[PATCH] D128189: [AMDGPU] Increase instruction cache line size to 128 bytes for GFX11
Jay Foad via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon Jun 20 05:45:29 PDT 2022
foad created this revision.
foad added reviewers: Joe_Nash, rampitec, piotr.
Herald added subscribers: kosarev, jsilvanus, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, jvesely, kzhuravl, arsenm.
Herald added a project: All.
foad requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D128189
Files:
llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp
llvm/test/CodeGen/AMDGPU/s_code_end.ll
Index: llvm/test/CodeGen/AMDGPU/s_code_end.ll
===================================================================
--- llvm/test/CodeGen/AMDGPU/s_code_end.ll
+++ llvm/test/CodeGen/AMDGPU/s_code_end.ll
@@ -41,10 +41,10 @@
; GCN-ASM-NEXT: [[END_LABEL3:\.Lfunc_end.*]]:
; GCN-ASM-NEXT: .size a_function, [[END_LABEL3]]-a_function
; GFX10END-ASM: .p2alignl 6, 3214868480
-; GFX11END-ASM: .p2alignl 6, 3214868480
+; GFX11END-ASM: .p2alignl 7, 3214868480
; GFX90AEND-ASM: .p2alignl 6, 3212836864
; GFX10END-ASM-NEXT: .fill 48, 4, 3214868480
-; GFX11END-ASM-NEXT: .fill 48, 4, 3214868480
+; GFX11END-ASM-NEXT: .fill 96, 4, 3214868480
; GFX90AEND-ASM-NEXT: .fill 256, 4, 3212836864
; GFX10NOEND-NOT: .fill
; GFX11NOEND-NOT: .fill
Index: llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp
===================================================================
--- llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp
+++ llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp
@@ -301,7 +301,7 @@
uint32_t Encoded_pad = Encoded_s_code_end;
// Instruction cache line size in bytes.
- const unsigned Log2CacheLineSize = 6;
+ const unsigned Log2CacheLineSize = AMDGPU::isGFX11Plus(STI) ? 7 : 6;
const unsigned CacheLineSize = 1u << Log2CacheLineSize;
// Extra padding amount in bytes to support prefetch mode 3.
@@ -824,7 +824,7 @@
uint32_t Encoded_pad = Encoded_s_code_end;
// Instruction cache line size in bytes.
- const unsigned Log2CacheLineSize = 6;
+ const unsigned Log2CacheLineSize = AMDGPU::isGFX11Plus(STI) ? 7 : 6;
const unsigned CacheLineSize = 1u << Log2CacheLineSize;
// Extra padding amount in bytes to support prefetch mode 3.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D128189.438356.patch
Type: text/x-patch
Size: 1783 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220620/7df1120e/attachment.bin>
More information about the llvm-commits
mailing list