-
Notifications
You must be signed in to change notification settings - Fork 34
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Graph fixes + other stuff #434
base: main
Are you sure you want to change the base?
Commits on May 12, 2023
-
Update OpenCL headers in include/CL
this is required for cl_khr_command_buffer
Configuration menu - View commit details
-
Copy full SHA for 450c3e6 - Browse repository at this point
Copy the full SHA 450c3e6View commit details -
Configuration menu - View commit details
-
Copy full SHA for 8714c1e - Browse repository at this point
Copy the full SHA 8714c1eView commit details -
Configuration menu - View commit details
-
Copy full SHA for d34db90 - Browse repository at this point
Copy the full SHA d34db90View commit details -
Configuration menu - View commit details
-
Copy full SHA for 525f0a4 - Browse repository at this point
Copy the full SHA 525f0a4View commit details -
This class is used to implement Graphs that execute "natively" in the backend, using OpenCL command-buffers or LevelZero command-lists and only synchronizing with the host when required. Fallback to original Graph is provided in CHIPGraphExec::launch()
Configuration menu - View commit details
-
Copy full SHA for c4f654e - Browse repository at this point
Copy the full SHA c4f654eView commit details -
cleanup/refactor OpenCL backend code
* turn CHIPEvent.Refc from pointer to integer, move it into CHIPEventLevel0 * disable increaseRefCount/decreaseRefCount for OpenCL, the OpenCL runtime already does the refcounting * remove naked pointers to OpenCL objects, instead use the smart pointers from opencl.hpp header
Configuration menu - View commit details
-
Copy full SHA for c36b530 - Browse repository at this point
Copy the full SHA c36b530View commit details -
add more error-checking & bugfixes
"ctest --timeout 120 -R Unit_hipGraph" with POCL (with the new cl_pocl_command_buffer_* extensions) reports 59% tests passed, 62 tests failed out of 152 additionally, samples/graph + samples/graphMatrixMultiply work using the "native graphs" (cl_command_buffer), not the original chip-spv's graph execution.
Configuration menu - View commit details
-
Copy full SHA for 4aeedfa - Browse repository at this point
Copy the full SHA 4aeedfaView commit details -
OpenCL backend: add support for cl_intel_unified_shared_memory
Previously, the OpenCL device was checked for fine-grained SVM support, and if it was unavailable, the backend would assume only coarse-grained support and insert SVMMap & SVMUnmap for each buffer used by a kernel. Since Intel OpenCL and PoCL both support the USM extension now, but don't support fine-grained SVM, this commit helps to avoid the overhead of the unnecessary Map & Unmap commands.
Configuration menu - View commit details
-
Copy full SHA for 689cd1f - Browse repository at this point
Copy the full SHA 689cd1fView commit details -
Configuration menu - View commit details
-
Copy full SHA for 377f98d - Browse repository at this point
Copy the full SHA 377f98dView commit details -
Configuration menu - View commit details
-
Copy full SHA for 09df3c8 - Browse repository at this point
Copy the full SHA 09df3c8View commit details -
samples/hip-cuda: convert tests to report also Best iteration time
original code reports Average iteration time
Configuration menu - View commit details
-
Copy full SHA for b7c0212 - Browse repository at this point
Copy the full SHA b7c0212View commit details -
Configuration menu - View commit details
-
Copy full SHA for a6e2750 - Browse repository at this point
Copy the full SHA a6e2750View commit details
Commits on May 14, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 10b2fa1 - Browse repository at this point
Copy the full SHA 10b2fa1View commit details -
Configuration menu - View commit details
-
Copy full SHA for 781d7a9 - Browse repository at this point
Copy the full SHA 781d7a9View commit details -
disable USM memory allocation types other than Shared
enabling Device/Host USM types make some Texture tests fail
Configuration menu - View commit details
-
Copy full SHA for 7264667 - Browse repository at this point
Copy the full SHA 7264667View commit details -
Configuration menu - View commit details
-
Copy full SHA for f298848 - Browse repository at this point
Copy the full SHA f298848View commit details
Commits on May 17, 2023
-
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for a9f09e7 - Browse repository at this point
Copy the full SHA a9f09e7View commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 7ee740f - Browse repository at this point
Copy the full SHA 7ee740fView commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 79edc52 - Browse repository at this point
Copy the full SHA 79edc52View commit details -
Style fix src/backend/OpenCL/CHIPBackendOpenCL.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for b840157 - Browse repository at this point
Copy the full SHA b840157View commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for ba74330 - Browse repository at this point
Copy the full SHA ba74330View commit details -
Style fix src/backend/OpenCL/CHIPBackendOpenCL.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 0aec729 - Browse repository at this point
Copy the full SHA 0aec729View commit details -
Style fix src/backend/OpenCL/CHIPBackendOpenCL.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for e806f19 - Browse repository at this point
Copy the full SHA e806f19View commit details -
Style fix src/backend/OpenCL/SVMemoryRegion.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 2f9f86e - Browse repository at this point
Copy the full SHA 2f9f86eView commit details -
Avoid unnecessary copy in Kernel->getName()
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for e6aa207 - Browse repository at this point
Copy the full SHA e6aa207View commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for fb5d7e2 - Browse repository at this point
Copy the full SHA fb5d7e2View commit details -
Style fix src/backend/OpenCL/CHIPBackendOpenCL.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 29651da - Browse repository at this point
Copy the full SHA 29651daView commit details -
Style fix src/backend/OpenCL/CHIPBackendOpenCL.cc
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 282a90c - Browse repository at this point
Copy the full SHA 282a90cView commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 1e1fafc - Browse repository at this point
Copy the full SHA 1e1fafcView commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 84ab96d - Browse repository at this point
Copy the full SHA 84ab96dView commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for d922f75 - Browse repository at this point
Copy the full SHA d922f75View commit details -
Configuration menu - View commit details
-
Copy full SHA for 79a6e38 - Browse repository at this point
Copy the full SHA 79a6e38View commit details -
Configuration menu - View commit details
-
Copy full SHA for d7a0e71 - Browse repository at this point
Copy the full SHA d7a0e71View commit details -
Configuration menu - View commit details
-
Copy full SHA for 582d4ff - Browse repository at this point
Copy the full SHA 582d4ffView commit details -
Co-authored-by: Henry Linjamäki <linehill@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 0f4f3c8 - Browse repository at this point
Copy the full SHA 0f4f3c8View commit details