* In a scenario where kernel is launched with hipExtLaunchKernelGGL and stop event is used, hipGraphInstantiate leaks. Since stop event is used, profiling is enabled and Timestamp (ReferencedCountedObject) is created, but it doesn't get released.
* The idea behind this solution is that profiling should be disabled when command is captured, hence the timestamp should not be created. Because information about capturing isn't available when kernel command is created, packet capturing state is used to determine whether to create a timestamp or not.
Change-Id: Ia23adac4592ded4fb5e236acf99e12e729f63692
[ROCm/clr commit: da5f1a6146]
Gfx12 has 16 bits for grid dim Y/Z. Detect gfxIp and return error if dim y/z > 16 bits
Change-Id: I43dd14affc9e4073d0b1232e7523967f0180fa31
[ROCm/clr commit: 0a918c8f96]
Although unpinned copies require synchronizations
in HIP, runtime can avoid syncs for H2D copies with
a staging buffer
Change-Id: If2203c6bc0cbd89742823688dc8e89e9acd873b2
[ROCm/clr commit: 29cc678d8d]
This reverts commit 2e7581a69a.
Changing the error code is considered as a breaking change,
so it should be done in major releases only.
The other reason for reverting the commit is that this change itself
is incorrect. Cuda behaves in the same way as hip when
pResDesc or pTexDesc are nullptr.
Change-Id: I3abee6b79279b81ab01c7f8466c7f8e3776c4109
[ROCm/clr commit: cfdc9dfc36]
1) Child Graph nodes need to have parent graph dependencies in waitlist.
2) Marker is placed on base stream with parent graph waitlist
Change-Id: Iec65a0171ea387be05b0733abcc708fb630e4be4
[ROCm/clr commit: 4d1ded9eaf]
The new set tracks only the queues that have a command
submitted to them. This allows for fast iteration
in waitActiveStreams.
Change-Id: I2c832eefa01280d9a87a5f57874d36d2e9441de7
[ROCm/clr commit: bcc545e6b8]
The variable is already set as cache, so that user can override.
But the hard coded setting is preventing override. Removed the same
Change-Id: I2aecc18ce4f1d1b523ba267ef1c8ef4ea1168d9c
[ROCm/clr commit: 4d0b815d06]
1) currently cpu wait is set to true, which makes the host wait for last
command in queue to finish even if the kernel execution has already
finished causing delay in device sync call.
2) device sync only needs to await completion when hw event
is not ready.
Change-Id: I91e3e89d39a1193ae06abac822cea8ae651493a5
[ROCm/clr commit: eb1089593e]
1) Since g_devices is not initialized when stream_per_thread constructor
is called on windows, m_streams is empty when hipDeviceReset is called.
2) clear_spt tries to access empty vector causing segfaults in
hipDeviceReset call.
3) on linux ROCCLR_INIT_PRIORITY makes sure that g_devices is initialized
first before tls constructor creates stream_per_thread object.
Change-Id: Ib2ba643d1278d820287ea3b242ed0878d7529165
[ROCm/clr commit: 450eca293b]
The amdgpu-arch tool is not supported for static build.
This commit adds changes to detect the build type during
cmake config and use the rocm_agent_enumerator for static build.
Change-Id: I8a295e01f54075507390ef540f16b28bb20237a9
[ROCm/clr commit: a02888af58]
This change adds a new HIP API `hipExtHostAlloc` which preserves
the functionality of `hipHostMalloc`.
Change-Id: I13504c6fc13465ddd7aed329795bb4f2fef1baff
[ROCm/clr commit: 2c84211b58]
Integration into pytorch pointed out some issues, value narrowing, to
fix this we are now using unions. Also removed check for -munsafe*
compiler flag. The check is now just on builtin detection.
Change-Id: I49364503fa429bd862952f9b29879072afa6d553
[ROCm/clr commit: bb52d9ed62]
SPT is destroyed with hipDeviceReset(). If a
stream is created right after reset, the same
object id could be reused. Later SPT destructor
incorrectly verifies that the stream is valid
referring to the reused object id causing the
corruption.
Change-Id: I3b1f7ffdf8bab874dca7b8fde22318162997b8f6
[ROCm/clr commit: f6a68b3c2e]
This change adds fixes in optimized multistream path for childGraph uses cases.
1) For childgraph nodes, rely on runNodes() only to process
the childgraph and skip calls to createCommand and enqueueCommands.
This ensures that the start/end markers are enqueued correctly
with respect to the childGraph commands.
In addition, the runNodes() for the childgraph should be called after
the dependency walkthrough to make sure that the subgraph is executed once.
2) Nodes with no outgoing edges should be marked
as a leafs regardless of which stream they are assigned to.
This is to ensure that marker dependencies from nodes
that run on non-zero stream to subgraph leafs that run on zero stream
are still set up correctly.
Change-Id: I4a5f4f3b0e0d01e515cdcb045b46c2798f291255
[ROCm/clr commit: 464b99373b]
When source or destination pitch is set to zero in hip_Memcpy2D struct
it should default to WidthInBytes + [src/dst]XInBytes
Change-Id: Id57b53cab40ba72ced231258da9356554c4868c3
[ROCm/clr commit: 7a1e818c82]
- Fixes -0.0 and +0.0 comparison. For atomicMax if the value on
address is -0.0 and on val is +0.0, gfx90a's unsafe atomics will swap
them. This behavior should be consistent with cas loop as well.
- _system variants of atomicMax and atomicMin are resulting in
incorrect output. Updated these to use the similar implementation as
atomicMax and atomicMin.
Change-Id: I20df36ee29ae0434a6b564f2ba71193fe41cfa59
[ROCm/clr commit: d69cc35750]
Satyanvesh Dittakavi
Jaydeep Patel
Branislav Brzak
Anusha GodavarthySurya
Vladana Stojiljkovic
Rahul Manocha
Jonathan R. Madsen
pghafari
Jatin Chaudhary
German Andryeyev
Saleel Kudchadker
kjayapra-amd
Marko Arandjelovic
Ioannis Assiouras
Ranjith Ramakrishnan
Ajay
Julia Jiang
victzhan
Tao Sang
Sameer Sahasrabuddhe
Sourabh Betigeri