Changes:
- Added rsmi_dev_partition_id_get() -> uses fallback described
below for devices which support partition updates.
- Updated/added to tests for partitions to reflect these changes.
Due to driver changes in KFD, some devices may report bits [31:28] or [2:0].
bits [63:32] = domain
bits [31:28] = partition id
bits [27:16] = reserved
bits [15:8] = Bus
bits [7:3] = Device
bits [2:0] = Function (partition id maybe in bits [2:0]) <-- Fallback for non SPX modes
Change-Id: Ia5641cfb8dbe2d1bff52f8eb81d5a159954528d3
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
The provides tag is required when the package provides a virtual package.
Package name along with version will be provided by default and the provides tag is not required for this.
Using the tag for providing the name, but without version was resulting in package upgrade issues.
Change-Id: I74506d8c3bbd75d028bcdc03525c29541dce2b4c
This patch fixes error:
error: assignment of member 'dismiss_' in read-only object
Reported by kind Gentoo folks:
<https://bugs.gentoo.org/918709>
Change-Id: I7cc593043e97402afd85593c528ace86952b1350
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
Updates the license to MIT
Code changes related to the following: None
Change-Id: I62d0a5f02a2d5e58c1952337dff54892793c16cf
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Checks returned error by rsmi_dev_od_volt_info_get() before assert
Code changes related to the following:
* Unit tests
Change-Id: Icc0f329e35992aae19f07243024521181467bcd3
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
When discover the amdgpu, if the assigned numbers are not consecutive,
not all GPU can be discovered. The code is change to discover the
GPU based on max card number.
Change-Id: I8b6a8b49594d6a54c7feb2645bedb83dc5c1b4cc
Fixes reading pp_od_clk_voltage new variable format and size.
Code changes related to the following:
* get_od_clk_volt_info()
* get_od_clk_volt_curve_regions()
* Unit tests
* CLI options restored: --showclkvolt, --showvc, --showvoltagerange, --setvc
* Rework: 48ddd9ab
* Bump CLI version
* CHANGELOG.md
Change-Id: I817ca224de923fdaa992df84592d63b4d5a12b22
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
This commit aligns the rsmiBindings.py.in file's
"notification_type_names" & "rsmi_evt_notification_type_t" with
those found in the rsmiBindings.py file.
Change-Id: I67f36606c505992fb98495651310bd70a1755033
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Fixes reading pp_od_clk_voltage new variable format and size.
Code changes related to the following:
* get_od_clk_volt_info()
* get_od_clk_volt_curve_regions()
* Unit tests
* CLI options removed: --showclkvolt, --showvc, --showvoltagerange, --setvc
Change-Id: Ieedb845eeadcea2f2e447ec576c253ad2a814176
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
This patch adds 'ring hang' enums to ROCM SMI LIB.
This event type name is KFD_SMI_EVENT_RING_HANG.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I9b886eb1fc027f03bcca1e5d1a89a2a186b64bf5
The environment variable RSMI_MUTEX_THREAD_ONLY=1 to enable thread only mutex.
The RSMI_INIT_FLAG_THRAD_ONLY_MUTEX can also be pass to rsmi_init()
to enable thread only mutex.
Change-Id: I2d9844039b774e386f03bb9bb130d8c342504ea6
Drops checks that are invalid with the new pp_od_clk_voltage format
Code changes related to the following:
* get_od_clk_volt_info()
* get_od_clk_volt_curve_regions()
Change-Id: I5ebe23aa0ed4ea77d5ab5a94ce34ad9b1b51281f
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Fixes TestMeasureApiExecutionTime test fails
Code changes related to the following:
* Unit tests
Change-Id: I6223078f219448deb6bfbd78edae371a5a4cf03c
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
* Updates:
- [CLI] Updated --showmemuse:
-> Add VRAM%, provide better context as "GPU Allocated Memory (VRAM%)"
-> Update "GPU memory use (%)" as
"GPU Memory Read/Write Activity(%)"
- [CLI] Updated --showmaxpower and rocm-smi (no arg)
-> Rounding was inconsistent with values past decimal.
This provides the floor value of the device
Change-Id: Ib76dea2cb8483a1d7f53df675b0a94d8d01c81b9
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
After the dead process is detected, pthread_mutex_consistent() will
be called. After that, the pthread_mutex_unlock() should also be
called to unlock it: "It is the responsibility of the application to
recover the state so it can be reused."
Change-Id: I45d3e2e68c3b06779f3acb1e908dbec0c6a39297
* Updates:
- [CHANGELOG.md] Provide 6.1 and 6.0 changes
- [README.md] Update readme with relavant changes
- [CLI] Updated --showpower to expand on types of power provided to users
Change-Id: Ic653cc81f80b7973654e2c23e1ab70567b930aa7
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
* Updates:
- [CLI] rocm-smi (no arg) and --showhw:
Now displays 'ID'/'PARTITION ID' from the pcie_id identifier
Helps users identify which partition # the device is
Information provided by KFD
Note: partition_id of 0, means a primary node (AKA root node),
ex. ASICs which do not have partitioning support will show 0
- [API] Fix partitions nodes which do not enumerate with domain:
Adding kfd's domain, allows ASICs which have domains
to enumerate in proper order.
Full pcie_id / bdf propagates to all partition nodes.
- [API] Update rsmi_dev_pci_id_get() to allow users to extract
partition_id from device
- [CLI] Added fix for devices which have modprobe failure,
but DRM does not come up properly. Even though driver shows
initialization was successful.
- [API/Utils] Overloaded print_int_as_hex() template:
Now accepts bitsize, and prints in smallest byte size
possible. Note: bitsize of < 8, please just print as decimial.
Change-Id: Ib0c6f73b2b9c9fea29442a39a669c432874382d8
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
- [CLI] Rounded VRAM output on CLI, no diffrence in output
- [python API] Fixed initializing calls which reuse initializeRsmi()
calls - now we set a global reference to rocmsmi to use
throughout API calls (see error below)
Traceback (most recent call last):
File "/home/charpoag/rocmsmi_pythonapi.py", line 9, in <module>
rocm_smi.initializeRsmi()
File "/opt/rocm/libexec/rocm_smi/rocm_smi.py", line 3531, in initializeRsmi
ret_init = rocmsmi.rsmi_init(0)
NameError: name 'rocmsmi' is not defined
Change-Id: I0eff3b8a432abf6d4344a02b9f638e1191c51a19
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Checks returned error by get_gpu_pci_bandwith() before assert
Code changes related to the following:
* Unit tests
Change-Id: Ia0fe64f168711147c5e66c7917cf633be40dee9f
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Checks and forces rereading gpu metrics unconditionally
Code changes related to the following:
* Device::dev_log_gpu_metrics()
* Examples
* Unit tests
Change-Id: Ic1c4f34a39f2bf197263f80ddbb84da26345807d
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Cleans up individual gpu metric APIs which will be implemented according to 'unified-headers' standards
Code changes related to the following:
* 'rsmi_dev_metrics_' APIs
* Functional tests
* Examples
Change-Id: I7d562a95889361ee6f8f7588f8a790f42c8eb262
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Updated:
* [CLI] Fixed vram % - printf style formatting causes many data errors
This fix updates to the recommended way of outputting formatted data.
https://docs.python.org/3/library/stdtypes.html#printf-style-string-formatting
* [API/CLI] Added gpu_id / GUID from kfd (rsmi_dev_guid_get)
-> CLI name: "GUID"
-> ROCm SMI calls: no arg, -i, --showhw, --showproduct
* [API/CLI] Added node_id from kfd (rsmi_dev_node_get)
-> CLI name: "Node"
-> ROCm SMI calls: no arg, --showhw, --showproduct
* [CLI] Added target gfx version from kfd
-> CLI name: "GFX Version" or "GFX VER"
-> ROCm SMI calls: --showhw, --showproduct
* [CLI] Base ROCm CLI
-> Removed - stacked id formatting:
This is to simplify identifiers helpful to users.
More identifiers can be found on -i --showhw, --showproduct
* [CLI] Update -i, --showhw, --showproduct, w/out arg
-> Card ID/DID/Model/SKU/VBIOS:
All unsupported values now display "N/A" instead
of "unknown" or "unsupported"
* [CLI] Showhw now expands data based on content
Change-Id: Ifb8586f9f545892b8a5aa7903608273cdd77e075
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
On some systems [rocm-smi --showpids] reports
get_compute_process_info_by_pid, Not supported on the given system
[PID] [PROCESS NAME] 1 UNKNOWN UNKNOWN UNKNOWN
get_compute_process_info_by_pid fails because cu_occupancy debugfs method
is not provided on some graphics cards and GFX revisions by design
Proposing a change to return success status when only cu_occupancy debugfs method
is not found and provide cu_occupancy invalidation value to mark only
this parameter as UNKNOWN
Change-Id: Iae37070d9bd19483b4e6c8ee24c7d9a4c92f00d7
Signed-off-by: Vladimir Stempen <Vladimir.Stempen@amd.com>
Reviewed-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
In addition to be able to set clock range, new setextremum option
is added to set only min/max clock as sometimes one of them may
not be supported.
Change-Id: I7c91ba308f3fc6c78efc88117509c515d403a6cb
Updates:
- [CLI] Switching to use generic rsmi_dev_power_get()
this is a backwards compatible function to
retrieve power values. More consistent than
previous fixes.
- [API] Update API for rsmi_dev_power_get()
Now provides @depricated for this function.
Providing notes on newer ASICS only support
current socket power, where as previous
ASICS only provided average power.
Change-Id: I34da0e925cf0b6c669bdd801b017f33f3b3ee86a
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Updates:
- [API] rsmi_dev_target_graphics_version_get, takes
reported value from KFD -> parses into human-readable
values. If device does not support, returns MAX UINT64
value and RSMI_STATUS_NOT_SUPPORTED.
Otherwise, puts into base10 format removing
extra 0's + putting in correct format. If user
provides nullptr, returning RSMI_STATUS_INVALID_ARGS.
- [Test/Example] sys_info_read updated to include
new rsmi_dev_target_graphics_version_get tests
Change-Id: I50f94e06b8733a5dec2eb08f284b44927f36abcd
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
The current code assume err_count sysfs only have 2 lines, which is
changed for umc_err_count by adding extra line for defer errors.
The code is changed to relax such check.
Change-Id: I1c469555a5d460d7bc4f4926245646c09c6a2056
Charis Poag
Bill(Shuzhou) Liu
Ranjith Ramakrishnan
Galantsev, Dmitrii
Sam Wu
Roopa Malavally
Joseph Macaranas
Oliveira, Daniel
Maisam Arif
Ori Messinger
WhiskyAKM
Junyi Hou
Daniel Martinez
David Galiffi
Istvan Kiss
Vladimir Stempen