERR could not successfully restart
|
||||
* the amdgpu driver
|
||||
*
|
||||
*/
|
||||
amdsmi_status_t amdsmi_reset_gpu_memory_partition(amdsmi_processor_handle processor_handle);
|
||||
|
||||
/** @} */ // end of memory_partition
|
||||
|
||||
/*****************************************************************************/
|
||||
|
||||
@@ -70,6 +70,7 @@ endif()
|
||||
configure_file(pyproject.toml.in ${PY_BUILD_DIR}/pyproject.toml @ONLY)
|
||||
configure_file(setup.cfg.in ${PY_BUILD_DIR}/setup.cfg @ONLY)
|
||||
configure_file(_version.py.in ${PY_PACKAGE_DIR}/_version.py @ONLY)
|
||||
configure_file(setup.py.in ${PY_BUILD_DIR}/setup.py @ONLY)
|
||||
|
||||
add_custom_target(
|
||||
python_wrapper
|
||||
@@ -103,6 +104,7 @@ add_custom_target(
|
||||
python_package ALL
|
||||
DEPENDS ${PY_BUILD_DIR}/pyproject.toml
|
||||
${PY_BUILD_DIR}/setup.cfg
|
||||
${PY_BUILD_DIR}/setup.py
|
||||
${PY_PACKAGE_DIR}/_version.py
|
||||
${PY_PACKAGE_DIR}/__init__.py
|
||||
${PY_PACKAGE_DIR}/amdsmi_exception.py
|
||||
@@ -116,6 +118,7 @@ install(
|
||||
FILES
|
||||
${CMAKE_CURRENT_BINARY_DIR}/${PY_BUILD_DIR}/pyproject.toml
|
||||
${CMAKE_CURRENT_BINARY_DIR}/${PY_BUILD_DIR}/setup.cfg
|
||||
${CMAKE_CURRENT_BINARY_DIR}/${PY_BUILD_DIR}/setup.py
|
||||
${CMAKE_CURRENT_BINARY_DIR}/${PY_PACKAGE_DIR}/_version.py
|
||||
DESTINATION ${PY_WRAPPER_INSTALL_DIR}
|
||||
COMPONENT dev)
|
||||
|
||||
@@ -3704,36 +3704,6 @@ except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_reset_gpu_compute_partition
|
||||
|
||||
Description: Reset the compute partitioning on the given GPU
|
||||
|
||||
Input parameters:
|
||||
|
||||
* `processor_handle` the device handle
|
||||
|
||||
Output: String of the partition type
|
||||
|
||||
Exceptions that can be thrown by `amdsmi_reset_gpu_compute_partition` function:
|
||||
|
||||
* `AmdSmiLibraryException`
|
||||
* `AmdSmiRetryException`
|
||||
* `AmdSmiParameterException`
|
||||
|
||||
Example:
|
||||
|
||||
```python
|
||||
try:
|
||||
devices = amdsmi_get_processor_handles()
|
||||
if len(devices) == 0:
|
||||
print("No GPUs on machine")
|
||||
else:
|
||||
for device in devices:
|
||||
amdsmi_reset_gpu_compute_partition(device)
|
||||
except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_get_gpu_memory_partition
|
||||
|
||||
Description: Get the memory partition from the given GPU
|
||||
@@ -3797,35 +3767,6 @@ except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_reset_gpu_memory_partition
|
||||
|
||||
Description: Reset the memory partitioning on the given GPU
|
||||
|
||||
Input parameters:
|
||||
|
||||
* `processor_handle` the device handle
|
||||
|
||||
Output: String of the partition type
|
||||
|
||||
Exceptions that can be thrown by `amdsmi_reset_gpu_memory_partition` function:
|
||||
|
||||
* `AmdSmiLibraryException`
|
||||
* `AmdSmiRetryException`
|
||||
* `AmdSmiParameterException`
|
||||
|
||||
Example:
|
||||
|
||||
```python
|
||||
try:
|
||||
devices = amdsmi_get_processor_handles()
|
||||
if len(devices) == 0:
|
||||
print("No GPUs on machine")
|
||||
else:
|
||||
for device in devices:
|
||||
amdsmi_reset_gpu_memory_partition(device)
|
||||
except AmdSmiException as e:
|
||||
print(e)
|
||||
```
|
||||
|
||||
### amdsmi_get_gpu_accelerator_partition_profile
|
||||
|
||||
|
||||
@@ -222,10 +222,8 @@ from .amdsmi_interface import amdsmi_get_link_topology_nearest
|
||||
# # Partition Functions
|
||||
from .amdsmi_interface import amdsmi_get_gpu_compute_partition
|
||||
from .amdsmi_interface import amdsmi_set_gpu_compute_partition
|
||||
from .amdsmi_interface import amdsmi_reset_gpu_compute_partition
|
||||
from .amdsmi_interface import amdsmi_get_gpu_memory_partition
|
||||
from .amdsmi_interface import amdsmi_set_gpu_memory_partition
|
||||
from .amdsmi_interface import amdsmi_reset_gpu_memory_partition
|
||||
from .amdsmi_interface import amdsmi_get_gpu_accelerator_partition_profile
|
||||
|
||||
# # Individual GPU Metrics Functions
|
||||
|
||||
@@ -2691,15 +2691,6 @@ def amdsmi_set_gpu_compute_partition(processor_handle: amdsmi_wrapper.amdsmi_pro
|
||||
)
|
||||
|
||||
|
||||
def amdsmi_reset_gpu_compute_partition(processor_handle: amdsmi_wrapper.amdsmi_processor_handle):
|
||||
if not isinstance(processor_handle, amdsmi_wrapper.amdsmi_processor_handle):
|
||||
raise AmdSmiParameterException(
|
||||
processor_handle, amdsmi_wrapper.amdsmi_processor_handle
|
||||
)
|
||||
|
||||
_check_res(amdsmi_wrapper.amdsmi_reset_gpu_compute_partition(processor_handle))
|
||||
|
||||
|
||||
def amdsmi_get_gpu_memory_partition(processor_handle: amdsmi_wrapper.amdsmi_processor_handle):
|
||||
if not isinstance(processor_handle, amdsmi_wrapper.amdsmi_processor_handle):
|
||||
raise AmdSmiParameterException(
|
||||
@@ -2737,15 +2728,6 @@ def amdsmi_set_gpu_memory_partition(processor_handle: amdsmi_wrapper.amdsmi_proc
|
||||
)
|
||||
|
||||
|
||||
def amdsmi_reset_gpu_memory_partition(processor_handle: amdsmi_wrapper.amdsmi_processor_handle):
|
||||
if not isinstance(processor_handle, amdsmi_wrapper.amdsmi_processor_handle):
|
||||
raise AmdSmiParameterException(
|
||||
processor_handle, amdsmi_wrapper.amdsmi_processor_handle
|
||||
)
|
||||
|
||||
_check_res(amdsmi_wrapper.amdsmi_reset_gpu_memory_partition(processor_handle))
|
||||
|
||||
|
||||
def amdsmi_get_gpu_accelerator_partition_profile(
|
||||
processor_handle: amdsmi_wrapper.amdsmi_processor_handle
|
||||
) -> Dict[str, Any]:
|
||||
|
||||
@@ -2349,18 +2349,12 @@ amdsmi_get_gpu_compute_partition.argtypes = [amdsmi_processor_handle, ctypes.POI
|
||||
amdsmi_set_gpu_compute_partition = _libraries['libamd_smi.so'].amdsmi_set_gpu_compute_partition
|
||||
amdsmi_set_gpu_compute_partition.restype = amdsmi_status_t
|
||||
amdsmi_set_gpu_compute_partition.argtypes = [amdsmi_processor_handle, amdsmi_compute_partition_type_t]
|
||||
amdsmi_reset_gpu_compute_partition = _libraries['libamd_smi.so'].amdsmi_reset_gpu_compute_partition
|
||||
amdsmi_reset_gpu_compute_partition.restype = amdsmi_status_t
|
||||
amdsmi_reset_gpu_compute_partition.argtypes = [amdsmi_processor_handle]
|
||||
amdsmi_get_gpu_memory_partition = _libraries['libamd_smi.so'].amdsmi_get_gpu_memory_partition
|
||||
amdsmi_get_gpu_memory_partition.restype = amdsmi_status_t
|
||||
amdsmi_get_gpu_memory_partition.argtypes = [amdsmi_processor_handle, ctypes.POINTER(ctypes.c_char), uint32_t]
|
||||
amdsmi_set_gpu_memory_partition = _libraries['libamd_smi.so'].amdsmi_set_gpu_memory_partition
|
||||
amdsmi_set_gpu_memory_partition.restype = amdsmi_status_t
|
||||
amdsmi_set_gpu_memory_partition.argtypes = [amdsmi_processor_handle, amdsmi_memory_partition_type_t]
|
||||
amdsmi_reset_gpu_memory_partition = _libraries['libamd_smi.so'].amdsmi_reset_gpu_memory_partition
|
||||
amdsmi_reset_gpu_memory_partition.restype = amdsmi_status_t
|
||||
amdsmi_reset_gpu_memory_partition.argtypes = [amdsmi_processor_handle]
|
||||
amdsmi_get_gpu_accelerator_partition_profile = _libraries['libamd_smi.so'].amdsmi_get_gpu_accelerator_partition_profile
|
||||
amdsmi_get_gpu_accelerator_partition_profile.restype = amdsmi_status_t
|
||||
amdsmi_get_gpu_accelerator_partition_profile.argtypes = [amdsmi_processor_handle, ctypes.POINTER(struct_amdsmi_accelerator_partition_profile_t), ctypes.POINTER(ctypes.c_uint32)]
|
||||
@@ -2899,10 +2893,9 @@ __all__ = \
|
||||
'amdsmi_process_info_t', 'amdsmi_processor_handle',
|
||||
'amdsmi_range_t', 'amdsmi_ras_err_state_t',
|
||||
'amdsmi_ras_feature_t', 'amdsmi_reg_type_t', 'amdsmi_reset_gpu',
|
||||
'amdsmi_reset_gpu_compute_partition', 'amdsmi_reset_gpu_fan',
|
||||
'amdsmi_reset_gpu_memory_partition',
|
||||
'amdsmi_reset_gpu_xgmi_error', 'amdsmi_retired_page_record_t',
|
||||
'amdsmi_set_clk_freq', 'amdsmi_set_cpu_core_boostlimit',
|
||||
'amdsmi_reset_gpu_fan', 'amdsmi_reset_gpu_xgmi_error',
|
||||
'amdsmi_retired_page_record_t', 'amdsmi_set_clk_freq',
|
||||
'amdsmi_set_cpu_core_boostlimit',
|
||||
'amdsmi_set_cpu_df_pstate_range',
|
||||
'amdsmi_set_cpu_gmi3_link_width_range',
|
||||
'amdsmi_set_cpu_pcie_link_rate',
|
||||
@@ -2970,5 +2963,6 @@ __all__ = \
|
||||
'struct_cache_', 'struct_engine_usage_', 'struct_fw_info_list_',
|
||||
'struct_memory_usage_', 'struct_nps_flags_',
|
||||
'struct_pcie_metric_', 'struct_pcie_static_',
|
||||
'struct_amdsmi_bdf_t','uint32_t', 'uint64_t', 'uint8_t',
|
||||
'struct_amdsmi_bdf_t', 'uint32_t', 'uint64_t', 'uint8_t',
|
||||
'union_amdsmi_bdf_t', 'union_amdsmi_nps_caps_t']
|
||||
|
||||
|
||||
@@ -16,7 +16,7 @@ readme = {file = "amdsmi/README.md", content-type = "text/markdown"}
|
||||
description = "AMDSMI Python LIB - AMD GPU Monitoring Library"
|
||||
requires-python = ">=3.6"
|
||||
dependencies = [
|
||||
"PyYAML >= 3.12",
|
||||
"PyYAML >= 3.0",
|
||||
]
|
||||
|
||||
[project.urls]
|
||||
|
||||
@@ -18,7 +18,7 @@ include_package_data = True
|
||||
packages = find:
|
||||
python_requires = >=3.6
|
||||
install_requires=
|
||||
PyYAML >= 3.12
|
||||
PyYAML >= 3.0
|
||||
|
||||
[options.package_data]
|
||||
* = *.so
|
||||
|
||||
@@ -0,0 +1,25 @@
|
||||
from setuptools import setup, find_packages
|
||||
import os
|
||||
|
||||
setup(
|
||||
name="amdsmi",
|
||||
version="@amd_smi_libraries_VERSION_STRING@",
|
||||
author="AMD",
|
||||
author_email="amd-smi.support@amd.com",
|
||||
description="AMDSMI Python LIB - AMD GPU Monitoring Library",
|
||||
url="https://github.com/ROCm/amdsmi",
|
||||
packages=find_packages(),
|
||||
install_requires=[
|
||||
"PyYAML>=3.0",
|
||||
],
|
||||
classifiers=[
|
||||
"Programming Language :: Python :: 3",
|
||||
],
|
||||
python_requires=">=3.6",
|
||||
include_package_data=True,
|
||||
package_data={
|
||||
'': ['*.so'],
|
||||
},
|
||||
zip_safe=False,
|
||||
license='amdsmi/LICENSE',
|
||||
)
|
||||
@@ -1431,12 +1431,6 @@ amdsmi_set_gpu_compute_partition(amdsmi_processor_handle processor_handle,
|
||||
static_cast<rsmi_compute_partition_type_t>(compute_partition));
|
||||
}
|
||||
|
||||
amdsmi_status_t
|
||||
amdsmi_reset_gpu_compute_partition(amdsmi_processor_handle processor_handle) {
|
||||
AMDSMI_CHECK_INIT();
|
||||
return rsmi_wrapper(rsmi_dev_compute_partition_reset, processor_handle);
|
||||
}
|
||||
|
||||
// Memory Partition functions
|
||||
amdsmi_status_t
|
||||
amdsmi_get_gpu_memory_partition(amdsmi_processor_handle processor_handle,
|
||||
@@ -1454,12 +1448,6 @@ amdsmi_set_gpu_memory_partition(amdsmi_processor_handle processor_handle,
|
||||
static_cast<rsmi_memory_partition_type_t>(memory_partition));
|
||||
}
|
||||
|
||||
amdsmi_status_t
|
||||
amdsmi_reset_gpu_memory_partition(amdsmi_processor_handle processor_handle) {
|
||||
AMDSMI_CHECK_INIT();
|
||||
return rsmi_wrapper(rsmi_dev_memory_partition_reset, processor_handle);
|
||||
}
|
||||
|
||||
amdsmi_status_t
|
||||
amdsmi_get_gpu_accelerator_partition_profile(amdsmi_processor_handle processor_handle,
|
||||
amdsmi_accelerator_partition_profile_t *profile,
|
||||
|
||||
Ссылка в новой задаче
Block a user