Submit explicit profile marker for hipEventRecord to record
timestamps. Enable explicit signal profiling if the API specifies
start and stop events.
Toggle this with env var HIP_FORCE_QUEUE_PROFILING=0
Change-Id: Iae449a63ec3ebf6c2880e65d7b1dd1031a29018f
Recently there were few OOM(out of Memory) issues poped. Adding trace will help know genuine OOM issue.
Change-Id: Ie2a56bba95cead24caf1af5c807e73c383c2f5d9
The current implementation creates default reference in the stack and assigns it to class member cuMasks_, so whenever the content of the stack changes, cuMask_ would change.
Change-Id: Iefab63c335d504b83c4ae90bd34ae76c6afb8f3c
Optimizaiton to remove extra syncs uncovered a bug with the cache
coherency layer, there runtime could lose the track of mem address
if coherency layer performed a sync.
Change-Id: I25647cfa4a4be9cdbd8577ff076a740bbdac79c8
rocm-utils is a meta-pacakge that pull in lot more pacakges than
what is required for hip.
Correct dependency to set as of today will be rocminfo.
Change-Id: I065ca74b4520e0c9ce183d82d6026554ff43e1f0
If the queue device doesn't match the device on any memory object, use the queue device from the memory object.
Change-Id: I5fdcf00494f8391574f4443332c01788b8da44ef
SWDEV-244798
If {src/dst} ptr is marked as hipMemoryTypeHost, check if the memory was
prepinned. If it was, upgrade the copy type to hipMemoryTypeDevice to
avoid extra pinning.
Change-Id: Id287ef5b14ae67dfbcf80c4caa1b08a311191948
When HIP_ENABLE_DEFERRED_LOADING=0, many global variables will be
referenced but they are not initialized in that early time. The patch
will use constexpr to initialze global constant varables in compile
time.
Change-Id: I9837c42e6bab38a894ece4e0f34b64f81b38fef0
When HIP_ENABLE_DEFERRED_LOADING=0, many global variables will be
referenced but they are not initialized in that early time. The patch
will use constexpr to initialze global constant varables in compile
time.
Change-Id: I9d538b7abc6a0ce700ec3332b97fc144db5fc1ef
COMGR_DYN_DLL is used by rocclr to load COMGR as dynamic library.
Need to remove this flag while linking COMGR as static lib by
ROCclr.
Default option is set to link COMGR as dynamic library
Change-Id: I5631a8c1d896bd4d7e5ecf5ef5fa4f631cd2244d
HIP or any ROCm component above HIP may not be calling
hsa-runtime directly. OpenCl and HIP are the two components
calling ROCclr and to bring in the transitive dependency of
thunk,ROCR,amd_comgr it is better to have the dependency
chain set correctly in the ROCclr cmake target. With this
change OpenCl or HIP should not be setting ROCR dependency
directly.
This helps to link OpenCl(libamdocl.so) link statically with
comgr,hsa,thunk.
Change-Id: I0d538b7abc6a0ce700ec3332b97fc144db5fc5ff
hipSetDevice is not used correctly to allocate on multiple devices in mGPU setup.
Due to which hipMalloc was called on the same device on multiple threads leading to out of memory issues on some devices with lesser memory.
Change-Id: I0e5b1bc028b9ecb11bd40c3a5edf715f8bd721ff
If numa lib is in building system, define ROCCLR_NUMA_SUPPORT to
support numa; otherwise, don't support numa.
Change-Id: I3848d7fdec5a3813ff1edad9b71ff04372dc0b9a
Jatin Chaudhary
Jason Tang
Julia jiang
Saleel Kudchadker
Payam
Sarbojit Sarkar
Anusha Godavarthy Surya
Chauncey Hui
German Andryeyev
Vlad Sytchenko
Aryan Salmanpour
Rahul Garg
Freddy Paul
Jason Tang
Tao Sang
kjayapra-amd
Alex Xie
Satyanvesh Dittakavi
Matt Arsenault