Wykres commitów

30 Commity

Autor SHA1 Wiadomość Data
Bill(Shuzhou) Liu 6340176b99 Unit tests for coarse grain utilization counters
The unit tests for GFX and Memory activity counters.

Change-Id: I968dabc9ef6de9d335d7f751b290fb713b51a79c
2021-04-14 10:53:55 -04:00
Divya Shikre aaf2120117 Update performance determinism api as per the modified sysfs interface.
Signed-off-by: Divya Shikre <DivyaUday.Shikre@amd.com>
Change-Id: Ib0ec5128819644a2ff6c916da9194a7fe1dad795
2021-04-07 16:38:48 -04:00
Divya Shikre 60d0f3052f Adding Performance Determinism Mode to rocm_smi lib, CLI & gtest.
A special mode of operation to achieve minimal performance variation by letting
the user have the ability to provide the desired frequency to be set as the soft limit.
The user can control the entry and exit to the mode via rocm-smi a mechanism to
enter / exit performance determinism mode as below.

Enter performance determinism mode:
- hold a lock
- write performance_determinism to power_dpm_force_performance_level
- write input clk_freq to pp_dpm_sclk
- release lock

Exit performance determinism_mode:
- hold a lock
- write auto to power_dpm_force_performance_level
- release lock

Signed-off-by: Divya Shikre <DivyaUday.Shikre@amd.com>
Change-Id: Ia24e27954cdf1c4337ffc83d8948fbdfaf4552d2
2020-12-02 11:11:00 -05:00
Divya Shikre f397cba414 Adding gtest for gpu metrics read
Signed-off-by: Divya Shikre <DivyaUday.Shikre@amd.com>
Change-Id: I66edb15c8b7380f3427822b33e845202bfac7a2b
2020-10-08 13:37:47 -04:00
Chris Freehill efc9b7658c Make verbosity level 0 completely quiet
Also, support --iterations flag for certain functions that will
likely be repeated frequently.

Change-Id: I7ed76835001b5cbca30042d6bf26484258c7b9a6
2020-06-17 21:26:12 -05:00
Divya Shikre 2805ed16a4 Adding current voltage feature & gtest.
Signed-off-by: Divya Shikre <DivyaUday.Shikre@amd.com>
Change-Id: Ic555a3af265e603419e2875d1989a366abc82596
2020-06-16 11:48:56 -04:00
Mike Li c7d349183a Add functions that are used to query Hardware topology.
Change-Id: I0f4cd02b237bde4d6dccfb0e83e65376ecb1cfaa
Signed-off-by: Mike Li <Tianxinmike.Li@amd.com>
2020-05-18 12:37:27 -04:00
Chris Freehill 8e03d10035 Add ref counting for rsmi init and shutdown
Also, clean lint from kfd_ioctl.h file.

Change-Id: I5a2ae127ab6ab6676a1b075ed10858d0ebfe13c1
2020-05-11 15:57:42 -05:00
Chris Freehill 2235ede34c Add event notification API
Change-Id: Ib6e8efbe6cdefaa7de1f74bd26993e9b4b011649
2020-05-06 14:07:25 -05:00
Chris Freehill 1c9ef44398 Add checking for no-longer-existing process in test
When getting process information for a process, it's possible
that between the time the process ID was discovered and when
we attempt to collect data for that process, that the process
ended. This change is meant to handle that in the test case.

* Also, fix compile warning by removing unused variable.

Change-Id: I62f9a84a63548c856f0661fef15b7d248a330c05
2020-04-10 08:51:44 -05:00
Chris Freehill f8b57c3b16 Add device mutual exclusion tests and related fixes
* Added a new test to verify mutual exclusion of access to device
  resources
* Added some missing acquiring of mutexes to some RSMI calls, as
  well as try-catch blocks.

Change-Id: I87aac009878a0b2d1f975e1d5b794d887bb23ff9
2020-04-08 15:05:11 -05:00
Chris Freehill 551b15182b Add functions that tell what capabilities are supported
The new functions added in this commit allow a caller to tell up
front what functions, function variants and monitors are
supported.

Also,
* fixed a few documentation/formatting issues
* fixed a process_info test issue

Change-Id: I2184ab1a4a6898f847e791f273e2185d556e78e9
2019-09-23 13:30:47 -05:00
Chris Freehill 73c54e1fd0 Add support for rsmi_dev_memory_reserved_pages_get()
Also, don't return an error for empty sysfs files. The reserved memory
page file will often have no lines. We don't want it to appear that
this function is not supported if the file is empty.

Change-Id: I1d28bb184ea587bb578fe71dd75adc2a812d09a8
2019-08-06 11:42:03 -05:00
Chris Freehill 557e1f5704 Add xgmi error_status and error_reset functions
Also, comment corrections and added check for invalid arguments

Change-Id: I891cbf9b37bfda629914a008811b840323872c02
2019-07-09 09:55:05 -04:00
Chris Freehill 9b93cbe21d Add initial support for getting process information
Added implementation of and tests for
rsmi_dev_compute_process_info_by_pid_get() and
rsmi_dev_compute_process_info_get()

Change-Id: I4c4f5f39fe6701da37916c9ad41449b5d35ac7af
2019-07-03 20:01:43 -05:00
Chris Freehill 1c5e090507 Add rsmi_dev_memory_busy_percent_get()
Change-Id: Ide683b6c72870af547331f4502c5bb8c445d61b5
2019-06-25 19:09:13 -05:00
Chris Freehill ea26baec20 Event counter support
XGMI related events are supported

Change-Id: If17036fe890c8be45da3654353599821b5828c14
2019-06-24 17:40:01 -05:00
Chris Freehill 908f07cb3b Revert "Event counter support"
This reverts commit 075833e9a5.
2019-06-21 22:07:40 -05:00
Chris Freehill 075833e9a5 Event counter support
XGMI related events are supported

Change-Id: Ic99d5a1847e8d28b22ad0b61cb9ea206eb878708
2019-06-21 18:27:50 -05:00
Chris Freehill a0817d6b13 By default, only consider AMD GPU's in RSMI device list (#36)
With newly added initialization parameters that can be
passed to rsmi_init(), you can tell RSMI to consider other
devices.

Also:
-fixed incorrect header file name that would break in C
builds
-modified rsmi_init() and rsmi_shut_down() to reinitialize and
clear static structures
2019-05-09 20:45:54 -05:00
Chris Freehill cbdfac7bdc Added new id and id name string look up functions
Also, updated docs with typo corrections and a new section
2019-03-15 16:21:37 -05:00
Chris Freehill ddd292f1b5 Add rsmi_dev_memory_total_get()and rsmi_dev_memory_usage_get() 2019-03-04 18:26:11 -06:00
Chris Freehill fb5f41fc10 Added rsmi_dev_error_count_get() 2019-03-01 16:33:11 -06:00
Chris Freehill 18ce553dce Add rsmi_dev_pci_throughput_get() 2019-02-27 15:10:26 -06:00
Chris Freehill 021f13a68f Add VBIOS version get function
Also, consolidate "get version" type function tests into 1 test.
2019-02-24 11:01:18 -06:00
Chris Freehill f3fa9a036c Add get_version test; remove sanity test
Also, don't fail pcie bandwidth test when the pp_dpm_pcie file
does not correctly show the current bandwidth.
2019-02-12 18:07:26 -06:00
Chris Freehill 17bf80dcb2 Break down monolithic test into many smaller tests
Also, added boot up default power profile, and modified to
accomodate new profile format
2019-02-11 22:53:24 -06:00
Chris Freehill dd450e963c Add dont_fail option to not fail entire test on a single failure 2019-02-09 12:18:49 -06:00
Chris Freehill 861c2c2e33 Add rsmi_dev_busy_percent_get()
Also: correct some comments, ifdef out unused code
2018-11-12 17:25:14 -06:00
Chris Freehill b9c6a5a443 Add gtest based testing 2018-10-24 11:57:03 -05:00