[llvm] [AMDGPU] Add VOP3 encoding for gfx13 (PR #196258)

[Matt Arsenault via llvm-commits]

================
@@ -9622,24 +9624,48 @@ void AMDGPUAsmParser::cvtVOP3P(MCInst &Inst, const OperandVector &Operands,
         Opc == AMDGPU::V_CVT_PK_FP8_F32_fake16_e64_dpp_gfx12 ||
         Opc == AMDGPU::V_CVT_PK_BF8_F32_fake16_e64_dpp8_gfx12 ||
         Opc == AMDGPU::V_CVT_PK_FP8_F32_fake16_e64_dpp8_gfx12 ||
+        Opc == AMDGPU::V_CVT_PK_BF8_F32_t16_e64_dpp_gfx13 ||
----------------
arsenm wrote:

We really need to do something about this function. This shouldn't require maintaining this use list of opcodes 

https://github.com/llvm/llvm-project/pull/196258