[clang] [compiler-rt] [llvm] [X86][AVX10.2] Support AVX10.2 option and VMPSADBW/VADDP[D,H,S] new instructions (PR #101452)

Phoebe Wang via llvm-commits llvm-commits at lists.llvm.org
Wed Jul 31 23:24:26 PDT 2024


================
@@ -926,7 +951,9 @@ void DisassemblerTables::emitContextTable(raw_ostream &o, unsigned &i) const {
       else
         o << "IC_VEX";
 
-      if ((index & ATTR_EVEX) && (index & ATTR_EVEXL2))
+      if ((index & ATTR_EVEXB) && (index & ATTR_EVEXU))
+        ; // Ignore ATTR_VEXL and ATTR_EVEXL2 under YMM rounding.
----------------
phoebewang wrote:

`L` and `L2` are repurposed as rouding bits, they have been test with {sae-*} tests.

https://github.com/llvm/llvm-project/pull/101452


More information about the llvm-commits mailing list