From 2c09e910545950f966d282a5d567ae85c8976252 Mon Sep 17 00:00:00 2001 From: Paul Walker Date: Sun, 21 Jun 2020 05:17:40 +0100 Subject: [PATCH] [MVT] Add missing floating point types for 1024/2048-bit vectors. Summary: This patch adds entries for: v64f16 v128f16 v64bf16 v128bf16 v32f64 Subscribers: dschuff, hiraditya, aheejin, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D82466 --- llvm/include/llvm/CodeGen/ValueTypes.td | 169 ++++++++++--------- llvm/include/llvm/Support/MachineValueType.h | 237 +++++++++++++++------------ llvm/lib/CodeGen/ValueTypes.cpp | 10 ++ llvm/utils/TableGen/CodeGenTarget.cpp | 5 + 4 files changed, 235 insertions(+), 186 deletions(-) diff --git a/llvm/include/llvm/CodeGen/ValueTypes.td b/llvm/include/llvm/CodeGen/ValueTypes.td index caa3d4daab3..dfa2e81608b 100644 --- a/llvm/include/llvm/CodeGen/ValueTypes.td +++ b/llvm/include/llvm/CodeGen/ValueTypes.td @@ -96,88 +96,93 @@ def v4f16 : ValueType<64 , 67>; // 4 x f16 vector value def v8f16 : ValueType<128, 68>; // 8 x f16 vector value def v16f16 : ValueType<256, 69>; // 16 x f16 vector value def v32f16 : ValueType<512, 70>; // 32 x f16 vector value -def v2bf16 : ValueType<32 , 71>; // 2 x bf16 vector value -def v3bf16 : ValueType<48 , 72>; // 3 x bf16 vector value -def v4bf16 : ValueType<64 , 73>; // 4 x bf16 vector value -def v8bf16 : ValueType<128, 74>; // 8 x bf16 vector value -def v16bf16 : ValueType<256, 75>; // 16 x bf16 vector value -def v32bf16 : ValueType<512, 76>; // 32 x bf16 vector value -def v1f32 : ValueType<32 , 77>; // 1 x f32 vector value -def v2f32 : ValueType<64 , 78>; // 2 x f32 vector value -def v3f32 : ValueType<96 , 79>; // 3 x f32 vector value -def v4f32 : ValueType<128, 80>; // 4 x f32 vector value -def v5f32 : ValueType<160, 81>; // 5 x f32 vector value -def v8f32 : ValueType<256, 82>; // 8 x f32 vector value -def v16f32 : ValueType<512, 83>; // 16 x f32 vector value -def v32f32 : ValueType<1024, 84>; // 32 x f32 vector value -def v64f32 : ValueType<2048, 85>; // 64 x f32 vector value -def v128f32 : ValueType<4096, 86>; // 128 x f32 vector value -def v256f32 : ValueType<8182, 87>; // 256 x f32 vector value -def v512f32 : ValueType<16384, 88>; // 512 x f32 vector value -def v1024f32 : ValueType<32768, 89>; // 1024 x f32 vector value -def v2048f32 : ValueType<65536, 90>; // 2048 x f32 vector value -def v1f64 : ValueType<64, 91>; // 1 x f64 vector value -def v2f64 : ValueType<128, 92>; // 2 x f64 vector value -def v4f64 : ValueType<256, 93>; // 4 x f64 vector value -def v8f64 : ValueType<512, 94>; // 8 x f64 vector value -def v16f64 : ValueType<1024, 95>; // 16 x f64 vector value - -def nxv1i1 : ValueType<1, 96>; // n x 1 x i1 vector value -def nxv2i1 : ValueType<2, 97>; // n x 2 x i1 vector value -def nxv4i1 : ValueType<4, 98>; // n x 4 x i1 vector value -def nxv8i1 : ValueType<8, 99>; // n x 8 x i1 vector value -def nxv16i1 : ValueType<16, 100>; // n x 16 x i1 vector value -def nxv32i1 : ValueType<32, 101>; // n x 32 x i1 vector value - -def nxv1i8 : ValueType<8, 102>; // n x 1 x i8 vector value -def nxv2i8 : ValueType<16, 103>; // n x 2 x i8 vector value -def nxv4i8 : ValueType<32, 104>; // n x 4 x i8 vector value -def nxv8i8 : ValueType<64, 105>; // n x 8 x i8 vector value -def nxv16i8 : ValueType<128, 106>; // n x 16 x i8 vector value -def nxv32i8 : ValueType<256, 107>; // n x 32 x i8 vector value - -def nxv1i16 : ValueType<16, 108>; // n x 1 x i16 vector value -def nxv2i16 : ValueType<32, 109>; // n x 2 x i16 vector value -def nxv4i16 : ValueType<64, 110>; // n x 4 x i16 vector value -def nxv8i16 : ValueType<128, 111>; // n x 8 x i16 vector value -def nxv16i16: ValueType<256, 112>; // n x 16 x i16 vector value -def nxv32i16: ValueType<512, 113>; // n x 32 x i16 vector value - -def nxv1i32 : ValueType<32, 114>; // n x 1 x i32 vector value -def nxv2i32 : ValueType<64, 115>; // n x 2 x i32 vector value -def nxv4i32 : ValueType<128, 116>; // n x 4 x i32 vector value -def nxv8i32 : ValueType<256, 117>; // n x 8 x i32 vector value -def nxv16i32: ValueType<512, 118>; // n x 16 x i32 vector value -def nxv32i32: ValueType<1024,119>; // n x 32 x i32 vector value - -def nxv1i64 : ValueType<64, 120>; // n x 1 x i64 vector value -def nxv2i64 : ValueType<128, 121>; // n x 2 x i64 vector value -def nxv4i64 : ValueType<256, 122>; // n x 4 x i64 vector value -def nxv8i64 : ValueType<512, 123>; // n x 8 x i64 vector value -def nxv16i64: ValueType<1024,124>; // n x 16 x i64 vector value -def nxv32i64: ValueType<2048,125>; // n x 32 x i64 vector value - -def nxv2f16 : ValueType<32 , 126>; // n x 2 x f16 vector value -def nxv4f16 : ValueType<64 , 127>; // n x 4 x f16 vector value -def nxv8f16 : ValueType<128, 128>; // n x 8 x f16 vector value -def nxv2bf16 : ValueType<32 , 129>; // n x 2 x bf16 vector value -def nxv4bf16 : ValueType<64 , 130>; // n x 4 x bf16 vector value -def nxv8bf16 : ValueType<128, 131>; // n x 8 x bf16 vector value -def nxv1f32 : ValueType<32 , 132>; // n x 1 x f32 vector value -def nxv2f32 : ValueType<64 , 133>; // n x 2 x f32 vector value -def nxv4f32 : ValueType<128, 134>; // n x 4 x f32 vector value -def nxv8f32 : ValueType<256, 135>; // n x 8 x f32 vector value -def nxv16f32 : ValueType<512, 136>; // n x 16 x f32 vector value -def nxv1f64 : ValueType<64, 137>; // n x 1 x f64 vector value -def nxv2f64 : ValueType<128, 138>; // n x 2 x f64 vector value -def nxv4f64 : ValueType<256, 139>; // n x 4 x f64 vector value -def nxv8f64 : ValueType<512, 140>; // n x 8 x f64 vector value - -def x86mmx : ValueType<64 , 141>; // X86 MMX value -def FlagVT : ValueType<0 , 142>; // Pre-RA sched glue -def isVoid : ValueType<0 , 143>; // Produces no value -def untyped: ValueType<8 , 144>; // Produces an untyped value -def exnref : ValueType<0 , 145>; // WebAssembly's exnref type +def v64f16 : ValueType<1024, 71>; // 64 x f16 vector value +def v128f16 : ValueType<2048, 72>; // 128 x f16 vector value +def v2bf16 : ValueType<32 , 73>; // 2 x bf16 vector value +def v3bf16 : ValueType<48 , 74>; // 3 x bf16 vector value +def v4bf16 : ValueType<64 , 75>; // 4 x bf16 vector value +def v8bf16 : ValueType<128, 76>; // 8 x bf16 vector value +def v16bf16 : ValueType<256, 77>; // 16 x bf16 vector value +def v32bf16 : ValueType<512, 78>; // 32 x bf16 vector value +def v64bf16 : ValueType<1024, 79>; // 64 x bf16 vector value +def v128bf16 : ValueType<2048, 80>; // 128 x bf16 vector value +def v1f32 : ValueType<32 , 81>; // 1 x f32 vector value +def v2f32 : ValueType<64 , 82>; // 2 x f32 vector value +def v3f32 : ValueType<96 , 83>; // 3 x f32 vector value +def v4f32 : ValueType<128, 84>; // 4 x f32 vector value +def v5f32 : ValueType<160, 85>; // 5 x f32 vector value +def v8f32 : ValueType<256, 86>; // 8 x f32 vector value +def v16f32 : ValueType<512, 87>; // 16 x f32 vector value +def v32f32 : ValueType<1024, 88>; // 32 x f32 vector value +def v64f32 : ValueType<2048, 89>; // 64 x f32 vector value +def v128f32 : ValueType<4096, 90>; // 128 x f32 vector value +def v256f32 : ValueType<8182, 91>; // 256 x f32 vector value +def v512f32 : ValueType<16384, 92>; // 512 x f32 vector value +def v1024f32 : ValueType<32768, 93>; // 1024 x f32 vector value +def v2048f32 : ValueType<65536, 94>; // 2048 x f32 vector value +def v1f64 : ValueType<64, 95>; // 1 x f64 vector value +def v2f64 : ValueType<128, 96>; // 2 x f64 vector value +def v4f64 : ValueType<256, 97>; // 4 x f64 vector value +def v8f64 : ValueType<512, 98>; // 8 x f64 vector value +def v16f64 : ValueType<1024, 99>; // 16 x f64 vector value +def v32f64 : ValueType<2048, 100>; // 32 x f64 vector value + +def nxv1i1 : ValueType<1, 101>; // n x 1 x i1 vector value +def nxv2i1 : ValueType<2, 102>; // n x 2 x i1 vector value +def nxv4i1 : ValueType<4, 103>; // n x 4 x i1 vector value +def nxv8i1 : ValueType<8, 104>; // n x 8 x i1 vector value +def nxv16i1 : ValueType<16, 105>; // n x 16 x i1 vector value +def nxv32i1 : ValueType<32, 106>; // n x 32 x i1 vector value + +def nxv1i8 : ValueType<8, 107>; // n x 1 x i8 vector value +def nxv2i8 : ValueType<16, 108>; // n x 2 x i8 vector value +def nxv4i8 : ValueType<32, 109>; // n x 4 x i8 vector value +def nxv8i8 : ValueType<64, 110>; // n x 8 x i8 vector value +def nxv16i8 : ValueType<128, 111>; // n x 16 x i8 vector value +def nxv32i8 : ValueType<256, 112>; // n x 32 x i8 vector value + +def nxv1i16 : ValueType<16, 113>; // n x 1 x i16 vector value +def nxv2i16 : ValueType<32, 114>; // n x 2 x i16 vector value +def nxv4i16 : ValueType<64, 115>; // n x 4 x i16 vector value +def nxv8i16 : ValueType<128, 116>; // n x 8 x i16 vector value +def nxv16i16: ValueType<256, 117>; // n x 16 x i16 vector value +def nxv32i16: ValueType<512, 118>; // n x 32 x i16 vector value + +def nxv1i32 : ValueType<32, 119>; // n x 1 x i32 vector value +def nxv2i32 : ValueType<64, 120>; // n x 2 x i32 vector value +def nxv4i32 : ValueType<128, 121>; // n x 4 x i32 vector value +def nxv8i32 : ValueType<256, 122>; // n x 8 x i32 vector value +def nxv16i32: ValueType<512, 123>; // n x 16 x i32 vector value +def nxv32i32: ValueType<1024,124>; // n x 32 x i32 vector value + +def nxv1i64 : ValueType<64, 125>; // n x 1 x i64 vector value +def nxv2i64 : ValueType<128, 126>; // n x 2 x i64 vector value +def nxv4i64 : ValueType<256, 127>; // n x 4 x i64 vector value +def nxv8i64 : ValueType<512, 128>; // n x 8 x i64 vector value +def nxv16i64: ValueType<1024,129>; // n x 16 x i64 vector value +def nxv32i64: ValueType<2048,130>; // n x 32 x i64 vector value + +def nxv2f16 : ValueType<32 , 131>; // n x 2 x f16 vector value +def nxv4f16 : ValueType<64 , 132>; // n x 4 x f16 vector value +def nxv8f16 : ValueType<128, 133>; // n x 8 x f16 vector value +def nxv2bf16 : ValueType<32 , 134>; // n x 2 x bf16 vector value +def nxv4bf16 : ValueType<64 , 135>; // n x 4 x bf16 vector value +def nxv8bf16 : ValueType<128, 136>; // n x 8 x bf16 vector value +def nxv1f32 : ValueType<32 , 137>; // n x 1 x f32 vector value +def nxv2f32 : ValueType<64 , 138>; // n x 2 x f32 vector value +def nxv4f32 : ValueType<128, 139>; // n x 4 x f32 vector value +def nxv8f32 : ValueType<256, 140>; // n x 8 x f32 vector value +def nxv16f32 : ValueType<512, 141>; // n x 16 x f32 vector value +def nxv1f64 : ValueType<64, 142>; // n x 1 x f64 vector value +def nxv2f64 : ValueType<128, 143>; // n x 2 x f64 vector value +def nxv4f64 : ValueType<256, 144>; // n x 4 x f64 vector value +def nxv8f64 : ValueType<512, 145>; // n x 8 x f64 vector value + +def x86mmx : ValueType<64 , 146>; // X86 MMX value +def FlagVT : ValueType<0 , 147>; // Pre-RA sched glue +def isVoid : ValueType<0 , 148>; // Produces no value +def untyped: ValueType<8 , 149>; // Produces an untyped value +def exnref : ValueType<0 , 150>; // WebAssembly's exnref type def token : ValueType<0 , 248>; // TokenTy def MetadataVT: ValueType<0, 249>; // Metadata diff --git a/llvm/include/llvm/Support/MachineValueType.h b/llvm/include/llvm/Support/MachineValueType.h index 9d2d0c85f58..e8789158687 100644 --- a/llvm/include/llvm/Support/MachineValueType.h +++ b/llvm/include/llvm/Support/MachineValueType.h @@ -123,91 +123,96 @@ namespace llvm { v8f16 = 68, // 8 x f16 v16f16 = 69, // 16 x f16 v32f16 = 70, // 32 x f16 - v2bf16 = 71, // 2 x bf16 - v3bf16 = 72, // 3 x bf16 - v4bf16 = 73, // 4 x bf16 - v8bf16 = 74, // 8 x bf16 - v16bf16 = 75, // 16 x bf16 - v32bf16 = 76, // 32 x bf16 - v1f32 = 77, // 1 x f32 - v2f32 = 78, // 2 x f32 - v3f32 = 79, // 3 x f32 - v4f32 = 80, // 4 x f32 - v5f32 = 81, // 5 x f32 - v8f32 = 82, // 8 x f32 - v16f32 = 83, // 16 x f32 - v32f32 = 84, // 32 x f32 - v64f32 = 85, // 64 x f32 - v128f32 = 86, // 128 x f32 - v256f32 = 87, // 256 x f32 - v512f32 = 88, // 512 x f32 - v1024f32 = 89, // 1024 x f32 - v2048f32 = 90, // 2048 x f32 - v1f64 = 91, // 1 x f64 - v2f64 = 92, // 2 x f64 - v4f64 = 93, // 4 x f64 - v8f64 = 94, // 8 x f64 - v16f64 = 95, // 16 x f64 + v64f16 = 71, // 64 x f16 + v128f16 = 72, // 128 x f16 + v2bf16 = 73, // 2 x bf16 + v3bf16 = 74, // 3 x bf16 + v4bf16 = 75, // 4 x bf16 + v8bf16 = 76, // 8 x bf16 + v16bf16 = 77, // 16 x bf16 + v32bf16 = 78, // 32 x bf16 + v64bf16 = 79, // 64 x bf16 + v128bf16 = 80, // 128 x bf16 + v1f32 = 81, // 1 x f32 + v2f32 = 82, // 2 x f32 + v3f32 = 83, // 3 x f32 + v4f32 = 84, // 4 x f32 + v5f32 = 85, // 5 x f32 + v8f32 = 86, // 8 x f32 + v16f32 = 87, // 16 x f32 + v32f32 = 88, // 32 x f32 + v64f32 = 89, // 64 x f32 + v128f32 = 90, // 128 x f32 + v256f32 = 91, // 256 x f32 + v512f32 = 92, // 512 x f32 + v1024f32 = 93, // 1024 x f32 + v2048f32 = 94, // 2048 x f32 + v1f64 = 95, // 1 x f64 + v2f64 = 96, // 2 x f64 + v4f64 = 97, // 4 x f64 + v8f64 = 98, // 8 x f64 + v16f64 = 99, // 16 x f64 + v32f64 = 100, // 32 x f64 FIRST_FP_FIXEDLEN_VECTOR_VALUETYPE = v2f16, - LAST_FP_FIXEDLEN_VECTOR_VALUETYPE = v16f64, + LAST_FP_FIXEDLEN_VECTOR_VALUETYPE = v32f64, FIRST_FIXEDLEN_VECTOR_VALUETYPE = v1i1, - LAST_FIXEDLEN_VECTOR_VALUETYPE = v16f64, - - nxv1i1 = 96, // n x 1 x i1 - nxv2i1 = 97, // n x 2 x i1 - nxv4i1 = 98, // n x 4 x i1 - nxv8i1 = 99, // n x 8 x i1 - nxv16i1 = 100, // n x 16 x i1 - nxv32i1 = 101, // n x 32 x i1 - - nxv1i8 = 102, // n x 1 x i8 - nxv2i8 = 103, // n x 2 x i8 - nxv4i8 = 104, // n x 4 x i8 - nxv8i8 = 105, // n x 8 x i8 - nxv16i8 = 106, // n x 16 x i8 - nxv32i8 = 107, // n x 32 x i8 - - nxv1i16 = 108, // n x 1 x i16 - nxv2i16 = 109, // n x 2 x i16 - nxv4i16 = 110, // n x 4 x i16 - nxv8i16 = 111, // n x 8 x i16 - nxv16i16 = 112, // n x 16 x i16 - nxv32i16 = 113, // n x 32 x i16 - - nxv1i32 = 114, // n x 1 x i32 - nxv2i32 = 115, // n x 2 x i32 - nxv4i32 = 116, // n x 4 x i32 - nxv8i32 = 117, // n x 8 x i32 - nxv16i32 = 118, // n x 16 x i32 - nxv32i32 = 119, // n x 32 x i32 - - nxv1i64 = 120, // n x 1 x i64 - nxv2i64 = 121, // n x 2 x i64 - nxv4i64 = 122, // n x 4 x i64 - nxv8i64 = 123, // n x 8 x i64 - nxv16i64 = 124, // n x 16 x i64 - nxv32i64 = 125, // n x 32 x i64 + LAST_FIXEDLEN_VECTOR_VALUETYPE = v32f64, + + nxv1i1 = 101, // n x 1 x i1 + nxv2i1 = 102, // n x 2 x i1 + nxv4i1 = 103, // n x 4 x i1 + nxv8i1 = 104, // n x 8 x i1 + nxv16i1 = 105, // n x 16 x i1 + nxv32i1 = 106, // n x 32 x i1 + + nxv1i8 = 107, // n x 1 x i8 + nxv2i8 = 108, // n x 2 x i8 + nxv4i8 = 109, // n x 4 x i8 + nxv8i8 = 110, // n x 8 x i8 + nxv16i8 = 111, // n x 16 x i8 + nxv32i8 = 112, // n x 32 x i8 + + nxv1i16 = 113, // n x 1 x i16 + nxv2i16 = 114, // n x 2 x i16 + nxv4i16 = 115, // n x 4 x i16 + nxv8i16 = 116, // n x 8 x i16 + nxv16i16 = 117, // n x 16 x i16 + nxv32i16 = 118, // n x 32 x i16 + + nxv1i32 = 119, // n x 1 x i32 + nxv2i32 = 120, // n x 2 x i32 + nxv4i32 = 121, // n x 4 x i32 + nxv8i32 = 122, // n x 8 x i32 + nxv16i32 = 123, // n x 16 x i32 + nxv32i32 = 124, // n x 32 x i32 + + nxv1i64 = 125, // n x 1 x i64 + nxv2i64 = 126, // n x 2 x i64 + nxv4i64 = 127, // n x 4 x i64 + nxv8i64 = 128, // n x 8 x i64 + nxv16i64 = 129, // n x 16 x i64 + nxv32i64 = 130, // n x 32 x i64 FIRST_INTEGER_SCALABLE_VECTOR_VALUETYPE = nxv1i1, LAST_INTEGER_SCALABLE_VECTOR_VALUETYPE = nxv32i64, - nxv2f16 = 126, // n x 2 x f16 - nxv4f16 = 127, // n x 4 x f16 - nxv8f16 = 128, // n x 8 x f16 - nxv2bf16 = 129, // n x 2 x bf16 - nxv4bf16 = 130, // n x 4 x bf16 - nxv8bf16 = 131, // n x 8 x bf16 - nxv1f32 = 132, // n x 1 x f32 - nxv2f32 = 133, // n x 2 x f32 - nxv4f32 = 134, // n x 4 x f32 - nxv8f32 = 135, // n x 8 x f32 - nxv16f32 = 136, // n x 16 x f32 - nxv1f64 = 137, // n x 1 x f64 - nxv2f64 = 138, // n x 2 x f64 - nxv4f64 = 139, // n x 4 x f64 - nxv8f64 = 140, // n x 8 x f64 + nxv2f16 = 131, // n x 2 x f16 + nxv4f16 = 132, // n x 4 x f16 + nxv8f16 = 133, // n x 8 x f16 + nxv2bf16 = 134, // n x 2 x bf16 + nxv4bf16 = 135, // n x 4 x bf16 + nxv8bf16 = 136, // n x 8 x bf16 + nxv1f32 = 137, // n x 1 x f32 + nxv2f32 = 138, // n x 2 x f32 + nxv4f32 = 139, // n x 4 x f32 + nxv8f32 = 140, // n x 8 x f32 + nxv16f32 = 141, // n x 16 x f32 + nxv1f64 = 142, // n x 1 x f64 + nxv2f64 = 143, // n x 2 x f64 + nxv4f64 = 144, // n x 4 x f64 + nxv8f64 = 145, // n x 8 x f64 FIRST_FP_SCALABLE_VECTOR_VALUETYPE = nxv2f16, LAST_FP_SCALABLE_VECTOR_VALUETYPE = nxv8f64, @@ -218,20 +223,20 @@ namespace llvm { FIRST_VECTOR_VALUETYPE = v1i1, LAST_VECTOR_VALUETYPE = nxv8f64, - x86mmx = 141, // This is an X86 MMX value + x86mmx = 146, // This is an X86 MMX value - Glue = 142, // This glues nodes together during pre-RA sched + Glue = 147, // This glues nodes together during pre-RA sched - isVoid = 143, // This has no value + isVoid = 148, // This has no value - Untyped = 144, // This value takes a register, but has - // unspecified type. The register class - // will be determined by the opcode. + Untyped = 149, // This value takes a register, but has + // unspecified type. The register class + // will be determined by the opcode. - exnref = 145, // WebAssembly's exnref type + exnref = 150, // WebAssembly's exnref type - FIRST_VALUETYPE = 1, // This is always the beginning of the list. - LAST_VALUETYPE = 146, // This always remains at the end of the list. + FIRST_VALUETYPE = 1, // This is always the beginning of the list. + LAST_VALUETYPE = 151, // This always remains at the end of the list. // This is the current maximum for LAST_VALUETYPE. // MVT::MAX_ALLOWED_VALUETYPE is used for asserts and to size bit vectors @@ -389,13 +394,17 @@ namespace llvm { bool is1024BitVector() const { return (SimpleTy == MVT::v1024i1 || SimpleTy == MVT::v128i8 || SimpleTy == MVT::v64i16 || SimpleTy == MVT::v32i32 || - SimpleTy == MVT::v16i64 || SimpleTy == MVT::v16f64); + SimpleTy == MVT::v16i64 || SimpleTy == MVT::v64f16 || + SimpleTy == MVT::v32f32 || SimpleTy == MVT::v16f64 || + SimpleTy == MVT::v64bf16); } /// Return true if this is a 2048-bit vector type. bool is2048BitVector() const { - return (SimpleTy == MVT::v256i8 || SimpleTy == MVT::v128i16 || - SimpleTy == MVT::v64i32 || SimpleTy == MVT::v32i64); + return (SimpleTy == MVT::v256i8 || SimpleTy == MVT::v128i16 || + SimpleTy == MVT::v64i32 || SimpleTy == MVT::v32i64 || + SimpleTy == MVT::v128f16 || SimpleTy == MVT::v64f32 || + SimpleTy == MVT::v32f64 || SimpleTy == MVT::v128bf16); } /// Return true if this is an overloaded type for TableGen. @@ -526,6 +535,8 @@ namespace llvm { case v8f16: case v16f16: case v32f16: + case v64f16: + case v128f16: case nxv2f16: case nxv4f16: case nxv8f16: return f16; @@ -535,6 +546,8 @@ namespace llvm { case v8bf16: case v16bf16: case v32bf16: + case v64bf16: + case v128bf16: case nxv2bf16: case nxv4bf16: case nxv8bf16: return bf16; @@ -562,6 +575,7 @@ namespace llvm { case v4f64: case v8f64: case v16f64: + case v32f64: case nxv1f64: case nxv2f64: case nxv4f64: @@ -589,11 +603,15 @@ namespace llvm { case v128i8: case v128i16: case v128i32: + case v128f16: + case v128bf16: case v128f32: return 128; case v64i1: case v64i8: case v64i16: case v64i32: + case v64f16: + case v64bf16: case v64f32: return 64; case v32i1: case v32i8: @@ -603,6 +621,7 @@ namespace llvm { case v32f16: case v32bf16: case v32f32: + case v32f64: case nxv32i1: case nxv32i8: case nxv32i16: @@ -855,15 +874,20 @@ namespace llvm { case v64i16: case v32i32: case v16i64: - case v16f64: - case v32f32: return TypeSize::Fixed(1024); + case v64f16: + case v64bf16: + case v32f32: + case v16f64: return TypeSize::Fixed(1024); case nxv32i32: case nxv16i64: return TypeSize::Scalable(1024); case v256i8: case v128i16: case v64i32: case v32i64: - case v64f32: return TypeSize::Fixed(2048); + case v128f16: + case v128bf16: + case v64f32: + case v32f64: return TypeSize::Fixed(2048); case nxv32i64: return TypeSize::Scalable(2048); case v128i32: case v128f32: return TypeSize::Fixed(4096); @@ -1033,20 +1057,24 @@ namespace llvm { if (NumElements == 1) return MVT::v1i128; break; case MVT::f16: - if (NumElements == 2) return MVT::v2f16; - if (NumElements == 3) return MVT::v3f16; - if (NumElements == 4) return MVT::v4f16; - if (NumElements == 8) return MVT::v8f16; - if (NumElements == 16) return MVT::v16f16; - if (NumElements == 32) return MVT::v32f16; + if (NumElements == 2) return MVT::v2f16; + if (NumElements == 3) return MVT::v3f16; + if (NumElements == 4) return MVT::v4f16; + if (NumElements == 8) return MVT::v8f16; + if (NumElements == 16) return MVT::v16f16; + if (NumElements == 32) return MVT::v32f16; + if (NumElements == 64) return MVT::v64f16; + if (NumElements == 128) return MVT::v128f16; break; case MVT::bf16: - if (NumElements == 2) return MVT::v2bf16; - if (NumElements == 3) return MVT::v3bf16; - if (NumElements == 4) return MVT::v4bf16; - if (NumElements == 8) return MVT::v8bf16; - if (NumElements == 16) return MVT::v16bf16; - if (NumElements == 32) return MVT::v32bf16; + if (NumElements == 2) return MVT::v2bf16; + if (NumElements == 3) return MVT::v3bf16; + if (NumElements == 4) return MVT::v4bf16; + if (NumElements == 8) return MVT::v8bf16; + if (NumElements == 16) return MVT::v16bf16; + if (NumElements == 32) return MVT::v32bf16; + if (NumElements == 64) return MVT::v64bf16; + if (NumElements == 128) return MVT::v128bf16; break; case MVT::f32: if (NumElements == 1) return MVT::v1f32; @@ -1070,6 +1098,7 @@ namespace llvm { if (NumElements == 4) return MVT::v4f64; if (NumElements == 8) return MVT::v8f64; if (NumElements == 16) return MVT::v16f64; + if (NumElements == 32) return MVT::v32f64; break; } return (MVT::SimpleValueType)(MVT::INVALID_SIMPLE_VALUE_TYPE); diff --git a/llvm/lib/CodeGen/ValueTypes.cpp b/llvm/lib/CodeGen/ValueTypes.cpp index 0fe2f73c9de..68e8acd218b 100644 --- a/llvm/lib/CodeGen/ValueTypes.cpp +++ b/llvm/lib/CodeGen/ValueTypes.cpp @@ -299,6 +299,10 @@ Type *EVT::getTypeForEVT(LLVMContext &Context) const { return FixedVectorType::get(Type::getHalfTy(Context), 16); case MVT::v32f16: return FixedVectorType::get(Type::getHalfTy(Context), 32); + case MVT::v64f16: + return FixedVectorType::get(Type::getBFloatTy(Context), 64); + case MVT::v128f16: + return FixedVectorType::get(Type::getBFloatTy(Context), 128); case MVT::v2bf16: return FixedVectorType::get(Type::getBFloatTy(Context), 2); case MVT::v3bf16: @@ -311,6 +315,10 @@ Type *EVT::getTypeForEVT(LLVMContext &Context) const { return FixedVectorType::get(Type::getBFloatTy(Context), 16); case MVT::v32bf16: return FixedVectorType::get(Type::getBFloatTy(Context), 32); + case MVT::v64bf16: + return FixedVectorType::get(Type::getBFloatTy(Context), 64); + case MVT::v128bf16: + return FixedVectorType::get(Type::getBFloatTy(Context), 128); case MVT::v1f32: return FixedVectorType::get(Type::getFloatTy(Context), 1); case MVT::v2f32: @@ -349,6 +357,8 @@ Type *EVT::getTypeForEVT(LLVMContext &Context) const { return FixedVectorType::get(Type::getDoubleTy(Context), 8); case MVT::v16f64: return FixedVectorType::get(Type::getDoubleTy(Context), 16); + case MVT::v32f64: + return FixedVectorType::get(Type::getDoubleTy(Context), 32); case MVT::nxv1i1: return ScalableVectorType::get(Type::getInt1Ty(Context), 1); case MVT::nxv2i1: diff --git a/llvm/utils/TableGen/CodeGenTarget.cpp b/llvm/utils/TableGen/CodeGenTarget.cpp index 78fb7328770..31f9f63d168 100644 --- a/llvm/utils/TableGen/CodeGenTarget.cpp +++ b/llvm/utils/TableGen/CodeGenTarget.cpp @@ -133,12 +133,16 @@ StringRef llvm::getEnumName(MVT::SimpleValueType T) { case MVT::v8f16: return "MVT::v8f16"; case MVT::v16f16: return "MVT::v16f16"; case MVT::v32f16: return "MVT::v32f16"; + case MVT::v64f16: return "MVT::v64f16"; + case MVT::v128f16: return "MVT::v128f16"; case MVT::v2bf16: return "MVT::v2bf16"; case MVT::v3bf16: return "MVT::v3bf16"; case MVT::v4bf16: return "MVT::v4bf16"; case MVT::v8bf16: return "MVT::v8bf16"; case MVT::v16bf16: return "MVT::v16bf16"; case MVT::v32bf16: return "MVT::v32bf16"; + case MVT::v64bf16: return "MVT::v64bf16"; + case MVT::v128bf16: return "MVT::v128bf16"; case MVT::v1f32: return "MVT::v1f32"; case MVT::v2f32: return "MVT::v2f32"; case MVT::v3f32: return "MVT::v3f32"; @@ -158,6 +162,7 @@ StringRef llvm::getEnumName(MVT::SimpleValueType T) { case MVT::v4f64: return "MVT::v4f64"; case MVT::v8f64: return "MVT::v8f64"; case MVT::v16f64: return "MVT::v16f64"; + case MVT::v32f64: return "MVT::v32f64"; case MVT::nxv1i1: return "MVT::nxv1i1"; case MVT::nxv2i1: return "MVT::nxv2i1"; case MVT::nxv4i1: return "MVT::nxv4i1"; -- 2.11.0