* feat: fetch OpenCL info
* feat: show opencl info in gui
* chore: move vulkan info to software page, refactor
* refactor: unify hw info page into one component
* chore: update generated-sources
* feat: show more information
* fix: disable dynamic feature in OpenCL
* Revert "fix: disable dynamic feature in OpenCL"
This reverts commit 18c9d93ed7.
* fix: add dependency
* fix: set correct dependencies
* fix: opencl package name
* fix: fedora package name
* fix: add rhel repo
* chore: update readme
* WIP: add initial intel xe/i915 support
* feat: basic gpu clock configuration
* wip i915 freq controls
* fix: applying frequency settings on i915
* feat: intel DRM info
* feat: show EUs and subslices in UI
* refactor: exclude all system-dependent info in tests with conditional compilation
* chore: ignore clippy errors on bindgen file
* fix: more test fixes
* feat: generate drm bindings for xe
* wip
* refactor: GPU controller initialization
* fixes
* chore: disable currently unused xe headers
* feat: dynamically load libdrm for intel
* chore: drop println
* Merge with master
* fix: tests
* chore: avoid error in vulkan when running tests
* feat: numerous i915 additions
* feat: show fan speed
* feat: hwmon monitoring on xe
* fix: avoid crashing history graph on empty plot with throttling data
* feat: report pstates
* doc: update readme
* chore: update API docs link
* feat: abstract GpuController
* test?
* fix: dont re-initialize nvidia gpu
* feat: more info on nvidia
* feat: core stats on nvidia
* fix: nvidia uses milliwatts not microwatts
* feat: include device info and stats responses in debug snapshot
* chore: avoid logging msg twice
* fix: correctly handle only secondary line series being present in a plot
* feat: more reporting
* feat: support for setting the power cap
* chore: avoid trying to initialize drm for non-amd gpus
* chore: avoid drawing primary plot label when only secondary values are present
* feat: hide unknown values on the info page
* feat: report cuda cores
* chore: limit power usage value accuracy
* Threaded plot render
* Better supersampler implementation
* Better to display nothing than do long freeze
* Fix plot throttling jumping around
* Further improve rendering by using filled legend
* Spawn render thread with minimum priority
* Optimize Cairo bindings
* Simplify code as we no longer need to track initial state
Signed-off-by: Alik Aslanyan <inline0@pm.me>
* Add plotters package override for opt-level 3 in release
* Immediately react to size changes of the widget, don't wait for new data
* feat: nvidia fan control
* Scale plots in GTK, instead of Cairo for Trillinear filtering, rewrite supersampling
* feat: pstate reporting
* doc: update README to mention that nvidia is now supported
* doc: add historical data to readme
* doc: improve screenshot alignemnt
* doc: add nvidia driver note
---------
Signed-off-by: Alik Aslanyan <inline0@pm.me>
Co-authored-by: Alik Aslanyan <inline0@pm.me>
libdrm was initially added as a way to get more information, but is now
used for some crucial functionality like GPU generation and VRAM type
detection. it should always be used to avoid these things not working
* feat: report git commit in version info, check commit for mismatches
* feat: use vergen to get commit info
* fix: install git in recipes
* fix: include git dir in pkger builds
* feat: add support for custom fan curves and static fan speed on RDNA3
* fix: use gpu-provided min and max fan speed values
* fix: set min temperature point too in static curve
* fix: reset fan curve when switching mode back
* fix: normalize fan speed values by taking min speed into account
* test
* Revert "test"
This reverts commit dbd37bdc0b.
* fix: dont use the upper range value as visibility trigger for oc scales
* fix visibility
* Revert "fix: normalize fan speed values by taking min speed into account"
This reverts commit 945cfd23e6.
* fix: get_adjustment_value
* doc: update README
Ilya Zlobintsev
Remigio
vanitiesandIlya Zlobintsev
Tai Lam
pastalian