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

Merge amd-staging into amd-master 20240321

Browse Source 
Signed-off-by: guanyu12 <guanyu12@amd.com>
Change-Id: I006fc6c187f134a4851e262fa53ab6bf8d58759d
This commit is contained in:
guanyu12
2024-03-21 14:03:51 +08:00
parent ab8ebd4dea c5acd4ee88
commit 8d4261c5c5
4 changed files with 321 additions and 154 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+122 -1
View File
@@ -1,6 +1,127 @@
# Change Log for ROCm SMI Library # Change Log for ROCm SMI Library
Full documentation for rocm_smi_lib is available at [https://docs.amd.com/](https://rocm.docs.amd.com/projects/rocm_smi_lib/en/latest/). Full documentation for rocm_smi_lib is available at [https://rocm.docs.amd.com/](https://rocm.docs.amd.com/projects/rocm_smi_lib/en/latest/).
***All information listed below is for reference and subject to change.***
## rocm_smi_lib for ROCm 6.1.0
### Added
- **Added support to set max/min clock level for sclk (`RSMI_CLK_TYPE_SYS`) or mclk (`RSMI_CLK_TYPE_MEM`)**
Users can now set a maximum or minimum sclk or mclk value through `rsmi_dev_clk_extremum_set()` API provided ASIC support. Alternatively, users can
use our Python CLI tool (`rocm-smi --setextremum max sclk 1500`). See example below.
```shell
$ sudo /opt/rocm/bin/rocm-smi --setextremum max sclk 2100
============================ ROCm System Management Interface ============================
******WARNING******
Operating your AMD GPU outside of official AMD specifications or outside of
factory settings, including but not limited to the conducting of overclocking,
over-volting or under-volting (including use of this interface software,
even if such software has been directly or indirectly provided by AMD or otherwise
affiliated in any way with AMD), may cause damage to your AMD GPU, system components
and/or result in system failure, as well as cause other problems.
DAMAGES CAUSED BY USE OF YOUR AMD GPU OUTSIDE OF OFFICIAL AMD SPECIFICATIONS OR
OUTSIDE OF FACTORY SETTINGS ARE NOT COVERED UNDER ANY AMD PRODUCT WARRANTY AND
MAY NOT BE COVERED BY YOUR BOARD OR SYSTEM MANUFACTURER'S WARRANTY.
Please use this utility with caution.
Do you accept these terms? [y/N] y
================================ Set Valid sclk Extremum =================================
GPU[0] : Successfully set max sclk to 2100(MHz)
GPU[1] : Successfully set max sclk to 2100(MHz)
GPU[2] : Successfully set max sclk to 2100(MHz)
GPU[3] : Successfully set max sclk to 2100(MHz)
================================== End of ROCm SMI Log ===================================
```
- **Added `rsmi_dev_target_graphics_version_get()`**
Users can now query through ROCm SMI API (`rsmi_dev_target_graphics_version_get()`) to retreive the target graphics version for a GPU device. Currently, this output is not supplied through our rocm-smi CLI.
### Changed
- **Removed non-unified API headers: Individual GPU metric APIs are no longer supported**
The individual metric APIs (`rsmi_dev_metrics_*`) were removed in order to keep updates easier for new GPU metric support. By providing a simple API (`rsmi_dev_gpu_metrics_info_get()`) with its reported device metrics, it is worth noting there is a risk for ABI break-age using `rsmi_dev_gpu_metrics_info_get()`. It is vital to understand, that ABI breaks are necessary (in some cases) in order to support newer ASICs and metrics for our customers. We will continue to support `rsmi_dev_gpu_metrics_info_get()` with these considerations and limitations in mind.
- **Depricated rsmi_dev_power_ave_get(), use newer API rsmi_dev_power_get()**
As outlined in change below for 6.0.0 (***Added a generic power API: rsmi_dev_power_get***), is now depricated. Please update your ROCm SMI API calls accordingly.
### Optimizations
- N/A
### Fixed
- Fix `--showpids` reporting `[PID] [PROCESS NAME] 1 UNKNOWN UNKNOWN UNKNOWN`
Output was failing because cu_occupancy debugfs method is not provided on some graphics cards by design. `get_compute_process_info_by_pid` was updated to reflect this and returns with output needed by CLI.
- Fix `rocm-smi --showpower` output was inconsistent on Navi32/31 devices
Updated to use `rsmi_dev_power_get()` within CLI to provide a consistent device power output. This was caused due to using the now depricated `rsmi_dev_average_power_get()` API.
- Fixed `rocm-smi --setcomputepartition` and `rocm-smi --resetcomputepartition` to notate if device is EBUSY
- Fixed `rocm-smi --setmemorypartition` and `rocm-smi --resetmemorypartition` read only SYSFS to return RSMI_STATUS_NOT_SUPPORTED
The `rsmi_dev_memory_partition_set` API is updated to handle the readonly SYSFS check. Corresponding tests and CLI (`rocm-smi --setmemorypartition` and `rocm-smi --resetmemorypartition`) calls were updated accordingly.
- Fix `rocm-smi --showclkvolt` and `rocm-smi --showvc` displaying 0 for overdrive and voltage curve is not supported
## rocm_smi_lib for ROCm 6.0.0
### Added
- **Added rocm-smi --version**
The SMI will report two "versions", ROCM-SMI version and other is ROCM-SMI-LIB version.
- The ROCM-SMI version is the CLI/tool version number with commit ID appended after `+` sign.
- The ROCM-SMI-LIB version is the library package version number.
```
$ rocm-smi --version
ROCM-SMI version: 2.0.0+8e78352
ROCM-SMI-LIB version: 6.0.0
```
- **Added support for gfx941/gfx942 metrics**
You can now query MI300 device metrics to get real-time information. Metrics include power, temperature, energy, and performance. Users can query through `rsmi_dev_gpu_metrics_info_get()`.
- **Compute and memory partition support**
Users can now view, set, and reset partitions. The topology display can provide a more in-depth look at the device's current configuration. If your ASIC supports these features, the following commands can help get started:
- `rocm-smi --showcomputepartition`
- `rocm-smi --setcomputepartition <SPX, DPX, CPX, TPX, QPX>`
- `rocm-smi --resetcomputepartition`
- `rocm-smi --showmemorypartition`
- `rocm-smi --setmemorypartition <NPS1, NPS2, NPS4, NPS8>`
- `rocm-smi --resetmemorypartition`
### Changed
- **GPU index sorting made consistent with other tools**
To ensure alignment with other ROCm software tools, GPU index sorting is optimized to use Bus:Device.Function (BDF) rather than the card number.
- **Increase max BDF ID length**
To allow for larger BDF data, we have increased the maximum BDF length from 256 to 512 buffer size.
- **Documentation is transitioning to Sphinx**
Sphinx allows us to generate code documentation easier for our users. Helps us provide centrized HTML documentation at single website location. Here customers can see how to use our software and tools.
- **Added a generic power API: `rsmi_dev_power_get()`**
Older ASICs provided average socket power, newer ASICs (MI300) provide current socket power. The generic API provides one interface to retreive either of these power readings, allowing backwards compatability.
- **Added flexible temperature readings (`rocm-smi` and `rocm-smi --showtempgraph`)**
Older ASICs provided edge temperature, newer ASICs (MI300) provide junction socket power (not edge). The rocm-smi CLI now provides a way to view which type of temperature is read across all sockets.
- **Added deep sleep frequency readings**
Newer ASICs (MI300) provide ability to know if a clock is in deep sleep.
### Optimizations
- Add new test to measure api execution time.
- Remove the shared mutex if no process is using it.
- Updated to C++17, gtest-1.14, and cmake 3.14
### Fixed
- Fix memory usage division by 0
- Fix missing firmware blocks (rocm-smi --showfw)
- Fix rocm-smi --showevents shows wrong gpuID
## rocm_smi_lib for ROCm 5.5.0 ## rocm_smi_lib for ROCm 5.5.0
README.md
+14 -1
View File
@@ -8,7 +8,20 @@ For additional information refer to [ROCm Documentation](https://rocm.docs.amd.c
The information contained herein is for informational purposes only, and is subject to change without notice. In addition, any stated support is planned and is also subject to change. While every precaution has been taken in the preparation of this document, it may contain technical inaccuracies, omissions and typographical errors, and AMD is under no obligation to update or otherwise correct this information. Advanced Micro Devices, Inc. makes no representations or warranties with respect to the accuracy or completeness of the contents of this document, and assumes no liability of any kind, including the implied warranties of noninfringement, merchantability or fitness for particular purposes, with respect to the operation or use of AMD hardware, software or other products described herein. The information contained herein is for informational purposes only, and is subject to change without notice. In addition, any stated support is planned and is also subject to change. While every precaution has been taken in the preparation of this document, it may contain technical inaccuracies, omissions and typographical errors, and AMD is under no obligation to update or otherwise correct this information. Advanced Micro Devices, Inc. makes no representations or warranties with respect to the accuracy or completeness of the contents of this document, and assumes no liability of any kind, including the implied warranties of noninfringement, merchantability or fitness for particular purposes, with respect to the operation or use of AMD hardware, software or other products described herein.
© 2022-2023 Advanced Micro Devices, Inc. All Rights Reserved. © 2022-2024 Advanced Micro Devices, Inc. All Rights Reserved.
## Installation
### Install amdgpu using ROCm
* Install amdgpu driver:
See example below, your release and link may differ. The `amdgpu-install --usecase=rocm` triggers both an amdgpu driver update and ROCm SMI packages to be installed on your device.
```shell
sudo apt update
wget https://repo.radeon.com/amdgpu-install/6.0.2/ubuntu/jammy/amdgpu-install_6.0.60002-1_all.deb
sudo apt install ./amdgpu-install_6.0.60002-1_all.deb
sudo amdgpu-install --usecase=rocm
```
* rocm-smi --help
## Building ROCm SMI ## Building ROCm SMI
python_smi_tools/README.md
+184 -151
View File
@@ -15,15 +15,13 @@ LD_LIBRARY_PATH should be set to the folder containing librocm_smi64.
## Version ## Version
The SMI will report a "version" which is the version of the kernel installed: The SMI will report two "versions", ROCM-SMI version and other is ROCM-SMI-LIB version.
- ROCM-SMI version is the CLI/tool version number with commit ID appended after + sign.
AMD ROCm System Management Interface v$(uname) - ROCM-SMI-LIB version is the library package version number.
```
For ROCk installations, this will be the AMDGPU module version (e.g. 5.0.71) ROCM-SMI version: 2.0.0+8e78352
For non-ROCk or monolithic ROCk installations, this will be the kernel version, ROCM-SMI-LIB version: 6.1.0
which will be equivalent to the following bash command: ```
$(uname -a) | cut -d ' ' -f 3)
## Usage ## Usage
@@ -33,169 +31,204 @@ For detailed and up to date usage information, we recommend consulting the help:
For convenience purposes, following is the output from the -h flag: For convenience purposes, following is the output from the -h flag:
usage: rocm-smi [-h] [-V] [-d DEVICE [DEVICE ...]] [--alldevices] [--showhw] [-a] [-i] [-v] [-e [EVENT ...]] ```
[--showdriverversion] [--showtempgraph] [--showfwinfo [BLOCK ...]] [--showmclkrange] [--showmemvendor] /opt/rocm/bin/rocm-smi -h
[--showsclkrange] [--showproductname] [--showserial] [--showuniqueid] [--showvoltagerange] [--showbus] usage: rocm-smi [-h] [-V] [-d DEVICE [DEVICE ...]] [--alldevices] [--showhw] [-a] [-i] [-v] [-e [EVENT [EVENT ...]]]
[--showpagesinfo] [--showpendingpages] [--showretiredpages] [--showunreservablepages] [-f] [-P] [-t] [--showdriverversion] [--showtempgraph] [--showfwinfo [BLOCK [BLOCK ...]]] [--showmclkrange]
[-u] [--showmemuse] [--showvoltage] [-b] [-c] [-g] [-l] [-M] [-m] [-o] [-p] [-S] [-s] [--showmemvendor] [--showsclkrange] [--showproductname] [--showserial] [--showuniqueid]
[--showmeminfo TYPE [TYPE ...]] [--showpids [VERBOSE]] [--showpidgpus [SHOWPIDGPUS ...]] [--showvoltagerange] [--showbus] [--showpagesinfo] [--showpendingpages] [--showretiredpages]
[--showreplaycount] [--showrasinfo [SHOWRASINFO ...]] [--showvc] [--showxgmierr] [--showtopo] [--showunreservablepages] [-f] [-P] [-t] [-u] [--showmemuse] [--showvoltage] [-b] [-c] [-g] [-l] [-M]
[-m] [-o] [-p] [-S] [-s] [--showmeminfo TYPE [TYPE ...]] [--showpids [VERBOSE]]
[--showpidgpus [SHOWPIDGPUS [SHOWPIDGPUS ...]]] [--showreplaycount]
[--showrasinfo [SHOWRASINFO [SHOWRASINFO ...]]] [--showvc] [--showxgmierr] [--showtopo]
[--showtopoaccess] [--showtopoweight] [--showtopohops] [--showtopotype] [--showtoponuma] [--showtopoaccess] [--showtopoweight] [--showtopohops] [--showtopotype] [--showtoponuma]
[--showenergycounter] [--shownodesbw] [--showcomputepartition] [--showmemorypartition] [-r] [--showenergycounter] [--shownodesbw] [--showcomputepartition] [--showmemorypartition] [-r]
[--resetfans] [--resetprofile] [--resetpoweroverdrive] [--resetxgmierr] [--resetperfdeterminism] [--resetfans] [--resetprofile] [--resetpoweroverdrive] [--resetxgmierr] [--resetperfdeterminism]
[--resetcomputepartition] [--resetmemorypartition] [--setclock TYPE LEVEL] [--setsclk LEVEL [LEVEL ...]] [--resetcomputepartition] [--resetmemorypartition] [--setclock TYPE LEVEL] [--setsclk LEVEL [LEVEL ...]]
[--setmclk LEVEL [LEVEL ...]] [--setpcie LEVEL [LEVEL ...]] [--setslevel SCLKLEVEL SCLK SVOLT] [--setmclk LEVEL [LEVEL ...]] [--setpcie LEVEL [LEVEL ...]] [--setslevel SCLKLEVEL SCLK SVOLT]
[--setmlevel MCLKLEVEL MCLK MVOLT] [--setvc POINT SCLK SVOLT] [--setsrange SCLKMIN SCLKMAX] [--setmlevel MCLKLEVEL MCLK MVOLT] [--setvc POINT SCLK SVOLT] [--setsrange SCLKMIN SCLKMAX]
[--setmrange MCLKMIN MCLKMAX] [--setfan LEVEL] [--setperflevel LEVEL] [--setoverdrive %] [--setextremum min|max sclk|mclk CLK] [--setmrange MCLKMIN MCLKMAX] [--setfan LEVEL]
[--setmemoverdrive %] [--setpoweroverdrive WATTS] [--setprofile SETPROFILE] [--setperfdeterminism SCLK] [--setperflevel LEVEL] [--setoverdrive %] [--setmemoverdrive %] [--setpoweroverdrive WATTS]
[--setprofile SETPROFILE] [--setperfdeterminism SCLK]
[--setcomputepartition {CPX,SPX,DPX,TPX,QPX,cpx,spx,dpx,tpx,qpx}] [--setcomputepartition {CPX,SPX,DPX,TPX,QPX,cpx,spx,dpx,tpx,qpx}]
[--setmemorypartition {NPS1,NPS2,NPS4,NPS8,nps1,nps2,nps4,nps8}] [--rasenable BLOCK ERRTYPE] [--setmemorypartition {NPS1,NPS2,NPS4,NPS8,nps1,nps2,nps4,nps8}] [--rasenable BLOCK ERRTYPE]
[--rasdisable BLOCK ERRTYPE] [--rasinject BLOCK] [--gpureset] [--load FILE | --save FILE] [--rasdisable BLOCK ERRTYPE] [--rasinject BLOCK] [--gpureset] [--load FILE | --save FILE]
[--autorespond RESPONSE] [--loglevel LEVEL] [--json] [--csv] [--autorespond RESPONSE] [--loglevel LEVEL] [--json] [--csv]
AMD ROCm System Management Interface | ROCM-SMI version: 1.5.0 AMD ROCm System Management Interface | ROCM-SMI version: 2.0.0+8e78352
options: optional arguments:
-h, --help show this help message and exit -h, --help show this help message and exit
--gpureset Reset specified GPU (One GPU must be specified) --gpureset Reset specified GPU (One GPU must be specified)
--load FILE Load Clock, Fan, Performance and Profile settings --load FILE Load Clock, Fan, Performance and Profile settings
from FILE from FILE
--save FILE Save Clock, Fan, Performance and Profile settings to --save FILE Save Clock, Fan, Performance and Profile settings to
FILE FILE
-V, --version Show version information -V, --version Show version information
-d DEVICE [DEVICE ...], --device DEVICE [DEVICE ...] Execute command on specified device -d DEVICE [DEVICE ...], --device DEVICE [DEVICE ...] Execute command on specified device
Display Options: Display Options:
--alldevices --alldevices
--showhw Show Hardware details --showhw Show Hardware details
-a, --showallinfo Show Temperature, Fan and Clock values -a, --showallinfo Show Temperature, Fan and Clock values
Topology: Topology:
-i, --showid Show DEVICE ID -i, --showid Show DEVICE ID
-v, --showvbios Show VBIOS version -v, --showvbios Show VBIOS version
-e [EVENT ...], --showevents [EVENT ...] Show event list -e [EVENT [EVENT ...]], --showevents [EVENT [EVENT ...]] Show event list
--showdriverversion Show kernel driver version --showdriverversion Show kernel driver version
--showtempgraph Show Temperature Graph --showtempgraph Show Temperature Graph
--showfwinfo [BLOCK ...] Show FW information --showfwinfo [BLOCK [BLOCK ...]] Show FW information
--showmclkrange Show mclk range --showmclkrange Show mclk range
--showmemvendor Show GPU memory vendor --showmemvendor Show GPU memory vendor
--showsclkrange Show sclk range --showsclkrange Show sclk range
--showproductname Show SKU/Vendor name --showproductname Show SKU/Vendor name
--showserial Show GPU's Serial Number --showserial Show GPU's Serial Number
--showuniqueid Show GPU's Unique ID --showuniqueid Show GPU's Unique ID
--showvoltagerange Show voltage range --showvoltagerange Show voltage range
--showbus Show PCI bus number --showbus Show PCI bus number
Pages information: Pages information:
--showpagesinfo Show retired, pending and unreservable pages --showpagesinfo Show retired, pending and unreservable pages
--showpendingpages Show pending retired pages --showpendingpages Show pending retired pages
--showretiredpages Show retired pages --showretiredpages Show retired pages
--showunreservablepages Show unreservable pages --showunreservablepages Show unreservable pages
Hardware-related information: Hardware-related information:
-f, --showfan Show current fan speed -f, --showfan Show current fan speed
-P, --showpower Show current Average Graphics Package Power -P, --showpower Show current Average or Socket Graphics Package Power
Consumption Consumption
-t, --showtemp Show current temperature -t, --showtemp Show current temperature
-u, --showuse Show current GPU use -u, --showuse Show current GPU use
--showmemuse Show current GPU memory used --showmemuse Show current GPU memory used
--showvoltage Show current GPU voltage --showvoltage Show current GPU voltage
Software-related/controlled information: Software-related/controlled information:
-b, --showbw Show estimated PCIe use -b, --showbw Show estimated PCIe use
-c, --showclocks Show current clock frequencies -c, --showclocks Show current clock frequencies
-g, --showgpuclocks Show current GPU clock frequencies -g, --showgpuclocks Show current GPU clock frequencies
-l, --showprofile Show Compute Profile attributes -l, --showprofile Show Compute Profile attributes
-M, --showmaxpower Show maximum graphics package power this GPU will -M, --showmaxpower Show maximum graphics package power this GPU will
consume consume
-m, --showmemoverdrive Show current GPU Memory Clock OverDrive level -m, --showmemoverdrive Show current GPU Memory Clock OverDrive level
-o, --showoverdrive Show current GPU Clock OverDrive level -o, --showoverdrive Show current GPU Clock OverDrive level
-p, --showperflevel Show current DPM Performance Level -p, --showperflevel Show current DPM Performance Level
-S, --showclkvolt Show supported GPU and Memory Clocks and Voltages -S, --showclkvolt Show supported GPU and Memory Clocks and Voltages
-s, --showclkfrq Show supported GPU and Memory Clock -s, --showclkfrq Show supported GPU and Memory Clock
--showmeminfo TYPE [TYPE ...] Show Memory usage information for given block(s) TYPE --showmeminfo TYPE [TYPE ...] Show Memory usage information for given block(s) TYPE
--showpids [VERBOSE] Show current running KFD PIDs (pass details to --showpids [VERBOSE] Show current running KFD PIDs (pass details to
VERBOSE for detailed information) VERBOSE for detailed information)
--showpidgpus [SHOWPIDGPUS ...] Show GPUs used by specified KFD PIDs (all if no arg --showpidgpus [SHOWPIDGPUS [SHOWPIDGPUS ...]] Show GPUs used by specified KFD PIDs (all if no arg
given) given)
--showreplaycount Show PCIe Replay Count --showreplaycount Show PCIe Replay Count
--showrasinfo [SHOWRASINFO ...] Show RAS enablement information and error counts for --showrasinfo [SHOWRASINFO [SHOWRASINFO ...]] Show RAS enablement information and error counts for
the specified block(s) (all if no arg given) the specified block(s) (all if no arg given)
--showvc Show voltage curve --showvc Show voltage curve
--showxgmierr Show XGMI error information since last read --showxgmierr Show XGMI error information since last read
--showtopo Show hardware topology information --showtopo Show hardware topology information
--showtopoaccess Shows the link accessibility between GPUs --showtopoaccess Shows the link accessibility between GPUs
--showtopoweight Shows the relative weight between GPUs --showtopoweight Shows the relative weight between GPUs
--showtopohops Shows the number of hops between GPUs --showtopohops Shows the number of hops between GPUs
--showtopotype Shows the link type between GPUs --showtopotype Shows the link type between GPUs
--showtoponuma Shows the numa nodes --showtoponuma Shows the numa nodes
--showenergycounter Energy accumulator that stores amount of energy --showenergycounter Energy accumulator that stores amount of energy
consumed consumed
--shownodesbw Shows the numa nodes --shownodesbw Shows the numa nodes
--showcomputepartition Shows current compute partitioning --showcomputepartition Shows current compute partitioning
--showmemorypartition Shows current memory partition --showmemorypartition Shows current memory partition
Set options: Set options:
--setclock TYPE LEVEL Set Clock Frequency Level(s) for specified clock --setclock TYPE LEVEL Set Clock Frequency Level(s) for specified clock
(requires manual Perf level) (requires manual Perf level)
--setsclk LEVEL [LEVEL ...] Set GPU Clock Frequency Level(s) (requires manual --setsclk LEVEL [LEVEL ...] Set GPU Clock Frequency Level(s) (requires manual
Perf level) Perf level)
--setmclk LEVEL [LEVEL ...] Set GPU Memory Clock Frequency Level(s) (requires --setmclk LEVEL [LEVEL ...] Set GPU Memory Clock Frequency Level(s) (requires
manual Perf level) manual Perf level)
--setpcie LEVEL [LEVEL ...] Set PCIE Clock Frequency Level(s) (requires manual --setpcie LEVEL [LEVEL ...] Set PCIE Clock Frequency Level(s) (requires manual
Perf level) Perf level)
--setslevel SCLKLEVEL SCLK SVOLT Change GPU Clock frequency (MHz) and Voltage (mV) for --setslevel SCLKLEVEL SCLK SVOLT Change GPU Clock frequency (MHz) and Voltage (mV) for
a specific Level a specific Level
--setmlevel MCLKLEVEL MCLK MVOLT Change GPU Memory clock frequency (MHz) and Voltage --setmlevel MCLKLEVEL MCLK MVOLT Change GPU Memory clock frequency (MHz) and Voltage
for (mV) a specific Level for (mV) a specific Level
--setvc POINT SCLK SVOLT Change SCLK Voltage Curve (MHz mV) for a specific --setvc POINT SCLK SVOLT Change SCLK Voltage Curve (MHz mV) for a specific
point point
--setsrange SCLKMIN SCLKMAX Set min and max SCLK speed --setsrange SCLKMIN SCLKMAX Set min and max SCLK speed
--setmrange MCLKMIN MCLKMAX Set min and max MCLK speed --setextremum min|max sclk|mclk CLK Set min/max of SCLK/MCLK speed
--setfan LEVEL Set GPU Fan Speed (Level or %) --setmrange MCLKMIN MCLKMAX Set min and max MCLK speed
--setperflevel LEVEL Set Performance Level --setfan LEVEL Set GPU Fan Speed (Level or %)
--setoverdrive % Set GPU OverDrive level (requires manual|high Perf --setperflevel LEVEL Set Performance Level
level) --setoverdrive % Set GPU OverDrive level (requires manual|high Perf
--setmemoverdrive % Set GPU Memory Overclock OverDrive level (requires level)
manual|high Perf level) --setmemoverdrive % Set GPU Memory Overclock OverDrive level (requires
--setpoweroverdrive WATTS Set the maximum GPU power using Power OverDrive in manual|high Perf level)
Watts --setpoweroverdrive WATTS Set the maximum GPU power using Power OverDrive in
--setprofile SETPROFILE Specify Power Profile level (#) or a quoted string of Watts
CUSTOM Profile attributes "# # # #..." (requires --setprofile SETPROFILE Specify Power Profile level (#) or a quoted string of
manual Perf level) CUSTOM Profile attributes "# # # #..." (requires
--setperfdeterminism SCLK Set clock frequency limit to get minimal performance manual Perf level)
variation --setperfdeterminism SCLK Set clock frequency limit to get minimal performance
--setcomputepartition {CPX,SPX,DPX,TPX,QPX,cpx,spx,dpx,tpx,qpx} Set compute partition variation
--setmemorypartition {NPS1,NPS2,NPS4,NPS8,nps1,nps2,nps4,nps8} Set memory partition --setcomputepartition {CPX,SPX,DPX,TPX,QPX,cpx,spx,dpx,tpx,qpx} Set compute partition
--rasenable BLOCK ERRTYPE Enable RAS for specified block and error type --setmemorypartition {NPS1,NPS2,NPS4,NPS8,nps1,nps2,nps4,nps8} Set memory partition
--rasdisable BLOCK ERRTYPE Disable RAS for specified block and error type --rasenable BLOCK ERRTYPE Enable RAS for specified block and error type
--rasinject BLOCK Inject RAS poison for specified block (ONLY WORKS ON --rasdisable BLOCK ERRTYPE Disable RAS for specified block and error type
UNSECURE BOARDS) --rasinject BLOCK Inject RAS poison for specified block (ONLY WORKS ON
UNSECURE BOARDS)
Reset options: Reset options:
-r, --resetclocks Reset clocks and OverDrive to default -r, --resetclocks Reset clocks and OverDrive to default
--resetfans Reset fans to automatic (driver) control --resetfans Reset fans to automatic (driver) control
--resetprofile Reset Power Profile back to default --resetprofile Reset Power Profile back to default
--resetpoweroverdrive Set the maximum GPU power back to the device deafult --resetpoweroverdrive Set the maximum GPU power back to the device deafult
state state
--resetxgmierr Reset XGMI error count --resetxgmierr Reset XGMI error count
--resetperfdeterminism Disable performance determinism --resetperfdeterminism Disable performance determinism
--resetcomputepartition Resets to boot compute partition state --resetcomputepartition Resets to boot compute partition state
--resetmemorypartition Resets to boot memory partition state --resetmemorypartition Resets to boot memory partition state
Auto-response options: Auto-response options:
--autorespond RESPONSE Response to automatically provide for all prompts --autorespond RESPONSE Response to automatically provide for all prompts
(NOT RECOMMENDED) (NOT RECOMMENDED)
Output options: Output options:
--loglevel LEVEL How much output will be printed for what program is --loglevel LEVEL How much output will be printed for what program is
doing, one of debug/info/warning/error/critical doing, one of debug/info/warning/error/critical
--json Print output in JSON format --json Print output in JSON format
--csv Print output in CSV format --csv Print output in CSV format
```
## Detailed Option Descriptions ## Detailed Option Descriptions
`--setextremum <min/max> <sclk or mclk> <value in MHz to set to>`
Provided ASIC support, users can now set a maximum or minimum sclk or mclk value through our Python CLI tool (`rocm-smi --setextremum max sclk 1500`). See example below.
```shell
$ sudo /opt/rocm/bin/rocm-smi --setextremum max sclk 2100
============================ ROCm System Management Interface ============================
******WARNING******
Operating your AMD GPU outside of official AMD specifications or outside of
factory settings, including but not limited to the conducting of overclocking,
over-volting or under-volting (including use of this interface software,
even if such software has been directly or indirectly provided by AMD or otherwise
affiliated in any way with AMD), may cause damage to your AMD GPU, system components
and/or result in system failure, as well as cause other problems.
DAMAGES CAUSED BY USE OF YOUR AMD GPU OUTSIDE OF OFFICIAL AMD SPECIFICATIONS OR
OUTSIDE OF FACTORY SETTINGS ARE NOT COVERED UNDER ANY AMD PRODUCT WARRANTY AND
MAY NOT BE COVERED BY YOUR BOARD OR SYSTEM MANUFACTURER'S WARRANTY.
Please use this utility with caution.
Do you accept these terms? [y/N] y
================================ Set Valid sclk Extremum =================================
GPU[0] : Successfully set max sclk to 2100(MHz)
GPU[1] : Successfully set max sclk to 2100(MHz)
GPU[2] : Successfully set max sclk to 2100(MHz)
GPU[3] : Successfully set max sclk to 2100(MHz)
================================== End of ROCm SMI Log ===================================
```
--setsclk/--setmclk # [# # ...]: --setsclk/--setmclk # [# # ...]:
This allows you to set a mask for the levels. For example, if a GPU has 8 clock levels, This allows you to set a mask for the levels. For example, if a GPU has 8 clock levels,
@@ -296,7 +329,7 @@ For convenience purposes, following is the output from the -h flag:
It is not possible to modify the non-CUSTOM Profiles. These are hard-coded by the kernel It is not possible to modify the non-CUSTOM Profiles. These are hard-coded by the kernel
-P, --showpower: -P, --showpower:
Show Average Graphics Package power consumption Show average or instantaneous socket graphics package power consumption
"Graphics Package" refers to the GPU plus any HBM (High-Bandwidth memory) modules, if present "Graphics Package" refers to the GPU plus any HBM (High-Bandwidth memory) modules, if present
@@ -356,8 +389,8 @@ the kernel debugfs file amdgpu_gpu_recover. Note that GPU reset will not always
manner in which the GPU is hung. manner in which the GPU is hung.
--showdriverversion: --showdriverversion:
This flag will print out the AMDGPU module version for amdgpu-pro or ROCK kernels. For other kernels, This flag will print out the AMDGPU module version for amdgpu-pro or ROCm kernels. For other kernels,
it will simply print out the name of the kernel (uname) it will simply print out the name of the kernel (`uname -r`)
--showserial: --showserial:
This flag will print out the serial number for the graphics card This flag will print out the serial number for the graphics card
@@ -424,4 +457,4 @@ The information contained herein is for informational purposes only, and is subj
AMD, the AMD Arrow logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies. AMD, the AMD Arrow logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies.
Copyright (c) 2014-2022 Advanced Micro Devices, Inc. All rights reserved. Copyright (c) 2014-2024 Advanced Micro Devices, Inc. All rights reserved.
python_smi_tools/rocm_smi.py
+1 -1
View File
@@ -3819,7 +3819,7 @@ if __name__ == '__main__':
groupDisplayPages.add_argument('--showretiredpages', help='Show retired pages', action='store_true') groupDisplayPages.add_argument('--showretiredpages', help='Show retired pages', action='store_true')
groupDisplayPages.add_argument('--showunreservablepages', help='Show unreservable pages', action='store_true') groupDisplayPages.add_argument('--showunreservablepages', help='Show unreservable pages', action='store_true')
groupDisplayHw.add_argument('-f', '--showfan', help='Show current fan speed', action='store_true') groupDisplayHw.add_argument('-f', '--showfan', help='Show current fan speed', action='store_true')
groupDisplayHw.add_argument('-P', '--showpower', help='Show current Average Graphics Package Power Consumption', groupDisplayHw.add_argument('-P', '--showpower', help='Show current average or instant socket graphics package power consumption',
action='store_true') action='store_true')
groupDisplayHw.add_argument('-t', '--showtemp', help='Show current temperature', action='store_true') groupDisplayHw.add_argument('-t', '--showtemp', help='Show current temperature', action='store_true')
groupDisplayHw.add_argument('-u', '--showuse', help='Show current GPU use', action='store_true') groupDisplayHw.add_argument('-u', '--showuse', help='Show current GPU use', action='store_true')