[llvm] f568763 - [AArch64] Fold more load.x into load.i with large offset
via llvm-commits
llvm-commits at lists.llvm.org
Thu Dec 21 02:56:39 PST 2023
Author: zhongyunde 00443407
Date: 2023-12-21T18:54:15+08:00
New Revision: f5687636415969e6d945659a0b78734abdfb0f06
URL: https://github.com/llvm/llvm-project/commit/f5687636415969e6d945659a0b78734abdfb0f06
DIFF: https://github.com/llvm/llvm-project/commit/f5687636415969e6d945659a0b78734abdfb0f06.diff
LOG: [AArch64] Fold more load.x into load.i with large offset
The list of load.x is refer to canFoldIntoAddrMode on D152828.
Also support LDRSroX missed in canFoldIntoAddrMode
Added:
Modified:
llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
llvm/lib/Target/AArch64/AArch64LoadStoreOptimizer.cpp
llvm/test/CodeGen/AArch64/arm64-addrmode.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
index 855c7f17102561..175f6ef49c3ba5 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
@@ -4094,7 +4094,20 @@ AArch64InstrInfo::getLdStAmountOp(const MachineInstr &MI) {
switch (MI.getOpcode()) {
default:
llvm_unreachable("Unexpected opcode");
+ case AArch64::LDRBroX:
case AArch64::LDRBBroX:
+ case AArch64::LDRSBXroX:
+ case AArch64::LDRSBWroX:
+ case AArch64::LDRHroX:
+ case AArch64::LDRHHroX:
+ case AArch64::LDRSHXroX:
+ case AArch64::LDRSHWroX:
+ case AArch64::LDRWroX:
+ case AArch64::LDRSroX:
+ case AArch64::LDRSWroX:
+ case AArch64::LDRDroX:
+ case AArch64::LDRXroX:
+ case AArch64::LDRQroX:
return MI.getOperand(4);
}
}
diff --git a/llvm/lib/Target/AArch64/AArch64LoadStoreOptimizer.cpp b/llvm/lib/Target/AArch64/AArch64LoadStoreOptimizer.cpp
index b435b3ce03e7ee..aa7a4bc235361f 100644
--- a/llvm/lib/Target/AArch64/AArch64LoadStoreOptimizer.cpp
+++ b/llvm/lib/Target/AArch64/AArch64LoadStoreOptimizer.cpp
@@ -180,7 +180,7 @@ struct AArch64LoadStoreOpt : public MachineFunctionPass {
// Scan the instruction list to find a register assigned with a const
// value that can be combined with the current instruction (a load or store)
- // using base addressing with writeback. Scan forwards.
+ // using base addressing with writeback. Scan backwards.
MachineBasicBlock::iterator
findMatchingConstOffsetBackward(MachineBasicBlock::iterator I, unsigned Limit,
unsigned &Offset);
@@ -221,7 +221,7 @@ struct AArch64LoadStoreOpt : public MachineFunctionPass {
// Find and merge a base register updates before or after a ld/st instruction.
bool tryToMergeLdStUpdate(MachineBasicBlock::iterator &MBBI);
- // Find and merge a index ldr/st instructions into a base ld/st instruction.
+ // Find and merge a index ldr/st instruction into a base ld/st instruction.
bool tryToMergeIndexLdSt(MachineBasicBlock::iterator &MBBI, int Scale);
bool optimizeBlock(MachineBasicBlock &MBB, bool EnableNarrowZeroStOpt);
@@ -511,8 +511,34 @@ static unsigned getBaseAddressOpcode(unsigned Opc) {
switch (Opc) {
default:
llvm_unreachable("Opcode has no base address equivalent!");
+ case AArch64::LDRBroX:
+ return AArch64::LDRBui;
case AArch64::LDRBBroX:
return AArch64::LDRBBui;
+ case AArch64::LDRSBXroX:
+ return AArch64::LDRSBXui;
+ case AArch64::LDRSBWroX:
+ return AArch64::LDRSBWui;
+ case AArch64::LDRHroX:
+ return AArch64::LDRHui;
+ case AArch64::LDRHHroX:
+ return AArch64::LDRHHui;
+ case AArch64::LDRSHXroX:
+ return AArch64::LDRSHXui;
+ case AArch64::LDRSHWroX:
+ return AArch64::LDRSHWui;
+ case AArch64::LDRWroX:
+ return AArch64::LDRWui;
+ case AArch64::LDRSroX:
+ return AArch64::LDRSui;
+ case AArch64::LDRSWroX:
+ return AArch64::LDRSWui;
+ case AArch64::LDRDroX:
+ return AArch64::LDRDui;
+ case AArch64::LDRXroX:
+ return AArch64::LDRXui;
+ case AArch64::LDRQroX:
+ return AArch64::LDRQui;
}
}
@@ -764,10 +790,31 @@ static bool isMergeableIndexLdSt(MachineInstr &MI, int &Scale) {
default:
return false;
// Scaled instructions.
- // TODO: Add more index address loads/stores.
+ // TODO: Add more index address stores.
+ case AArch64::LDRBroX:
case AArch64::LDRBBroX:
+ case AArch64::LDRSBXroX:
+ case AArch64::LDRSBWroX:
Scale = 1;
return true;
+ case AArch64::LDRHroX:
+ case AArch64::LDRHHroX:
+ case AArch64::LDRSHXroX:
+ case AArch64::LDRSHWroX:
+ Scale = 2;
+ return true;
+ case AArch64::LDRWroX:
+ case AArch64::LDRSroX:
+ case AArch64::LDRSWroX:
+ Scale = 4;
+ return true;
+ case AArch64::LDRDroX:
+ case AArch64::LDRXroX:
+ Scale = 8;
+ return true;
+ case AArch64::LDRQroX:
+ Scale = 16;
+ return true;
}
}
diff --git a/llvm/test/CodeGen/AArch64/arm64-addrmode.ll b/llvm/test/CodeGen/AArch64/arm64-addrmode.ll
index 2181eaaee7db68..bfef61abd8c129 100644
--- a/llvm/test/CodeGen/AArch64/arm64-addrmode.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-addrmode.ll
@@ -239,9 +239,8 @@ define i32 @LdOffset_i8_zext32(ptr %a) {
define i32 @LdOffset_i8_sext32(ptr %a) {
; CHECK-LABEL: LdOffset_i8_sext32:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #56952 // =0xde78
-; CHECK-NEXT: movk w8, #15, lsl #16
-; CHECK-NEXT: ldrsb w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #253, lsl #12 // =1036288
+; CHECK-NEXT: ldrsb w0, [x8, #3704]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i8, ptr %a, i64 1039992
%val = load i8, ptr %arrayidx, align 1
@@ -266,9 +265,8 @@ define i64 @LdOffset_i8_zext64(ptr %a) {
define i64 @LdOffset_i8_sext64(ptr %a) {
; CHECK-LABEL: LdOffset_i8_sext64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #56952 // =0xde78
-; CHECK-NEXT: movk w8, #15, lsl #16
-; CHECK-NEXT: ldrsb x0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #253, lsl #12 // =1036288
+; CHECK-NEXT: ldrsb x0, [x8, #3704]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i8, ptr %a, i64 1039992
%val = load i8, ptr %arrayidx, align 1
@@ -280,9 +278,8 @@ define i64 @LdOffset_i8_sext64(ptr %a) {
define i16 @LdOffset_i16(ptr %a) {
; CHECK-LABEL: LdOffset_i16:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrh w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrh w0, [x8, #7408]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
%val = load i16, ptr %arrayidx, align 2
@@ -293,9 +290,8 @@ define i16 @LdOffset_i16(ptr %a) {
define i32 @LdOffset_i16_zext32(ptr %a) {
; CHECK-LABEL: LdOffset_i16_zext32:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrh w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrh w0, [x8, #7408]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
%val = load i16, ptr %arrayidx, align 2
@@ -307,9 +303,8 @@ define i32 @LdOffset_i16_zext32(ptr %a) {
define i32 @LdOffset_i16_sext32(ptr %a) {
; CHECK-LABEL: LdOffset_i16_sext32:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrsh w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrsh w0, [x8, #7408]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
%val = load i16, ptr %arrayidx, align 2
@@ -321,9 +316,8 @@ define i32 @LdOffset_i16_sext32(ptr %a) {
define i64 @LdOffset_i16_zext64(ptr %a) {
; CHECK-LABEL: LdOffset_i16_zext64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrh w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrh w0, [x8, #7408]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
%val = load i16, ptr %arrayidx, align 2
@@ -335,9 +329,8 @@ define i64 @LdOffset_i16_zext64(ptr %a) {
define i64 @LdOffset_i16_sext64(ptr %a) {
; CHECK-LABEL: LdOffset_i16_sext64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrsh x0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrsh x0, [x8, #7408]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
%val = load i16, ptr %arrayidx, align 2
@@ -349,9 +342,8 @@ define i64 @LdOffset_i16_sext64(ptr %a) {
define i32 @LdOffset_i32(ptr %a) {
; CHECK-LABEL: LdOffset_i32:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #31200 // =0x79e0
-; CHECK-NEXT: movk w8, #63, lsl #16
-; CHECK-NEXT: ldr w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #1012, lsl #12 // =4145152
+; CHECK-NEXT: ldr w0, [x8, #14816]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i32, ptr %a, i64 1039992
%val = load i32, ptr %arrayidx, align 4
@@ -362,9 +354,8 @@ define i32 @LdOffset_i32(ptr %a) {
define i64 @LdOffset_i32_zext64(ptr %a) {
; CHECK-LABEL: LdOffset_i32_zext64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #31200 // =0x79e0
-; CHECK-NEXT: movk w8, #63, lsl #16
-; CHECK-NEXT: ldr w0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #1012, lsl #12 // =4145152
+; CHECK-NEXT: ldr w0, [x8, #14816]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i32, ptr %a, i64 1039992
%val = load i32, ptr %arrayidx, align 2
@@ -376,9 +367,8 @@ define i64 @LdOffset_i32_zext64(ptr %a) {
define i64 @LdOffset_i32_sext64(ptr %a) {
; CHECK-LABEL: LdOffset_i32_sext64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #31200 // =0x79e0
-; CHECK-NEXT: movk w8, #63, lsl #16
-; CHECK-NEXT: ldrsw x0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #1012, lsl #12 // =4145152
+; CHECK-NEXT: ldrsw x0, [x8, #14816]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i32, ptr %a, i64 1039992
%val = load i32, ptr %arrayidx, align 2
@@ -390,9 +380,8 @@ define i64 @LdOffset_i32_sext64(ptr %a) {
define i64 @LdOffset_i64(ptr %a) {
; CHECK-LABEL: LdOffset_i64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #62400 // =0xf3c0
-; CHECK-NEXT: movk w8, #126, lsl #16
-; CHECK-NEXT: ldr x0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #2024, lsl #12 // =8290304
+; CHECK-NEXT: ldr x0, [x8, #29632]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i64, ptr %a, i64 1039992
%val = load i64, ptr %arrayidx, align 4
@@ -403,9 +392,8 @@ define i64 @LdOffset_i64(ptr %a) {
define <2 x i32> @LdOffset_v2i32(ptr %a) {
; CHECK-LABEL: LdOffset_v2i32:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #62400 // =0xf3c0
-; CHECK-NEXT: movk w8, #126, lsl #16
-; CHECK-NEXT: ldr d0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #2024, lsl #12 // =8290304
+; CHECK-NEXT: ldr d0, [x8, #29632]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds <2 x i32>, ptr %a, i64 1039992
%val = load <2 x i32>, ptr %arrayidx, align 4
@@ -416,9 +404,8 @@ define <2 x i32> @LdOffset_v2i32(ptr %a) {
define <2 x i64> @LdOffset_v2i64(ptr %a) {
; CHECK-LABEL: LdOffset_v2i64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #59264 // =0xe780
-; CHECK-NEXT: movk w8, #253, lsl #16
-; CHECK-NEXT: ldr q0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #4048, lsl #12 // =16580608
+; CHECK-NEXT: ldr q0, [x8, #59264]
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds <2 x i64>, ptr %a, i64 1039992
%val = load <2 x i64>, ptr %arrayidx, align 4
@@ -429,9 +416,8 @@ define <2 x i64> @LdOffset_v2i64(ptr %a) {
define double @LdOffset_i8_f64(ptr %a) {
; CHECK-LABEL: LdOffset_i8_f64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #56952 // =0xde78
-; CHECK-NEXT: movk w8, #15, lsl #16
-; CHECK-NEXT: ldrsb w8, [x0, x8]
+; CHECK-NEXT: add x8, x0, #253, lsl #12 // =1036288
+; CHECK-NEXT: ldrsb w8, [x8, #3704]
; CHECK-NEXT: scvtf d0, w8
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i8, ptr %a, i64 1039992
@@ -444,9 +430,8 @@ define double @LdOffset_i8_f64(ptr %a) {
define double @LdOffset_i16_f64(ptr %a) {
; CHECK-LABEL: LdOffset_i16_f64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #48368 // =0xbcf0
-; CHECK-NEXT: movk w8, #31, lsl #16
-; CHECK-NEXT: ldrsh w8, [x0, x8]
+; CHECK-NEXT: add x8, x0, #506, lsl #12 // =2072576
+; CHECK-NEXT: ldrsh w8, [x8, #7408]
; CHECK-NEXT: scvtf d0, w8
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i16, ptr %a, i64 1039992
@@ -459,9 +444,8 @@ define double @LdOffset_i16_f64(ptr %a) {
define double @LdOffset_i32_f64(ptr %a) {
; CHECK-LABEL: LdOffset_i32_f64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #31200 // =0x79e0
-; CHECK-NEXT: movk w8, #63, lsl #16
-; CHECK-NEXT: ldr s0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #1012, lsl #12 // =4145152
+; CHECK-NEXT: ldr s0, [x8, #14816]
; CHECK-NEXT: ucvtf d0, d0
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i32, ptr %a, i64 1039992
@@ -474,9 +458,8 @@ define double @LdOffset_i32_f64(ptr %a) {
define double @LdOffset_i64_f64(ptr %a) {
; CHECK-LABEL: LdOffset_i64_f64:
; CHECK: // %bb.0:
-; CHECK-NEXT: mov w8, #62400 // =0xf3c0
-; CHECK-NEXT: movk w8, #126, lsl #16
-; CHECK-NEXT: ldr d0, [x0, x8]
+; CHECK-NEXT: add x8, x0, #2024, lsl #12 // =8290304
+; CHECK-NEXT: ldr d0, [x8, #29632]
; CHECK-NEXT: scvtf d0, d0
; CHECK-NEXT: ret
%arrayidx = getelementptr inbounds i64, ptr %a, i64 1039992
More information about the llvm-commits
mailing list