arch-gcn3, gpu-compute: Implement out-of-range accesses
authorMichael LeBeane <Michael.Lebeane@amd.com>
Thu, 31 May 2018 17:03:38 +0000 (13:03 -0400)
committerAnthony Gutierrez <anthony.gutierrez@amd.com>
Mon, 13 Jul 2020 19:48:00 +0000 (19:48 +0000)
commited7daa10aae7639ff3435910721d0b0c292d5d53
tree90f077e5d77fac7ef5dc4830f9f4f84e88df75e7
parentf8e295922b9b59c7157dbd79cd0bba734497ff3a
arch-gcn3, gpu-compute: Implement out-of-range accesses

Certain buffer out-of-range memory accesses should be special
cased and not generate memory accesses. This patch implements
those special cases and supresses lanes from accessing memory
when the calculated address falls in an ISA-specified out-of-range
condition.

Change-Id: I8298f861c6b59587789853a01e503ba7d98cb13d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/29935
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Matt Sinclair <mattdsinclair@gmail.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
src/arch/gcn3/insts/instructions.cc
src/arch/gcn3/insts/op_encodings.hh
src/gpu-compute/global_memory_pipeline.cc