Similar to #152524 - add handling for __builtin_ia32_pmulhrs to VectorExprEvaluator::VisitCallExpr and InterBuiltin.cpp ``` _mm_mulhrs_pi16 mm_mulhrs_epi16 mm256_mulhrs_epi16 mm512_mulhrs_epi16 mm_mask_mulhrs_epi16 mm256_mask_mulhrs_epi16 mm512_mask_mulhrs_epi16 mm_maskz_mulhrs_epi16 mm256_maskz_mulhrs_epi16 mm512_maskz_mulhrs_epi16 ``` Further info on how to constant fold PMULHRSW can be found in simplifyX86pmulh in X86InstCombineIntrinsic.cpp