6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-12 13:30:51 +00:00
Code Issues Projects Releases Wiki Activity
91ccead42a
llvm-6502/lib/Target/NVPTX/NVPTXISelLowering.h

542 lines
14 KiB
C
Raw Normal View History

This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
//===-- NVPTXISelLowering.h - NVPTX DAG Lowering Interface ------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines the interfaces that NVPTX uses to lower LLVM code into a
// selection DAG.
//
//===----------------------------------------------------------------------===//
Canonicalize header guards into a common format. Add header guards to files that were missing guards. Remove #endif comments as they don't seem common in LLVM (we can easily add them back if we decide they're useful) Changes made by clang-tidy with minor tweaks. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@215558 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-13 16:26:38 +00:00
#ifndef LLVM_LIB_TARGET_NVPTX_NVPTXISELLOWERING_H
#define LLVM_LIB_TARGET_NVPTX_NVPTXISELLOWERING_H
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
#include "NVPTX.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/Target/TargetLowering.h"
namespace llvm {
namespace NVPTXISD {
enum NodeType {
// Start the numbering from where ISD NodeType finishes.
FIRST_NUMBER = ISD::BUILTIN_OP_END,
Wrapper,
CALL,
RET_FLAG,
LOAD_PARAM,
DeclareParam,
DeclareScalarParam,
DeclareRetParam,
DeclareRet,
DeclareScalarRet,
PrintCall,
PrintCallUni,
CallArgBegin,
CallArg,
LastCallArg,
CallArgEnd,
CallVoid,
CallVal,
CallSymbol,
Prototype,
MoveParam,
PseudoUseParam,
RETURN,
CallSeqBegin,
CallSeqEnd,
[NVPTX] Fix handling of indirect calls Using a special machine node is cleaner than an InlineAsm node, and fixes an assertion failure in InstrEmitter git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194810 91177308-0d34-0410-b5e6-96231b3b80d8
2013-11-15 12:30:04 +00:00
CallPrototype,
[NVPTX] Add support for [SHL,SRA,SRL]_PARTS git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211936 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:35:40 +00:00
FUN_SHFL_CLAMP,
FUN_SHFR_CLAMP,
[NVPTX] Implement fma and imad contraction as target DAGCombiner patterns This also introduces DAGCombiner patterns for mul.wide to multiply two smaller integers and produce a larger integer git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211935 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:35:37 +00:00
MUL_WIDE_SIGNED,
MUL_WIDE_UNSIGNED,
IMAD,
[NVPTX] Disable vector registers Vectors were being manually scalarized by the backend. Instead, let the target-independent code do all of the work. The manual scalarization was from a time before good target-independent support for scalarization in LLVM. However, this forces us to specially-handle vector loads and stores, which we can turn into PTX instructions that produce/consume multiple operands. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174968 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-12 14:18:49 +00:00
Dummy,
LoadV2 = ISD::FIRST_TARGET_MEMORY_OPCODE,
LoadV4,
LDGV2, // LDG.v2
LDGV4, // LDG.v4
LDUV2, // LDU.v2
LDUV4, // LDU.v4
StoreV2,
[NVPTX] Remove i8 register class. PTX support for i8 (.b8, .u8, .s8) is rather poor and we're better off just ignoring it and letting LLVM expand all i8 ops out to i16. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@185174 91177308-0d34-0410-b5e6-96231b3b80d8
2013-06-28 17:57:59 +00:00
StoreV4,
LoadParam,
LoadParamV2,
LoadParamV4,
StoreParam,
StoreParamV2,
StoreParamV4,
StoreParamS32, // to sext and store a <32bit value, not used currently
StoreParamU32, // to zext and store a <32bit value, not used currently
StoreRetval,
StoreRetvalV2,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
StoreRetvalV4,
// Texture intrinsics
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex1DFloatS32,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Tex1DFloatFloat,
Tex1DFloatFloatLevel,
Tex1DFloatFloatGrad,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex1DS32S32,
Tex1DS32Float,
Tex1DS32FloatLevel,
Tex1DS32FloatGrad,
Tex1DU32S32,
Tex1DU32Float,
Tex1DU32FloatLevel,
Tex1DU32FloatGrad,
Tex1DArrayFloatS32,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Tex1DArrayFloatFloat,
Tex1DArrayFloatFloatLevel,
Tex1DArrayFloatFloatGrad,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex1DArrayS32S32,
Tex1DArrayS32Float,
Tex1DArrayS32FloatLevel,
Tex1DArrayS32FloatGrad,
Tex1DArrayU32S32,
Tex1DArrayU32Float,
Tex1DArrayU32FloatLevel,
Tex1DArrayU32FloatGrad,
Tex2DFloatS32,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Tex2DFloatFloat,
Tex2DFloatFloatLevel,
Tex2DFloatFloatGrad,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex2DS32S32,
Tex2DS32Float,
Tex2DS32FloatLevel,
Tex2DS32FloatGrad,
Tex2DU32S32,
Tex2DU32Float,
Tex2DU32FloatLevel,
Tex2DU32FloatGrad,
Tex2DArrayFloatS32,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Tex2DArrayFloatFloat,
Tex2DArrayFloatFloatLevel,
Tex2DArrayFloatFloatGrad,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex2DArrayS32S32,
Tex2DArrayS32Float,
Tex2DArrayS32FloatLevel,
Tex2DArrayS32FloatGrad,
Tex2DArrayU32S32,
Tex2DArrayU32Float,
Tex2DArrayU32FloatLevel,
Tex2DArrayU32FloatGrad,
Tex3DFloatS32,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Tex3DFloatFloat,
Tex3DFloatFloatLevel,
Tex3DFloatFloatGrad,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Tex3DS32S32,
Tex3DS32Float,
Tex3DS32FloatLevel,
Tex3DS32FloatGrad,
Tex3DU32S32,
Tex3DU32Float,
Tex3DU32FloatLevel,
Tex3DU32FloatGrad,
TexCubeFloatFloat,
TexCubeFloatFloatLevel,
TexCubeS32Float,
TexCubeS32FloatLevel,
TexCubeU32Float,
TexCubeU32FloatLevel,
TexCubeArrayFloatFloat,
TexCubeArrayFloatFloatLevel,
TexCubeArrayS32Float,
TexCubeArrayS32FloatLevel,
TexCubeArrayU32Float,
TexCubeArrayU32FloatLevel,
Tld4R2DFloatFloat,
Tld4G2DFloatFloat,
Tld4B2DFloatFloat,
Tld4A2DFloatFloat,
Tld4R2DS64Float,
Tld4G2DS64Float,
Tld4B2DS64Float,
Tld4A2DS64Float,
Tld4R2DU64Float,
Tld4G2DU64Float,
Tld4B2DU64Float,
Tld4A2DU64Float,
TexUnified1DFloatS32,
TexUnified1DFloatFloat,
TexUnified1DFloatFloatLevel,
TexUnified1DFloatFloatGrad,
TexUnified1DS32S32,
TexUnified1DS32Float,
TexUnified1DS32FloatLevel,
TexUnified1DS32FloatGrad,
TexUnified1DU32S32,
TexUnified1DU32Float,
TexUnified1DU32FloatLevel,
TexUnified1DU32FloatGrad,
TexUnified1DArrayFloatS32,
TexUnified1DArrayFloatFloat,
TexUnified1DArrayFloatFloatLevel,
TexUnified1DArrayFloatFloatGrad,
TexUnified1DArrayS32S32,
TexUnified1DArrayS32Float,
TexUnified1DArrayS32FloatLevel,
TexUnified1DArrayS32FloatGrad,
TexUnified1DArrayU32S32,
TexUnified1DArrayU32Float,
TexUnified1DArrayU32FloatLevel,
TexUnified1DArrayU32FloatGrad,
TexUnified2DFloatS32,
TexUnified2DFloatFloat,
TexUnified2DFloatFloatLevel,
TexUnified2DFloatFloatGrad,
TexUnified2DS32S32,
TexUnified2DS32Float,
TexUnified2DS32FloatLevel,
TexUnified2DS32FloatGrad,
TexUnified2DU32S32,
TexUnified2DU32Float,
TexUnified2DU32FloatLevel,
TexUnified2DU32FloatGrad,
TexUnified2DArrayFloatS32,
TexUnified2DArrayFloatFloat,
TexUnified2DArrayFloatFloatLevel,
TexUnified2DArrayFloatFloatGrad,
TexUnified2DArrayS32S32,
TexUnified2DArrayS32Float,
TexUnified2DArrayS32FloatLevel,
TexUnified2DArrayS32FloatGrad,
TexUnified2DArrayU32S32,
TexUnified2DArrayU32Float,
TexUnified2DArrayU32FloatLevel,
TexUnified2DArrayU32FloatGrad,
TexUnified3DFloatS32,
TexUnified3DFloatFloat,
TexUnified3DFloatFloatLevel,
TexUnified3DFloatFloatGrad,
TexUnified3DS32S32,
TexUnified3DS32Float,
TexUnified3DS32FloatLevel,
TexUnified3DS32FloatGrad,
TexUnified3DU32S32,
TexUnified3DU32Float,
TexUnified3DU32FloatLevel,
TexUnified3DU32FloatGrad,
TexUnifiedCubeFloatFloat,
TexUnifiedCubeFloatFloatLevel,
TexUnifiedCubeS32Float,
TexUnifiedCubeS32FloatLevel,
TexUnifiedCubeU32Float,
TexUnifiedCubeU32FloatLevel,
TexUnifiedCubeArrayFloatFloat,
TexUnifiedCubeArrayFloatFloatLevel,
TexUnifiedCubeArrayS32Float,
TexUnifiedCubeArrayS32FloatLevel,
TexUnifiedCubeArrayU32Float,
TexUnifiedCubeArrayU32FloatLevel,
Tld4UnifiedR2DFloatFloat,
Tld4UnifiedG2DFloatFloat,
Tld4UnifiedB2DFloatFloat,
Tld4UnifiedA2DFloatFloat,
Tld4UnifiedR2DS64Float,
Tld4UnifiedG2DS64Float,
Tld4UnifiedB2DS64Float,
Tld4UnifiedA2DS64Float,
Tld4UnifiedR2DU64Float,
Tld4UnifiedG2DU64Float,
Tld4UnifiedB2DU64Float,
Tld4UnifiedA2DU64Float,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
// Surface intrinsics
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld1DI8Clamp,
Suld1DI16Clamp,
Suld1DI32Clamp,
Suld1DI64Clamp,
Suld1DV2I8Clamp,
Suld1DV2I16Clamp,
Suld1DV2I32Clamp,
Suld1DV2I64Clamp,
Suld1DV4I8Clamp,
Suld1DV4I16Clamp,
Suld1DV4I32Clamp,
Suld1DArrayI8Clamp,
Suld1DArrayI16Clamp,
Suld1DArrayI32Clamp,
Suld1DArrayI64Clamp,
Suld1DArrayV2I8Clamp,
Suld1DArrayV2I16Clamp,
Suld1DArrayV2I32Clamp,
Suld1DArrayV2I64Clamp,
Suld1DArrayV4I8Clamp,
Suld1DArrayV4I16Clamp,
Suld1DArrayV4I32Clamp,
Suld2DI8Clamp,
Suld2DI16Clamp,
Suld2DI32Clamp,
Suld2DI64Clamp,
Suld2DV2I8Clamp,
Suld2DV2I16Clamp,
Suld2DV2I32Clamp,
Suld2DV2I64Clamp,
Suld2DV4I8Clamp,
Suld2DV4I16Clamp,
Suld2DV4I32Clamp,
Suld2DArrayI8Clamp,
Suld2DArrayI16Clamp,
Suld2DArrayI32Clamp,
Suld2DArrayI64Clamp,
Suld2DArrayV2I8Clamp,
Suld2DArrayV2I16Clamp,
Suld2DArrayV2I32Clamp,
Suld2DArrayV2I64Clamp,
Suld2DArrayV4I8Clamp,
Suld2DArrayV4I16Clamp,
Suld2DArrayV4I32Clamp,
Suld3DI8Clamp,
Suld3DI16Clamp,
Suld3DI32Clamp,
Suld3DI64Clamp,
Suld3DV2I8Clamp,
Suld3DV2I16Clamp,
Suld3DV2I32Clamp,
Suld3DV2I64Clamp,
Suld3DV4I8Clamp,
Suld3DV4I16Clamp,
Suld3DV4I32Clamp,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld1DI8Trap,
Suld1DI16Trap,
Suld1DI32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld1DI64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld1DV2I8Trap,
Suld1DV2I16Trap,
Suld1DV2I32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld1DV2I64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld1DV4I8Trap,
Suld1DV4I16Trap,
Suld1DV4I32Trap,
Suld1DArrayI8Trap,
Suld1DArrayI16Trap,
Suld1DArrayI32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld1DArrayI64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld1DArrayV2I8Trap,
Suld1DArrayV2I16Trap,
Suld1DArrayV2I32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld1DArrayV2I64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld1DArrayV4I8Trap,
Suld1DArrayV4I16Trap,
Suld1DArrayV4I32Trap,
Suld2DI8Trap,
Suld2DI16Trap,
Suld2DI32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld2DI64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld2DV2I8Trap,
Suld2DV2I16Trap,
Suld2DV2I32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld2DV2I64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld2DV4I8Trap,
Suld2DV4I16Trap,
Suld2DV4I32Trap,
Suld2DArrayI8Trap,
Suld2DArrayI16Trap,
Suld2DArrayI32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld2DArrayI64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld2DArrayV2I8Trap,
Suld2DArrayV2I16Trap,
Suld2DArrayV2I32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld2DArrayV2I64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld2DArrayV4I8Trap,
Suld2DArrayV4I16Trap,
Suld2DArrayV4I32Trap,
Suld3DI8Trap,
Suld3DI16Trap,
Suld3DI32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld3DI64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld3DV2I8Trap,
Suld3DV2I16Trap,
Suld3DV2I32Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld3DV2I64Trap,
[NVPTX] Add preliminary intrinsics and codegen support for textures/surfaces This commit adds intrinsics and codegen support for the surface read/write and texture read instructions that take an explicit sampler parameter. Codegen operates on image handles at the PTX level, but falls back to direct replacement of handles with kernel arguments if image handles are not enabled. Note that image handles are explicitly disabled for all target architectures in this change (to be enabled later). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205907 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-09 15:39:15 +00:00
Suld3DV4I8Trap,
Suld3DV4I16Trap,
[NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch This also uses TSFlags to mark machine instructions that are surface/texture accesses, as well as the vector width for surface operations. This is used to simplify some of the switch statements that need to detect surface/texture instructions git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213256 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 11:59:04 +00:00
Suld3DV4I32Trap,
Suld1DI8Zero,
Suld1DI16Zero,
Suld1DI32Zero,
Suld1DI64Zero,
Suld1DV2I8Zero,
Suld1DV2I16Zero,
Suld1DV2I32Zero,
Suld1DV2I64Zero,
Suld1DV4I8Zero,
Suld1DV4I16Zero,
Suld1DV4I32Zero,
Suld1DArrayI8Zero,
Suld1DArrayI16Zero,
Suld1DArrayI32Zero,
Suld1DArrayI64Zero,
Suld1DArrayV2I8Zero,
Suld1DArrayV2I16Zero,
Suld1DArrayV2I32Zero,
Suld1DArrayV2I64Zero,
Suld1DArrayV4I8Zero,
Suld1DArrayV4I16Zero,
Suld1DArrayV4I32Zero,
Suld2DI8Zero,
Suld2DI16Zero,
Suld2DI32Zero,
Suld2DI64Zero,
Suld2DV2I8Zero,
Suld2DV2I16Zero,
Suld2DV2I32Zero,
Suld2DV2I64Zero,
Suld2DV4I8Zero,
Suld2DV4I16Zero,
Suld2DV4I32Zero,
Suld2DArrayI8Zero,
Suld2DArrayI16Zero,
Suld2DArrayI32Zero,
Suld2DArrayI64Zero,
Suld2DArrayV2I8Zero,
Suld2DArrayV2I16Zero,
Suld2DArrayV2I32Zero,
Suld2DArrayV2I64Zero,
Suld2DArrayV4I8Zero,
Suld2DArrayV4I16Zero,
Suld2DArrayV4I32Zero,
Suld3DI8Zero,
Suld3DI16Zero,
Suld3DI32Zero,
Suld3DI64Zero,
Suld3DV2I8Zero,
Suld3DV2I16Zero,
Suld3DV2I32Zero,
Suld3DV2I64Zero,
Suld3DV4I8Zero,
Suld3DV4I16Zero,
Suld3DV4I32Zero
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
};
}
Move NVPTX subtarget dependent variables from the target machine to the subtarget. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211860 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 04:33:14 +00:00
class NVPTXSubtarget;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
//===--------------------------------------------------------------------===//
// TargetLowering Implementation
//===--------------------------------------------------------------------===//
class NVPTXTargetLowering : public TargetLowering {
public:
Improve some const-correctness to remove a -Wcast-qual warning. No functional changes intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@214503 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-01 12:34:58 +00:00
explicit NVPTXTargetLowering(const NVPTXTargetMachine &TM);
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SDValue LowerOperation(SDValue Op, SelectionDAG &DAG) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
SDValue LowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerGlobalAddress(const GlobalValue *GV, int64_t Offset,
SelectionDAG &DAG) const;
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
const char *getTargetNodeName(unsigned Opcode) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
bool isTypeSupportedInIntrinsic(MVT VT) const;
[NVPTX] Run clang-format on all NVPTX sources. Hopefully this resolves any outstanding style issues and gives us an automated way of ensuring we conform to the style guidelines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178415 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-30 14:29:21 +00:00
bool getTgtMemIntrinsic(IntrinsicInfo &Info, const CallInst &I,
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
unsigned Intrinsic) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
/// isLegalAddressingMode - Return true if the addressing mode represented
/// by AM is legal for this target, for a load/store of the specified type
/// Used to guide target specific optimizations, like loop strength
/// reduction (LoopStrengthReduce.cpp) and memory optimization for
/// address mode (CodeGenPrepare.cpp)
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
bool isLegalAddressingMode(const AddrMode &AM, Type *Ty) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
/// getFunctionAlignment - Return the Log2 alignment of this function.
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
unsigned getFunctionAlignment(const Function *F) const;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[NVPTX] Handle all possible vector types in getSetCCResultType, not just the ones representable as MVTs git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211947 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:36:08 +00:00
EVT getSetCCResultType(LLVMContext &Ctx, EVT VT) const override {
Allow targets to prefer TypeSplitVector over TypePromoteInteger when computing the legalization method for vectors For some targets, it is desirable to prefer scalarizing <N x i1> instead of promoting to a larger legal type, such as <N x i32>. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168882 91177308-0d34-0410-b5e6-96231b3b80d8
2012-11-29 14:26:24 +00:00
if (VT.isVector())
[NVPTX] Handle all possible vector types in getSetCCResultType, not just the ones representable as MVTs git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211947 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:36:08 +00:00
return EVT::getVectorVT(Ctx, MVT::i1, VT.getVectorNumElements());
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
return MVT::i1;
}
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
ConstraintType
getConstraintType(const std::string &Constraint) const override;
[NVPTX] Run clang-format on all NVPTX sources. Hopefully this resolves any outstanding style issues and gives us an automated way of ensuring we conform to the style guidelines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178415 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-30 14:29:21 +00:00
std::pair<unsigned, const TargetRegisterClass *>
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
getRegForInlineAsmConstraint(const std::string &Constraint,
MVT VT) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SDValue LowerFormalArguments(
[NVPTX] Run clang-format on all NVPTX sources. Hopefully this resolves any outstanding style issues and gives us an automated way of ensuring we conform to the style guidelines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178415 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-30 14:29:21 +00:00
SDValue Chain, CallingConv::ID CallConv, bool isVarArg,
Track IR ordering of SelectionDAG nodes 2/4. Change SelectionDAG::getXXXNode() interfaces as well as call sites of these functions to pass in SDLoc instead of DebugLoc. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182703 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-25 02:42:55 +00:00
const SmallVectorImpl<ISD::InputArg> &Ins, SDLoc dl, SelectionDAG &DAG,
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SmallVectorImpl<SDValue> &InVals) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SDValue LowerCall(CallLoweringInfo &CLI,
SmallVectorImpl<SDValue> &InVals) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
std::string getPrototype(Type *, const ArgListTy &,
const SmallVectorImpl<ISD::OutputArg> &,
[NVPTX] Remove i8 register class. PTX support for i8 (.b8, .u8, .s8) is rather poor and we're better off just ignoring it and letting LLVM expand all i8 ops out to i16. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@185174 91177308-0d34-0410-b5e6-96231b3b80d8
2013-06-28 17:57:59 +00:00
unsigned retAlignment,
const ImmutableCallSite *CS) const;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SDValue
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
LowerReturn(SDValue Chain, CallingConv::ID CallConv, bool isVarArg,
const SmallVectorImpl<ISD::OutputArg> &Outs,
Track IR ordering of SelectionDAG nodes 2/4. Change SelectionDAG::getXXXNode() interfaces as well as call sites of these functions to pass in SDLoc instead of DebugLoc. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182703 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-25 02:42:55 +00:00
const SmallVectorImpl<SDValue> &OutVals, SDLoc dl,
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
SelectionDAG &DAG) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
void LowerAsmOperandForConstraint(SDValue Op, std::string &Constraint,
std::vector<SDValue> &Ops,
SelectionDAG &DAG) const override;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
Improve some const-correctness to remove a -Wcast-qual warning. No functional changes intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@214503 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-01 12:34:58 +00:00
const NVPTXTargetMachine *nvTM;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
// PTX always uses 32-bit shift amounts
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
MVT getScalarShiftAmountTy(EVT LHSTy) const override { return MVT::i32; }
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@212242 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-03 00:23:43 +00:00
TargetLoweringBase::LegalizeTypeAction
getPreferredVectorAction(EVT VT) const override;
Allow targets to prefer TypeSplitVector over TypePromoteInteger when computing the legalization method for vectors For some targets, it is desirable to prefer scalarizing <N x i1> instead of promoting to a larger legal type, such as <N x i32>. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168882 91177308-0d34-0410-b5e6-96231b3b80d8
2012-11-29 14:26:24 +00:00
[NVPTX] Improve handling of FP fusion We now consider the FPOpFusion flag when determining whether to fuse ops. We also explicitly emit add.rn when fusion is disabled to prevent ptxas from fusing the operations on its own. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213287 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 18:10:09 +00:00
bool allowFMA(MachineFunction &MF, CodeGenOpt::Level OptLevel) const;
Add override to overriden virtual methods, remove virtual keywords. No functionality change. Changes made by clang-tidy + some manual cleanup. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@217028 91177308-0d34-0410-b5e6-96231b3b80d8
2014-09-03 11:41:21 +00:00
bool isFMAFasterThanFMulAndFAdd(EVT) const override { return true; }
[NVPTX] Improve handling of FP fusion We now consider the FPOpFusion flag when determining whether to fuse ops. We also explicitly emit add.rn when fusion is disabled to prevent ptxas from fusing the operations on its own. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213287 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 18:10:09 +00:00
Override the TLI callback enableAggressiveFMAFusion and return true. Indeed, fmul, fmadd and fadd nodes cost the same number of cycles, so we can enable more combining heuristics to produce more fmadd nodes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225984 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-14 14:47:24 +00:00
bool enableAggressiveFMAFusion(EVT VT) const override { return true; }
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
private:
[NVPTX] Run clang-format on all NVPTX sources. Hopefully this resolves any outstanding style issues and gives us an automated way of ensuring we conform to the style guidelines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178415 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-30 14:29:21 +00:00
const NVPTXSubtarget &nvptxSubtarget; // cache the subtarget here
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
[NVPTX] Run clang-format on all NVPTX sources. Hopefully this resolves any outstanding style issues and gives us an automated way of ensuring we conform to the style guidelines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178415 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-30 14:29:21 +00:00
SDValue getExtSymb(SelectionDAG &DAG, const char *name, int idx,
EVT = MVT::i32) const;
[NVPTX] Default pointer type doesn't make sense for getParamSymbol() git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@184831 91177308-0d34-0410-b5e6-96231b3b80d8
2013-06-25 12:22:21 +00:00
SDValue getParamSymbol(SelectionDAG &DAG, int idx, EVT) const;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
SDValue getParamHelpSymbol(SelectionDAG &DAG, int idx);
SDValue LowerCONCAT_VECTORS(SDValue Op, SelectionDAG &DAG) const;
[NVPTX] Implement custom lowering of loads/stores for i1 Loads from i1 become loads from i8 followed by trunc Stores to i1 become zext to i8 followed by store to i8 Fixes PR13291 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167948 91177308-0d34-0410-b5e6-96231b3b80d8
2012-11-14 19:19:16 +00:00
SDValue LowerLOAD(SDValue Op, SelectionDAG &DAG) const;
[NVPTX] Disable vector registers Vectors were being manually scalarized by the backend. Instead, let the target-independent code do all of the work. The manual scalarization was from a time before good target-independent support for scalarization in LLVM. However, this forces us to specially-handle vector loads and stores, which we can turn into PTX instructions that produce/consume multiple operands. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174968 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-12 14:18:49 +00:00
SDValue LowerLOADi1(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerSTORE(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerSTOREi1(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerSTOREVector(SDValue Op, SelectionDAG &DAG) const;
[NVPTX] Add support for [SHL,SRA,SRL]_PARTS git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211936 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:35:40 +00:00
SDValue LowerShiftRightParts(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerShiftLeftParts(SDValue Op, SelectionDAG &DAG) const;
[C++11] Add 'override' keywords and remove 'virtual'. Additionally add 'final' and leave 'virtual' on some methods that are marked virtual without overriding anything and have no obvious overrides themselves. NVPTX edition git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207505 91177308-0d34-0410-b5e6-96231b3b80d8
2014-04-29 07:57:44 +00:00
void ReplaceNodeResults(SDNode *N, SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG) const override;
[NVPTX] Implement fma and imad contraction as target DAGCombiner patterns This also introduces DAGCombiner patterns for mul.wide to multiply two smaller integers and produce a larger integer git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211935 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-27 18:35:37 +00:00
SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const override;
[NVPTX] Remove i8 register class. PTX support for i8 (.b8, .u8, .s8) is rather poor and we're better off just ignoring it and letting LLVM expand all i8 ops out to i16. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@185174 91177308-0d34-0410-b5e6-96231b3b80d8
2013-06-28 17:57:59 +00:00
unsigned getArgumentAlignment(SDValue Callee, const ImmutableCallSite *CS,
Type *Ty, unsigned Idx) const;
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it. The new target machines are: nvptx (old ptx32) => 32-bit PTX nvptx64 (old ptx64) => 64-bit PTX The sources are based on the internal NVIDIA NVPTX back-end, and contain more functionality than the current PTX back-end currently provides. NV_CONTRIB git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00
};
} // namespace llvm
Canonicalize header guards into a common format. Add header guards to files that were missing guards. Remove #endif comments as they don't seem common in LLVM (we can easily add them back if we decide they're useful) Changes made by clang-tidy with minor tweaks. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@215558 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-13 16:26:38 +00:00
#endif
Reference in New Issue Copy Permalink