[llvm] [AMDGPU] Compiler should synthesize private buffer resource descriptor from flat_scratch_init (PR #79586)
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Thu Feb 8 03:27:18 PST 2024
================
@@ -5468,9 +5468,13 @@ If the *Target Properties* column of :ref:`amdgpu-processor-table` specifies
Instead the flat SCRATCH instructions are used.
Otherwise, Private Segment Buffer SGPR register is used to initialize 4 SGPRs
-that are used as a V# to access scratch. CP uses the value provided by the
-runtime. It is used, together with Scratch Wavefront Offset as an offset, to
-access the private memory space using a segment address. See
+that are used as a V# to access scratch.
+FIXME: The compiler synthetizes the initialization value for the Private Segment
----------------
arsenm wrote:
Not sure why this is "FIXME". Also typo 'synthetizes'
https://github.com/llvm/llvm-project/pull/79586
More information about the llvm-commits
mailing list