hipEventRecord should always create a new marker so it can track work going on at the time the API is called.
Change-Id: I10ce98044be894fbacab8798441ec3d3f2753b93
[ROCm/clr commit: eb132ccef8]
Those APIs should look at the device associated with the stream first.
If that stream is null then get the current device ID.
Change-Id: Iedde1d1644818ba64f128b988f0bd9674f5b8ad6
[ROCm/clr commit: 8f5a70a150]
The issues of the following functions have been fixed.
hipModuleLoad: Make Hip-Vdi runtime able to read code object module
generated by Hcc compiler.
hipLaunchKernel: Use introspect method to find function if it cannot
be found from platform state instance.
Change-Id: Id740e5a96614ec6a0b6c704f8f74600bfdc4983e
[ROCm/clr commit: 62ef029288]
Fix the install directory for libamdhip64.so and create libhiprtc.so symlink
Change-Id: Id731bfa18bb3585c3f9e3ae6697b4f4687c49195
[ROCm/clr commit: c98a17c80f]
- The known target checking should skip `gfx000` as well as it won't be
used in real compilation command formation. The avoid generating
annoying warning on `gfx000`.
[ROCm/clr commit: 59bd608bfe]
SWDEV-125823 - Fix the build issue due to API interface change.
- `hipOccupancyMaxActiveBlocksPerMultiprocessor` interface is revised
and the runtime needs updating.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#60 edit
Change-Id: Ia7901b0dbbfd37977ce4adf2ae1a821aba0ac044
[ROCm/clr commit: 657734689d]
* Device texture functions should not normalize the sampled pixel. This is already done by HW.
* Add support to use h/w capability for normalized float data convertion for driver API's
Co-authored-by: ansurya <50609411+ansurya@users.noreply.github.com>
[ROCm/clr commit: fa363ae721]
* Fix bug in LaunchKernel test
Instead of passing the address of the gpu buffer, pass the address
of the pointer that holds the address of the gpu buffer
* Fix hipLaunchKernel's kernarg buffer construction.
The hipLaunchKernel implementation should rely on ihipModuleLaunchKernel
to construct the kernarg buffer correctly based on kernel metadata.
* Fix a bug in get_functions where the Kernel_descriptor wasn't constructed with the correct kernarg layout information.
* Fix a bug in kernarg layout parsing dealing with kernel without any arg
* teach ihipModuleLaunchKernel to handle kernel without any arg
* Add a more interesting test
[ROCm/clr commit: 14e235378f]
* Added support to disable test based on HIP Runtime(HCC/VDI) and Compiler(hcc/clang)
* Updated README.md, HIT.cmake for new options EXCLUDE_HIP_RUNTIME and EXCLUDE_HIP_COMPILER
[ROCm/clr commit: 696c85aec9]
* Add missing texturePitchAlignment member to the hipDeviceProp_t struct.
* Add missing hipDeviceAttributeTexturePitchAlignment enumerator to the hipDeviceAttribute_t enum.
* Initialize texturePitchAlignment to 256. This works for gfx9+, but is technically overaligned in most cases for pre-gfx9.
* Add the texturePitchAlignment property to the NVCC path.
[ROCm/clr commit: 6da0c82b78]
Payam
Christophe Paquot
Tao Sang
Aaron Enye Shi
Laurent Morichetti
Mark Searles
Jatin Chaudhary
Maneesh Gupta
Alex Voicu
mhbliao
Saleel Kudchadker
Michael LIAO
Siu Chi Chan
vsytch
lthakur
Sarbojit2019
Daniil Fukalov
Rahul Garg
kpyzhov
ansurya
gandryey
Evgeny Mankov
jujiang-del
satyanveshd