Fixed order of vector concat for NEON uminp.

UoB-HPC · Dec 17, 2024 · c0b2316 · c0b2316
1 parent eea6365
commit c0b2316
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/src/include/simeng/arch/aarch64/helpers/neon.hh b/src/include/simeng/arch/aarch64/helpers/neon.hh
@@ -592,8 +592,8 @@ RegisterValue vecUMinP(srcValContainer& sourceValues) {
 
   // Concatenate the vectors
   T temp[2 * I];
-  memcpy(temp, m, sizeof(T) * I);
-  memcpy(temp + (sizeof(T) * I), n, sizeof(T) * I);
+  memcpy(temp, n, sizeof(T) * I);
+  memcpy(temp + (sizeof(T) * I), m, sizeof(T) * I);
 
   T out[I];
   for (int i = 0; i < I; i++) {

diff --git a/test/regression/aarch64/instructions/neon.cc b/test/regression/aarch64/instructions/neon.cc
@@ -2684,8 +2684,8 @@ TEST_P(InstNeon, uminp) {
 
   )");
   CHECK_NEON(2, uint8_t,
-             {0x00, 0x11, 0x22, 0x44, 0xEE, 0xCC, 0xAA, 0x88, 0x00, 0xAA, 0xBB,
-              0xDD, 0x01, 0x03, 0x05, 0x07});
+             {0x00, 0xAA, 0xBB, 0xDD, 0x01, 0x03, 0x05, 0x07, 0x00, 0x11, 0x22,
+              0x44, 0xEE, 0xCC, 0xAA, 0x88});
 }
 TEST_P(InstNeon, umaxp) {
   // umaxp vd.16b vn.16b vm.16b