[LLVM][AVX512] PSRLWI Chnage imm8 to int

[delena]

LGTM

llvm/trunk/include/llvm/IR/IntrinsicsX86.td

let TargetPrefix = "x86" in {  // All intrinsics start with "llvm.x86.".
   def int_x86_avx512_mask_psrl_w_256 : GCCBuiltin<"__builtin_ia32_psrlw256_mask">,
               Intrinsic<[llvm_v16i16_ty], [llvm_v16i16_ty,
                          llvm_v8i16_ty, llvm_v16i16_ty,  llvm_i16_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psrl_w_512 : GCCBuiltin<"__builtin_ia32_psrlw512_mask">,
               Intrinsic<[llvm_v32i16_ty], [llvm_v32i16_ty,
                          llvm_v8i16_ty, llvm_v32i16_ty,  llvm_i32_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psrl_wi_128 : GCCBuiltin<"__builtin_ia32_psrlwi128_mask">,
               Intrinsic<[llvm_v8i16_ty], [llvm_v8i16_ty,
-                         llvm_i8_ty, llvm_v8i16_ty,  llvm_i8_ty], [IntrNoMem]>;
+                         llvm_i32_ty, llvm_v8i16_ty,  llvm_i8_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psrl_wi_256 : GCCBuiltin<"__builtin_ia32_psrlwi256_mask">,
               Intrinsic<[llvm_v16i16_ty], [llvm_v16i16_ty,
-                         llvm_i8_ty, llvm_v16i16_ty,  llvm_i16_ty], [IntrNoMem]>;
+                         llvm_i32_ty, llvm_v16i16_ty,  llvm_i16_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psrl_wi_512 : GCCBuiltin<"__builtin_ia32_psrlwi512_mask">,
               Intrinsic<[llvm_v32i16_ty], [llvm_v32i16_ty,
-                         llvm_i8_ty, llvm_v32i16_ty,  llvm_i32_ty], [IntrNoMem]>;
+                         llvm_i32_ty, llvm_v32i16_ty,  llvm_i32_ty], [IntrNoMem]>;
 
   def int_x86_avx512_mask_psra_w_128 : GCCBuiltin<"__builtin_ia32_psraw128_mask">,
               Intrinsic<[llvm_v8i16_ty], [llvm_v8i16_ty,
                          llvm_v8i16_ty, llvm_v8i16_ty, llvm_i8_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psra_w_256 : GCCBuiltin<"__builtin_ia32_psraw256_mask">,
               Intrinsic<[llvm_v16i16_ty], [llvm_v16i16_ty,
                          llvm_v8i16_ty, llvm_v16i16_ty, llvm_i16_ty], [IntrNoMem]>;
   def int_x86_avx512_mask_psra_w_512 : GCCBuiltin<"__builtin_ia32_psraw512_mask">,

llvm/trunk/lib/Target/X86/X86IntrinsicsInfo.h

   X86_INTRINSIC_DATA(avx512_mask_psrl_q_128,    INTR_TYPE_2OP_MASK, X86ISD::VSRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_q_256,    INTR_TYPE_2OP_MASK, X86ISD::VSRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_qi_128,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_qi_256,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_qi_512,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_w_128,    INTR_TYPE_2OP_MASK, X86ISD::VSRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_w_256,    INTR_TYPE_2OP_MASK, X86ISD::VSRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrl_w_512,    INTR_TYPE_2OP_MASK, X86ISD::VSRL, 0),
-  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_128,   INTR_TYPE_2OP_MASK, X86ISD::VSRLI, 0),
-  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_256,   INTR_TYPE_2OP_MASK, X86ISD::VSRLI, 0),
-  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_512,   INTR_TYPE_2OP_MASK, X86ISD::VSRLI, 0),
+  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_128,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
+  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_256,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
+  X86_INTRINSIC_DATA(avx512_mask_psrl_wi_512,   INTR_TYPE_2OP_IMM8_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrli_d,       VSHIFT_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrli_q,       VSHIFT_MASK, X86ISD::VSRLI, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv_d,       INTR_TYPE_2OP_MASK, ISD::SRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv_q,       INTR_TYPE_2OP_MASK, ISD::SRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv16_hi,    INTR_TYPE_2OP_MASK, ISD::SRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv2_di,     INTR_TYPE_2OP_MASK, ISD::SRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv32hi,     INTR_TYPE_2OP_MASK, ISD::SRL, 0),
   X86_INTRINSIC_DATA(avx512_mask_psrlv4_di,     INTR_TYPE_2OP_MASK, ISD::SRL, 0),

llvm/trunk/test/CodeGen/X86/avx512bw-intrinsics.ll

; AVX512F-32-NEXT:    retl
   %res = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3)
   %res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
   %res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
   %res3 = add <32 x i16> %res, %res1
   %res4 = add <32 x i16> %res3, %res2
   ret <32 x i16> %res4
 }
 
-declare <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16>, i8, <32 x i16>, i32)
+declare <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16>, i32, <32 x i16>, i32)
 
-define <32 x i16>@test_int_x86_avx512_mask_psrl_wi_512(<32 x i16> %x0, i8 %x1, <32 x i16> %x2, i32 %x3) {
+define <32 x i16>@test_int_x86_avx512_mask_psrl_wi_512(<32 x i16> %x0, i32 %x1, <32 x i16> %x2, i32 %x3) {
 ; AVX512BW-LABEL: test_int_x86_avx512_mask_psrl_wi_512:
 ; AVX512BW:       ## BB#0:
 ; AVX512BW-NEXT:    kmovd %esi, %k1
 ; AVX512BW-NEXT:    vpsrlw $3, %zmm0, %zmm1 {%k1}
 ; AVX512BW-NEXT:    vpsrlw $3, %zmm0, %zmm2 {%k1} {z}
 ; AVX512BW-NEXT:    vpsrlw $3, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpaddw %zmm0, %zmm1, %zmm0
 ; AVX512BW-NEXT:    vpaddw %zmm2, %zmm0, %zmm0
 ; AVX512BW-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: test_int_x86_avx512_mask_psrl_wi_512:
 ; AVX512F-32:       # BB#0:
 ; AVX512F-32-NEXT:    kmovd {{[0-9]+}}(%esp), %k1
 ; AVX512F-32-NEXT:    vpsrlw $3, %zmm0, %zmm1 {%k1}
 ; AVX512F-32-NEXT:    vpsrlw $3, %zmm0, %zmm2 {%k1} {z}
 ; AVX512F-32-NEXT:    vpsrlw $3, %zmm0, %zmm0
 ; AVX512F-32-NEXT:    vpaddw %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    vpaddw %zmm2, %zmm0, %zmm0
 ; AVX512F-32-NEXT:    retl
-  %res = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i8 3, <32 x i16> %x2, i32 %x3)
-  %res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i8 3, <32 x i16> %x2, i32 -1)
-  %res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i8 3, <32 x i16> zeroinitializer, i32 %x3)
+  %res = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 %x3)
+  %res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 -1)
+  %res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> zeroinitializer, i32 %x3)
   %res3 = add <32 x i16> %res, %res1
   %res4 = add <32 x i16> %res3, %res2
   ret <32 x i16> %res4
 }
 
 declare <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16>, <32 x i16>, <32 x i16>, i32)
 
 define <32 x i16>@test_int_x86_avx512_mask_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {

llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll

; CHECK-NEXT:    retq
   %res = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3)
   %res1 = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 -1)
   %res2 = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
   %res3 = add <16 x i16> %res, %res1
   %res4 = add <16 x i16> %res3, %res2
   ret <16 x i16> %res4
 }
 
-declare <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16>, i8, <8 x i16>, i8)
+declare <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16>, i32, <8 x i16>, i8)
 
-define <8 x i16>@test_int_x86_avx512_mask_psrl_wi_128(<8 x i16> %x0, i8 %x1, <8 x i16> %x2, i8 %x3) {
+define <8 x i16>@test_int_x86_avx512_mask_psrl_wi_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
 ; CHECK-LABEL: test_int_x86_avx512_mask_psrl_wi_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %esi, %k1
 ; CHECK-NEXT:    vpsrlw $3, %xmm0, %xmm1 {%k1}
 ; CHECK-NEXT:    vpsrlw $3, %xmm0, %xmm2 {%k1} {z}
 ; CHECK-NEXT:    vpsrlw $3, %xmm0, %xmm0
 ; CHECK-NEXT:    vpaddw %xmm0, %xmm1, %xmm0
 ; CHECK-NEXT:    vpaddw %xmm0, %xmm2, %xmm0
 ; CHECK-NEXT:    retq
-  %res = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i8 3, <8 x i16> %x2, i8 %x3)
-  %res1 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i8 3, <8 x i16> %x2, i8 -1)
-  %res2 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i8 3, <8 x i16> zeroinitializer, i8 %x3)
+  %res = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
+  %res1 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
+  %res2 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
   %res3 = add <8 x i16> %res, %res1
   %res4 = add <8 x i16> %res2, %res3
   ret <8 x i16> %res4
 }
 
-declare <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16>, i8, <16 x i16>, i16)
+declare <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16>, i32, <16 x i16>, i16)
 
-define <16 x i16>@test_int_x86_avx512_mask_psrl_wi_256(<16 x i16> %x0, i8 %x1, <16 x i16> %x2, i16 %x3) {
+define <16 x i16>@test_int_x86_avx512_mask_psrl_wi_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
 ; CHECK-LABEL: test_int_x86_avx512_mask_psrl_wi_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %esi, %k1
 ; CHECK-NEXT:    vpsrlw $3, %ymm0, %ymm1 {%k1}
 ; CHECK-NEXT:    vpsrlw $3, %ymm0, %ymm2 {%k1} {z}
 ; CHECK-NEXT:    vpsrlw $3, %ymm0, %ymm0
 ; CHECK-NEXT:    vpaddw %ymm0, %ymm1, %ymm0
 ; CHECK-NEXT:    vpaddw %ymm2, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
-  %res = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i8 3, <16 x i16> %x2, i16 %x3)
-  %res1 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i8 3, <16 x i16> %x2, i16 -1)
-  %res2 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i8 3, <16 x i16> zeroinitializer, i16 %x3)
+  %res = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
+  %res1 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
+  %res2 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
   %res3 = add <16 x i16> %res, %res1
   %res4 = add <16 x i16> %res3, %res2
   ret <16 x i16> %res4
 }
 
 declare <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16>, <16 x i16>, <16 x i16>, i16)
 
 define <16 x i16>@test_int_x86_avx512_mask_psrlv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {