Razvan Lupusoru 33c4e3e2bc

[flang][cuda] Avoid inserting GetDeviceAddress call in offload regions (#175225) ...

In regions destined for GPU offload, computing an address_of means
getting device address directly - no need (and actually incorrect) to
insert a runtime call to get the address. This was already working for
regions such as `gpu.launch` - but now it applies to acc regions as
well.

2026-01-09 12:20:18 -08:00

..

CUFAddConstructor.cpp

[flang][cuda][NFC] Move CUDA Fortran passes into CUDA directory (#174657)

2026-01-06 14:15:07 -08:00

CUFAllocationConversion.cpp

Reland "[flang][cuda] Add support for derived-type initialization on device #172568" (#174107)

2025-12-31 11:26:17 -08:00

CUFComputeSharedMemoryOffsetsAndSize.cpp

[flang][cuda][NFC] Move CUDA Fortran passes into CUDA directory (#174657)

2026-01-06 14:15:07 -08:00

CUFDeviceFuncTransform.cpp

[flang][cuda] Add CUFDeviceFuncTransform pass (#174487)

2026-01-05 15:12:19 -08:00

CUFDeviceGlobal.cpp

[flang][cuda][NFC] Move CUDA Fortran passes into CUDA directory (#174657)

2026-01-06 14:15:07 -08:00

CUFFunctionRewrite.cpp

[flang][cuda] Add CUFFunctionRewrite pass (#174650)

2026-01-06 13:30:03 -08:00

CUFGPUToLLVMConversion.cpp

[flang][cuda][NFC] Move CUDA Fortran passes into CUDA directory (#174657)

2026-01-06 14:15:07 -08:00

CUFLaunchAttachAttr.cpp

[flang][cuda] Add CUFLaunchAttachAttr pass (#174465)

2026-01-05 13:05:34 -08:00

CUFOpConversion.cpp

[flang][cuda] Avoid inserting GetDeviceAddress call in offload regions (#175225)

2026-01-09 12:20:18 -08:00

CUFPredefinedVarToGPU.cpp

[flang][cuda] Add pass to transform predefined variables (#174451)

2026-01-05 10:10:37 -08:00