[clang] 0ad19a8 - [CUDA, NVPTX] Corrected fragment size for tf32 LD B matrix.

[Artem Belevich via cfe-commits]

Author: JackAKirk
Date: 2022-01-25T11:29:19-08:00
New Revision: 0ad19a833177861be55fefaff725ab89c8695d01

URL: https://github.com/llvm/llvm-project/commit/0ad19a833177861be55fefaff725ab89c8695d01
DIFF: https://github.com/llvm/llvm-project/commit/0ad19a833177861be55fefaff725ab89c8695d01.diff

LOG: [CUDA,NVPTX] Corrected fragment size for tf32 LD B matrix.

Signed-off-by: JackAKirk <jack.kirk at codeplay.com>

Reviewed By: tra

Differential Revision: https://reviews.llvm.org/D118023

Added: 
    

Modified: 
    clang/lib/CodeGen/CGBuiltin.cpp

Removed: 
    


################################################################################
diff  --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index cd35e7cbe76f7..a80a55e054a3b 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -17190,7 +17190,7 @@ static NVPTXMmaLdstInfo getNVPTXMmaLdstInfo(unsigned BuiltinID) {
   case NVPTX::BI__mma_tf32_m16n16k8_ld_a:
     return MMA_LDST(4, m16n16k8_load_a_tf32);
   case NVPTX::BI__mma_tf32_m16n16k8_ld_b:
-    return MMA_LDST(2, m16n16k8_load_b_tf32);
+    return MMA_LDST(4, m16n16k8_load_b_tf32);
   case NVPTX::BI__mma_tf32_m16n16k8_ld_c:
     return MMA_LDST(8, m16n16k8_load_c_f32);