- Support graph with different types of nodes with single
branch when DEBUG_CLR_GRAPH_PACKET_CAPTURE flag is enabled
Change-Id: I149a8629769cd0d5849ffefb04f1352668a685b6
[ROCm/clr commit: 38d2c56784]
- Address corner cases that can arise with the new
hipMemcpyDeviceToDeviceNoCU enum
- Better log
Change-Id: I6035b901f8d616741054b7a5ff4f67956329ac57
[ROCm/clr commit: 5662d4037c]
Addresses the below scenarios like parameters mismatch
in memcopy node, difference in the count of nodes, difference
in the dependencies of the nodes.
Change-Id: I31c6516fb27cc1007809f1b50306fdb0c2356ccc
[ROCm/clr commit: f16d336e32]
When kernel function expects no parameters no error should be returned
if both kernelParams and extra arguments are set to null.
Change-Id: I5941bcc400b6fb380e623bdae0233ae3e4f73815
[ROCm/clr commit: 315082e554]
The change enables VM support in graphs on Windows. That allows
to avoid caching of all allocations at the cost of map/unmap
overhead during memory create/destroy.
Change-Id: I792be00fba099e5e5d3cd44a963e1dfd6976a86d
[ROCm/clr commit: 04b696abee]
GraphMemcpyNodeSetParamsFrom/ToSymbol APIs neew to check device id for
original src/dst is same as what is passed in while set.
Change-Id: If0b610808223dce9115562bb5e9b31c8eaa2df22
[ROCm/clr commit: b6aa27d4a3]
Avoid syncing blocking streams with the default stream,
since that introduces extra command dependencies and
doesn't allow to destroy memory after last submission
Change-Id: I618e9bd2091c4cf9157125612d8c4759030c5a80
[ROCm/clr commit: 1e88d2c52f]
Hardcoding check for accessPolicyMaxWindowSize
since its returned from Hardware thats not available yet
Change-Id: I3a09fc9ace3027c7d5533bbbc7da5e18f43591ec
[ROCm/clr commit: c603d5004a]
- if graph is already instantiated there cant be a second instance in
case nodes are free or alloc
Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: I51e2ca8ade24799da96ed126a26c5ea3bad6f452
[ROCm/clr commit: f6c36be185]
- for to/from symbols nodes the correct kind is not showing
Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: Idf203976b71fcca67b2dd989de0daba4d0554b6d
[ROCm/clr commit: ae631aaa32]
- Use correct accessor to fetch memory objects. This checks the svm map
and arena maps
Change-Id: I84515330bb530cfe2b39abf30e1e659938f06806
[ROCm/clr commit: 8028d327e9]
- test fails intermittently because object is not cleared correctly
Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: I88daf3dc08bb83d6d3f047ff48a63c8f856fb0bf
[ROCm/clr commit: a69db9b426]
MemPool was designed to use hip::Stream, but graph implementation uses
amd::HostQueue. Hence switch graph to hip::Stream management.
Change-Id: Ia319389de45e4c3c6043d17473279a6f27a13140
[ROCm/clr commit: 0fe03b6494]
- remove node dependency before checking parents
- reduce edge level acording to new value of node
Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: Id4bff1684f7e0b42beeebc4d2e009bfdb507fb5f
[ROCm/clr commit: 34d087da93]
Add memory allocation support in graph. Current implementation uses
cache from mempool to hold the allocations which belong to the graph.
Also the resource tracking is disabled at this moment because mempool
operates with hip::Stream objects, but graph has execution with
amd::HostQueue objects.
Change-Id: I54fe3250126d24f5a26ada975f37d429bb4ef17b
[ROCm/clr commit: eef47ca24a]
sdashmiz
Anusha GodavarthySurya
Saleel Kudchadker
Satyanvesh Dittakavi
Ioannis Assiouras
Jaydeep Patel
German
Ajay
Sourabh Betigeri
Jiabao Xie
pghafari
Tao Sang