* [hip] Docs: Overhaul HW implementation page
* Update hardware implementation and glossary
* Update programming model
* Add performance optimization
* Split into how-to and understanding
---------
Signed-off-by: Jan Stephan <jan.stephan@amd.com>
Co-authored-by: Jan Stephan <jan.stephan@amd.com>
Co-authored-by: Julia Jiang <julia.jiang@amd.com>
* rocrtst: Updated CMakeFiles to find_package instead of hardcoded
This is to support TheROCK build environment
* rocrtst: Fix CMake to use find_package() instead of hardcoded ENV paths
Fixed CMake style issues from previos first commit's code review
* rocrtst: Fix rocrtst NUMA dependency detection to use find_package
Also added handling of missing headers
* rocrtst: Fix NUMA and hwloc detection for cross-platform builds
---------
Co-authored-by: Shweta Khatri <shweta.khatri@amd.com>
* Fixed parsing of env var lists which were overwriting the mutable env var string and polluting future parses.
* Fixed all tests to obey UT_DATATYPES and UT_REDOPS filters.
* Allow tests to bail early via `GTEST_SKIP` if UT_DATATYPES or UT_REDOPS filters give a test size of zero. This allows tests to run much faster with filters on.
* Wrapped the support checks in helper functions on `TestBed`.
* [hip-tests] Update API coverage report generator
Updates the HIP API coverage tool. It now takes
extra arguments for the location of the catch test folder
and for the working directory. This avoids issues where the output
of the executable is dependent on the path where it is being
executed from.
Also updates CmakeLists.txt to integrate seamlessly with the
hip-tests project and avoid using commands which rely on
relative paths.
* Remove double new line
* Remove Cmake option to generate coverage
Removes Cmake option to generate coverage. Instead, explicitly removes
the gen_coverage target from all (this is already the default but
doing it explicitly prevents confusion).
* SWDEV-558848 - vmm api support for rocr on windows
* Fixes to VMM handle Map/Unmap Set/Get Access
* Fix GetShareableHandle to use pointer for shareable handle
* Update os specific map/unmap memory calls
* clang format update
* Minor syntax fixes from code review
Co-authored-by: Yiannis Papadopoulos <102817138+ypapadop-amd@users.noreply.github.com>
---------
Co-authored-by: Rahul Manocha <rmanocha@amd.com>
Co-authored-by: Yiannis Papadopoulos <102817138+ypapadop-amd@users.noreply.github.com>
* Remove TCP_TCP_LATENCY_sum counter for MI300
* Remove TCP_TCP_LATENCY_sum counter which is unsupported for MI300 per register specification
* Remove VL1 Lat metric from memory chart section (block 3) for MI 300
since it uses TCP_TCP_LATENCY_sum counter which is unsupported
* Remove references to TCP_TCP_LATENCY_sum
* Update CHANGELOG
* reword changelog
* Read the ids_flags when fetching GPU info
The ids_flags contains the flags that can help identify if a GPU
is a dGPU or an APU.
* Show correct memory pool for APUs
The kernel policy for APUs will be to choose the bigger pool of
memory (GTT or VRAM) for KFD work. Adjust the policy for the monitor
and default commands to show the right memory pool when using an APU.
* Add ls statement for debugging /opt directory file naming
* Update ROCM_VERSION from 7.0.0 to 7.1.1 in SDK CI
* Update amdgpu debian package for Ubuntu in Dockerfile.ci
* disable HIP/CLR build in codeql (#2242)
---------
Co-authored-by: Venkateshwar Reddy Kandula <Venkateshwarreddy.Kandula@amd.com>
* Change rocprofiler-sdk CMake compatibility to AnyNewerVersion
Update CMake package version compatibility from SameMinorVersion to
AnyNewerVersion to allow downstream packages (like RDC) to use newer
versions of rocprofiler-sdk without requiring exact minor version match.
This fixes compatibility issues where RDC requests 1.0.0 but finds 1.1.0.
* Update projects/rocprofiler-sdk/cmake/rocprofiler_config_install.cmake
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Change rocpd and roctx CMake compatibility to SameMajorVersion
Update COMPATIBILITY setting from SameMinorVersion to SameMajorVersion
for both rocpd and roctx packages to allow compatibility across major
version boundaries.
---------
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* SWDEV-554626 - return hipErrorInvalidDeviceFunction when we can not load module
Return correct error code when modules are empty
* Match the error codes
* Revert the error code
* Add functional test for barrier_all_on_stream
* Add rocshmem_barrier_all_on_stream support for GDA and RO backends
Implements rocshmem_barrier_all_on_stream operation for
GPU Direct Access and Reverse Offload backends.
Previously, rocshmem_barrier_all_on_stream was only supported for IPC backend.
* Add functional test for rocshmem_broadcastmem_on_stream
* Add host-side rocshmem_broadcastmem_on_stream API
Implement stream-based broadcast collective operation
- Add rocshmem_broadcastmem_on_stream host API and kernel implementation
- Add functional test TeamBroadcastmemOnStreamTester with multi-stream
support and correctness verification
- Use per-workgroup contexts to avoid contention across parallel streams
API:
rocshmem_broadcastmem_on_stream(team, dest, source, nelems, pe_root, stream)
* Add functional test for rocshmem_getmem_on_stream
* Add host-side rocshmem_getmem_on_stream API
Implement stream-based point-to-point RMA get operation
- Add rocshmem_getmem_on_stream host API and kernel implementation
- Support for asynchronous getmem operations on HIP streams
- Add backend support for GDA, RO, and IPC contexts
- Use work-group collective getmem for efficient memory transfer
API:
rocshmem_getmem_on_stream(dest, source, nelems, pe, stream)
(AI Assist)
* Add host-side rocshmem_putmem_on_stream API
- Add rocshmem_putmem_on_stream for asynchronous remote writes
- Support for concurrent RMA operations on HIP streams
- Add backend support for GDA, RO, and IPC contexts
- Use work-group device collective operation
API:
rocshmem_putmem_on_stream(dest, source, bytes, pe, stream)
(AI Assist)
* Add functional test for rocshmem_putmem_on_stream
* Add host-side rocshmem_putmem_signal_on_stream API
Enables asynchronous putmem operations with signaling on HIP streams.
The implementation includes:
- Kernel wrapper rocshmem_putmem_signal_kernel
- Host interface putmem_signal_on_stream method
- Context layer support across all backends (IPC, GDA, RO)
- Public API
Function signature:
void rocshmem_putmem_signal_on_stream(void *dest, const void *source,
size_t bytes, uint64_t *sig_addr,
uint64_t signal, int sig_op,
int pe, hipStream_t stream);
* Add functional test for rocshmem_putmem_signal_on_stream
* Add host-side rocshmem_signal_wait_until_on_stream API
Enables asynchronous signal wait operations on HIP streams.
The implementation includes:
- Kernel wrapper rocshmem_signal_wait_until_kernel
- Host interface signal_wait_until_on_stream method
- Context layer support across all backends (IPC, GDA, RO)
- Native uint64_t support in wait_until API (generated from P2P_SYNC.py)
Function signature:
void rocshmem_signal_wait_until_on_stream(uint64_t *sig_addr, int cmp,
uint64_t cmp_value,
hipStream_t stream);
(AI Assist)
* Add functional test for rocshmem_signal_wait_until_on_stream
* Add documentation for stream API functions
This commit adds API documentation for the following host-side
stream functions:
- rocshmem_barrier_all_on_stream (collective routines)
- rocshmem_broadcastmem_on_stream (collective routines)
- rocshmem_getmem_on_stream (RMA operations)
- rocshmem_putmem_on_stream (RMA operations)
- rocshmem_putmem_signal_on_stream (signaling operations)
- rocshmem_signal_wait_until_on_stream (point-to-point sync)
The documentation includes function signatures, parameter descriptions,
and detailed explanations of asynchronous behavior and stream handling.
(AI Assist)
* Rename "bytes" -> "nelems"
* Add "_TEST_" to the variables used in tests
* Remove incorrect hipStreamDefault usage
hipStreamDefault is not a default stream. This is a flag.
If stream == nullptr, then just pass it to kernel. It will launch the kernel on the default stream
* Don't require powercap support
APUs don't necessarily support setting a power cap from sysfs.
Ignore failures of the file missing.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
* Show edge temperature in default output if hotspot is missing
APUs don't have a hotspot temperature, they have an edge though.
Use that.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
* Format all "power" keys as watts
There will be more power keys when APU support is added, so format
them properly.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
* Don't show power limit in output if it's invalid
APUs can't set power limit using power_cap1 interface. The limit
will be 0 and thus the UX looks weird in default output.
Only add the `/power_limit` if it's valid.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
* Unify sizes of `amdsmi_power_info_t`
Sizes are used inconsistently. This causes tools to not show
N/A when they should. Make them unified.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
---------
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
These have an external source of truth.
Also drop the non-existent hipblaslt which isn't in rocm-systems.
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
* While reusing signals, its possible we can come across a timestamp
that can contain several signals, like when profiling a graph. Reading
timestamps from all signals can make the call severely CPU bound.
Instead cache only that signal so as to avoid the overhead for critical
path.
* Let functional tests build without external MPI
* Fix error conditions when using uuid startup with internal MPI
* Do not abort if libibverbs is not found but not using GDA
* Enabled RO functional test initialized with TEST_UUID
* Reduce load time for ro backend_can_run and prevent mpilib_dlclose
crashing
* Fix case TEST_UUID=1, ROCSHMEM_BACKEND='' (autoloading gda)
Adel Johar
koushikbillakanti-amd
SakaSitharammurthy
systems-assistant[bot]
yugang-amd
shwetakhatri-amd
Geo Min
David Galiffi
Aurelien Bouteiller
German Andryeyev
Pengda Xie
corey-derochie-amd
Fábio Mestre
Yiltan
Rahul Manocha
Venkateshwar Reddy Kandula
jamessiddeley-amd
vedithal-amd
cfallows-amd
Mario Limonciello
Maisam Arif
Jason Bonnell
Dominic Widdows
Benjamin Welton
Swati Rawat
Larry Meadows
Victor Zhang
Jatin Chaudhary
Anatolii Rozanov
Dimple Prajapati
Avinash Kethineedi
Mario Limonciello
Bindhiya Kanangot Balakrishnan
SaleelK
Yazen AL Musaffar
Dmitrii