You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
-**`amd-smi set --power-cap` now requires sepcification of the power cap type**.
83
+
- Command now takes the form: `amd-smi set --power-cap <power-cap-type> <new-cap>`
84
+
- Acceptable power cap types are "ppt0" and "ppt1"
85
+
86
+
```console
87
+
$ sudo amd-smi set --power-cap ppt1 1150
88
+
GPU: 0
89
+
POWERCAP: Successfully set ppt1 power cap to 1150W
90
+
...
91
+
```
73
92
74
-
```console
75
-
$ sudo amd-smi set --power-cap ppt1 1150
76
-
GPU: 0
77
-
POWERCAP: Successfully set PPT1 power cap to 1150W
78
-
...
79
-
```
80
93
-**`amd-smi reset --power-cap` will attempt to reset both power caps**.
81
94
- When using the reset command, both PPT0 and PPT1 power caps will be reset to their default values. If a device only has PPT0, then only PPT0 will be reset.
82
95
Ex.
@@ -1402,7 +1415,7 @@ Functions affected by struct change are:
1402
1415
- **Corrected CLI CPU argument name**.
1403
1416
- `--cpu-pwr-svi-telemtry-rails` to `--cpu-pwr-svi-telemetry-rails`
1404
1417
1405
-
- **Added amdgpu driver version and amd_hsmp driver version to `amd-smi version` command**.
1418
+
- **Added amdgpu driver version and amd_hsmp driver version to `amd-smi version` command**.
1406
1419
- The `amd-smi version` command can now also display the amdgpu driver version using the `-g` flag.
1407
1420
- The amd_hsmp driver version can also be displayed using the `-c` flag.
1408
1421
- The new default for the `version` command is to display all the version information, including both amdgpu and amd_hsmp driver versions.
logging.info('Unable to detect any CPU devices, check amd_hsmp version and module status (sudo modprobe amd_hsmp)')
99
+
logging.info('Unable to detect any CPU devices, check amd_hsmp (or) hsmp_acpi version and module status (sudo modprobe amd_hsmp (or) sudo modprobe hsmp_acpi)')
# No CPU's found post amd_hsmp driver initialization
115
-
logging.error('Unable to detect any CPU devices, check amd_hsmp version and module status (sudo modprobe amd_hsmp)')
115
+
logging.error('Unable to detect any CPU devices, check amd_hsmp (or) hsmp_acpi version and module status (sudo modprobe amd_hsmp (or) sudo modprobe hsmp_acpi)')
To run the AMD SMI library, the `amdgpu` driver and the `amd_hsmp` driver need
41
-
to be installed. Optionally, `libdrm` can be installed to query firmware
40
+
To run the AMD SMI library, the `amdgpu` driver and the `amd_hsmp` or `hsmp_acpi` driver need to be installed. Optionally, `libdrm` can be installed to query firmware
Copy file name to clipboardExpand all lines: docs/reference/amdsmi-py-api.md
+14-5Lines changed: 14 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1310,12 +1310,8 @@ Input parameters:
1310
1310
*`cursor` the zero based index at which to start retrieving cper entries; default value is 0; for example, if there are 10 cper entries available, then with a cursor value of 8, it will retrieve the last two cper entries only
1311
1311
1312
1312
Output: Dictionary with fields, updated cursor, and a dictionary of the cper_data, status_code
1313
-
status_code:
1314
-
AMDSMI_STATUS_SUCCESS: If all entries were retrieved successfully
1315
-
AMDSMI_STATUS_MORE_DATA: If some of the entries were retrieved and:
1316
-
* A subsequent call to the API with the updated cursor will result in the fetching the next batch of entries, or
1317
-
* Increasing the input buffer_size will allow more entries to be fetched with the same cursor
1318
1313
1314
+
Output1: Dictionary with fields
1319
1315
Field | Description
1320
1316
---|---
1321
1317
`error_severity` | The severity of the CPER error ex: `non_fatal_uncorrected`, `fatal`, `non_fatal_corrected`. |
@@ -1326,12 +1322,25 @@ Field | Description
1326
1322
`signature_end` | A marker value (typically `0xFFFFFFFF`) confirming the integrity of the signature. |
1327
1323
`sec_cnt` | The count of sections included in the CPER entry. |
1328
1324
`record_length` | The total length in bytes of the CPER entry. |
1325
+
`serial_number` | The product serial number. Exists in raw entries in C++ API |
1329
1326
`platform_id` | A character array identifying the GPU or platform. |
1330
1327
`creator_id` | A character array indicating the creator of the CPER entry. |
1331
1328
`record_id` | A unique identifier for the CPER entry. |
1332
1329
`flags` | Reserved flags related to the CPER entry. |
1333
1330
`persistence_info` | Reserved information related to persistence. |
1334
1331
1332
+
Output2: Updated cursor (int type)
1333
+
* Cursor is the index of the next cper entry in the GPU ring buffer. For example, if 10 entries were fetched successfully, the value of cursor will be 11 upon return from the API. Subsequent call to the API with cursor value of 11 should fetch the next entry
1334
+
1335
+
Output3: A list of dictionaries, each dictionary containing the CPER record and its size:
1336
+
* {"bytes": <rawbytes>, "size": <numberofbytes>}
1337
+
1338
+
Output4: status_code
1339
+
AMDSMI_STATUS_SUCCESS: If all entries were retrieved successfully
1340
+
AMDSMI_STATUS_MORE_DATA: If some of the entries were retrieved and:
1341
+
* A subsequent call to the API with the updated cursor will result in the fetching the next batch of entries, or
1342
+
* Increasing the input buffer_size will allow more entries to be fetched with the same cursor
1343
+
1335
1344
Exceptions that can be thrown by `amdsmi_get_gpu_cper_entries` function:
0 commit comments