[llvm] [LLVM] Improve the DemandedBits Analysis (PR #148853)
Panagiotis K via llvm-commits
llvm-commits at lists.llvm.org
Wed Jul 16 00:40:57 PDT 2025
https://github.com/karouzakisp updated https://github.com/llvm/llvm-project/pull/148853
>From 78db1799107b2b3043b063150fd78626ba2fb2ea Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Tue, 15 Jul 2025 18:01:29 +0300
Subject: [PATCH 1/6] [LLVM] DemandedBits: Propagate demanded bits through
div/rem ops
---
llvm/lib/Analysis/DemandedBits.cpp | 54 +++++
llvm/test/Analysis/DemandedBits/div_rem.ll | 266 +++++++++++++++++++++
2 files changed, 320 insertions(+)
create mode 100644 llvm/test/Analysis/DemandedBits/div_rem.ll
diff --git a/llvm/lib/Analysis/DemandedBits.cpp b/llvm/lib/Analysis/DemandedBits.cpp
index 6694d5cc06c8c..7327bc4706ffc 100644
--- a/llvm/lib/Analysis/DemandedBits.cpp
+++ b/llvm/lib/Analysis/DemandedBits.cpp
@@ -36,6 +36,7 @@
#include "llvm/Support/Casting.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/KnownBits.h"
+#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include <algorithm>
#include <cstdint>
@@ -246,6 +247,59 @@ void DemandedBits::determineLiveOperandBits(
else
AB &= ~(Known.One & ~Known2.One);
break;
+ case Instruction::SRem:
+ case Instruction::URem:
+ case Instruction::UDiv:
+ case Instruction::SDiv: {
+ auto Opc = UserI->getOpcode();
+ auto IsDiv = Opc == Instruction::UDiv || Opc == Instruction::SDiv;
+ bool IsSigned = Opc == Instruction::SDiv || Opc == Instruction::SRem;
+ if (OperandNo == 0) {
+ const APInt *DivAmnt;
+ if (match(UserI->getOperand(1), m_APInt(DivAmnt))) {
+ uint64_t D = DivAmnt->getZExtValue();
+ if (isPowerOf2_64(D)) {
+ unsigned Sh = Log2_64(D);
+ if (IsDiv) {
+ AB = AOut.shl(Sh);
+ } else {
+ AB = AOut & APInt::getLowBitsSet(BitWidth, Sh);
+ }
+ } else if (IsDiv) { // Non power of 2 constant div
+ // x = q * C + r;
+ // q = x / C;
+ // We think of it like grade school division in base 2.
+ //
+ // x = [ unused | window m-bits | ... | needed bits ]
+ // ^ each step emits 1 quotient bit
+ // |
+ // |
+ // C fits in m = ⌈log₂ C⌉ bits
+ // Each new quotient bit consumes the window of m low bits and
+ // shifts one position left.
+
+ // To produce the first LowQ quotient/rem bits we slide the window
+ // LowQ times --> need at most LowQ + m low bits of the dividend.
+ // Need = LowQ + Ceil(log2(C)) (+1 sign bit for
+ // sdiv/srem). For example : Assume x = b7 b6 b5 b4 b3 b2 b1 b0.
+ // LowQ = 4, C = 5 and ceil(log_2(C)) = 3.
+ // step 0: b2 b1 b0, produces quotient q[0].
+ // step 1: b3 b2 b1, produces quotient q[1].
+ // step 2: b4 b3 b2, produces quotient q[2].
+ // step 3: b5 b4 b3, produces quotient q[3].
+ // k = LowQ - 1;
+ // TopIndex = k + m-1 = 3 + 2 = 5;
+ // The dividend bits b5...b0 are enough we don't care for b6 and b7.
+ unsigned LowQ = AOut.getActiveBits();
+ unsigned Need = LowQ + Log2_64_Ceil(D);
+ if (IsSigned)
+ Need++;
+ AB = APInt::getLowBitsSet(BitWidth, std::min(BitWidth, Need));
+ }
+ }
+ }
+ break;
+ }
case Instruction::Xor:
case Instruction::PHI:
AB = AOut;
diff --git a/llvm/test/Analysis/DemandedBits/div_rem.ll b/llvm/test/Analysis/DemandedBits/div_rem.ll
new file mode 100644
index 0000000000000..c27e88f321ac4
--- /dev/null
+++ b/llvm/test/Analysis/DemandedBits/div_rem.ll
@@ -0,0 +1,266 @@
+; RUN: opt -S -disable-output -passes="print<demanded-bits>" < %s 2>&1 | FileCheck %s
+
+define i8 @test_sdiv_const_amount_4(i32 %a) {
+; CHECK-LABEL: 'test_sdiv_const_amount_4'
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div = sdiv i32 %a, 4
+; CHECK-DAG: DemandedBits: 0x3fc for %a in %div = sdiv i32 %a, 4
+; CHECK-DAG: DemandedBits: 0xffffffff for 4 in %div = sdiv i32 %a, 4
+;
+ %div = sdiv i32 %a, 4
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_sdiv_const_amount_5(i32 %a) {
+; CHECK-LABEL: 'test_sdiv_const_amount_5'
+; CHECK-DAG: DemandedBits: 0xff for %div = sdiv i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xfff for %a in %div = sdiv i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for 5 in %div = sdiv i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = sdiv i32 %a, 5
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_sdiv_const_amount_8(i32 %a) {
+; CHECK-LABEL: 'test_sdiv_const_amount_8'
+; CHECK-DAG: DemandedBits: 0xff for %div = sdiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0x7f8 for %a in %div = sdiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %div = sdiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = sdiv i32 %a, 8
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_sdiv_const_amount_9(i32 %a) {
+; CHECK-LABEL: 'test_sdiv_const_amount_9'
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xfff for %a in %div = udiv i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for 9 in %div = udiv i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = udiv i32 %a, 9
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_sdiv(i32 %a, i32 %b) {
+; CHECK-LABEL: 'test_sdiv'
+; CHECK-DAG: DemandedBits: 0xff for %div = sdiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %div = sdiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %b in %div = sdiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = sdiv i32 %a, %b
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_udiv_const_amount_4(i32 %a) {
+; CHECK-LABEL: 'test_udiv_const_amount_4'
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 4
+; CHECK-DAG: DemandedBits: 0x3fc for %a in %div = udiv i32 %a, 4
+; CHECK-DAG: DemandedBits: 0xffffffff for 4 in %div = udiv i32 %a, 4
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = udiv i32 %a, 4
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_udiv_const_amount_5(i32 %a) {
+; CHECK-LABEL: 'test_udiv_const_amount_5'
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 5
+; CHECK-DAG: DemandedBits: 0x7ff for %a in %div = udiv i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for 5 in %div = udiv i32 %a, 5
+;
+ %div = udiv i32 %a, 5
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_udiv_const_amount_8(i32 %a) {
+; CHECK-LABEL: 'test_udiv_const_amount_8'
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0x7f8 for %a in %div = udiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %div = udiv i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = udiv i32 %a, 8
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_udiv_const_amount_9(i32 %a) {
+; CHECK-LABEL: 'test_udiv_const_amount_9'
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xfff for %a in %div = udiv i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for 9 in %div = udiv i32 %a, 9
+;
+ %div = udiv i32 %a, 9
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_udiv(i32 %a, i32 %b) {
+; CHECK-LABEL: 'test_udiv'
+; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %div = udiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %b in %div = udiv i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
+;
+ %div = udiv i32 %a, %b
+ %div.t = trunc i32 %div to i8
+ ret i8 %div.t
+}
+
+define i8 @test_srem_const_amount_4(i32 %a) {
+; CHECK-LABEL: 'test_srem_const_amount_4'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 4
+; CHECK-DAG: DemandedBits: 0x3 for %a in %rem = srem i32 %a, 4
+; CHECK-DAG: DemandedBits: 0xffffffff for 4 in %rem = srem i32 %a, 4
+;
+ %rem = srem i32 %a, 4
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_srem_const_amount_5(i32 %a) {
+; CHECK-LABEL: 'test_srem_const_amount_5'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = srem i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for 5 in %rem = srem i32 %a, 5
+;
+ %rem = srem i32 %a, 5
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_srem_const_amount_8(i32 %a) {
+; CHECK-LABEL: 'test_srem_const_amount_8'
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0x7 for %a in %rem = srem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %rem = srem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+;
+ %rem = srem i32 %a, 8
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_srem_const_amount_9(i32 %a) {
+; CHECK-LABEL: 'test_srem_const_amount_9'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = srem i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for 9 in %rem = srem i32 %a, 9
+;
+ %rem = srem i32 %a, 9
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_srem(i32 %a, i32 %b) {
+; CHECK-LABEL: 'test_srem'
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = srem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %b in %rem = srem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+;
+ %rem = srem i32 %a, %b
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_urem_const_amount_4(i32 %a) {
+; CHECK-LABEL: 'test_urem_const_amount_4'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = urem i32 %a, 4
+; CHECK-DAG: DemandedBits: 0x3 for %a in %rem = urem i32 %a, 4
+; CHECK-DAG: DemandedBits: 0xffffffff for 4 in %rem = urem i32 %a, 4
+;
+ %rem = urem i32 %a, 4
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+
+
+
+
+
+define i8 @test_urem_const_amount_5(i32 %a) {
+; CHECK-LABEL: 'test_urem_const_amount_5'
+; CHECK-DAG: DemandedBits: 0xff for %rem = urem i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = urem i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xffffffff for 5 in %rem = urem i32 %a, 5
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+;
+ %rem = urem i32 %a, 5
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_urem_const_amount_8(i32 %a) {
+; CHECK-LABEL: 'test_urem_const_amount_8'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = urem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0x7 for %a in %rem = urem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %rem = urem i32 %a, 8
+;
+ %rem = urem i32 %a, 8
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_urem_const_amount_9(i32 %a) {
+; CHECK-LABEL: 'test_urem_const_amount_9'
+; CHECK-DAG: DemandedBits: 0xff for %rem = urem i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = urem i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xffffffff for 9 in %rem = urem i32 %a, 9
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+;
+ %rem = urem i32 %a, 9
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
+define i8 @test_urem(i32 %a, i32 %b) {
+; CHECK-LABEL: 'test_urem'
+; CHECK-DAG: DemandedBits: 0xff for %rem = urem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = urem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xffffffff for %b in %rem = urem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+;
+ %rem = urem i32 %a, %b
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
>From 6d9b2710df83bf206c6019c6e30baff84391b4fd Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Wed, 16 Jul 2025 00:13:28 +0300
Subject: [PATCH 2/6] [LLVM] used APInt's API to be safe with from larger than
64 bits div/rems
---
llvm/lib/Analysis/DemandedBits.cpp | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/llvm/lib/Analysis/DemandedBits.cpp b/llvm/lib/Analysis/DemandedBits.cpp
index 7327bc4706ffc..260076b01470c 100644
--- a/llvm/lib/Analysis/DemandedBits.cpp
+++ b/llvm/lib/Analysis/DemandedBits.cpp
@@ -258,8 +258,8 @@ void DemandedBits::determineLiveOperandBits(
const APInt *DivAmnt;
if (match(UserI->getOperand(1), m_APInt(DivAmnt))) {
uint64_t D = DivAmnt->getZExtValue();
- if (isPowerOf2_64(D)) {
- unsigned Sh = Log2_64(D);
+ if (DivAmnt->isPowerOf2()) {
+ unsigned Sh = DivAmnt->countr_zero();
if (IsDiv) {
AB = AOut.shl(Sh);
} else {
@@ -274,7 +274,7 @@ void DemandedBits::determineLiveOperandBits(
// ^ each step emits 1 quotient bit
// |
// |
- // C fits in m = ⌈log₂ C⌉ bits
+ // C fits in m = ⌈log2 C⌉ bits
// Each new quotient bit consumes the window of m low bits and
// shifts one position left.
@@ -291,7 +291,7 @@ void DemandedBits::determineLiveOperandBits(
// TopIndex = k + m-1 = 3 + 2 = 5;
// The dividend bits b5...b0 are enough we don't care for b6 and b7.
unsigned LowQ = AOut.getActiveBits();
- unsigned Need = LowQ + Log2_64_Ceil(D);
+ unsigned Need = LowQ + DivAmnt->ceilLogBase2();
if (IsSigned)
Need++;
AB = APInt::getLowBitsSet(BitWidth, std::min(BitWidth, Need));
>From d2df5277b6b26820653438c3afe93ff3bfcb9326 Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Wed, 16 Jul 2025 06:39:27 +0300
Subject: [PATCH 3/6] minor updates and added new test that was supposed to be
failing
---
llvm/lib/Analysis/DemandedBits.cpp | 6 ++----
llvm/test/Analysis/DemandedBits/div_rem.ll | 16 ++++++++++++++++
2 files changed, 18 insertions(+), 4 deletions(-)
diff --git a/llvm/lib/Analysis/DemandedBits.cpp b/llvm/lib/Analysis/DemandedBits.cpp
index 260076b01470c..1672d5aedccc9 100644
--- a/llvm/lib/Analysis/DemandedBits.cpp
+++ b/llvm/lib/Analysis/DemandedBits.cpp
@@ -36,7 +36,6 @@
#include "llvm/Support/Casting.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/KnownBits.h"
-#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include <algorithm>
#include <cstdint>
@@ -257,7 +256,6 @@ void DemandedBits::determineLiveOperandBits(
if (OperandNo == 0) {
const APInt *DivAmnt;
if (match(UserI->getOperand(1), m_APInt(DivAmnt))) {
- uint64_t D = DivAmnt->getZExtValue();
if (DivAmnt->isPowerOf2()) {
unsigned Sh = DivAmnt->countr_zero();
if (IsDiv) {
@@ -278,10 +276,10 @@ void DemandedBits::determineLiveOperandBits(
// Each new quotient bit consumes the window of m low bits and
// shifts one position left.
- // To produce the first LowQ quotient/rem bits we slide the window
+ // To produce the first LowQ quotient bits we slide the window
// LowQ times --> need at most LowQ + m low bits of the dividend.
// Need = LowQ + Ceil(log2(C)) (+1 sign bit for
- // sdiv/srem). For example : Assume x = b7 b6 b5 b4 b3 b2 b1 b0.
+ // sdiv). For example : Assume x = b7 b6 b5 b4 b3 b2 b1 b0.
// LowQ = 4, C = 5 and ceil(log_2(C)) = 3.
// step 0: b2 b1 b0, produces quotient q[0].
// step 1: b3 b2 b1, produces quotient q[1].
diff --git a/llvm/test/Analysis/DemandedBits/div_rem.ll b/llvm/test/Analysis/DemandedBits/div_rem.ll
index c27e88f321ac4..3040c305a215e 100644
--- a/llvm/test/Analysis/DemandedBits/div_rem.ll
+++ b/llvm/test/Analysis/DemandedBits/div_rem.ll
@@ -130,6 +130,22 @@ define i8 @test_udiv(i32 %a, i32 %b) {
ret i8 %div.t
}
+define i8 @test_srem_zext_trunc_const_amount2(i8 %a) {
+; CHECK-LABEL: 'test_srem_const_amount_4'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %ext, 2
+; CHECK-DAG: DemandedBits: 0xffffffff for %ext in %rem = srem i32 %ext, 2
+; CHECK-DAG: DemandedBits: 0xffffffff for 2 in %rem = srem i32 %ext, 2
+; CHECK-DAG: DemandedBits: 0xffffffff for %ext = sext i8 %a to i32
+; CHECK-DAG: DemandedBits: 0xff for %a in %ext = sext i8 %a to i32
+;
+ %ext = sext i8 %a to i32
+ %rem = srem i32 %ext, 2
+ %rem.t = trunc i32 %rem to i8
+ ret i8 %rem.t
+}
+
define i8 @test_srem_const_amount_4(i32 %a) {
; CHECK-LABEL: 'test_srem_const_amount_4'
; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
>From e1c0bde2dcac121785ef63c57b923c9ffe3b67e7 Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Wed, 16 Jul 2025 06:43:34 +0300
Subject: [PATCH 4/6] fixed Check-Label
---
llvm/test/Analysis/DemandedBits/div_rem.ll | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/llvm/test/Analysis/DemandedBits/div_rem.ll b/llvm/test/Analysis/DemandedBits/div_rem.ll
index 3040c305a215e..e4d76b370b0dd 100644
--- a/llvm/test/Analysis/DemandedBits/div_rem.ll
+++ b/llvm/test/Analysis/DemandedBits/div_rem.ll
@@ -131,7 +131,7 @@ define i8 @test_udiv(i32 %a, i32 %b) {
}
define i8 @test_srem_zext_trunc_const_amount2(i8 %a) {
-; CHECK-LABEL: 'test_srem_const_amount_4'
+; CHECK-LABEL: 'test_srem_zext_trunc_const_amount2'
; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %ext, 2
>From 179eb2228e8cc356b1ae42027b71216467f3b5f0 Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Wed, 16 Jul 2025 07:51:30 +0300
Subject: [PATCH 5/6] fixed srem, we need to preserve the sign bit, otherwise
we might lose it
---
llvm/lib/Analysis/DemandedBits.cpp | 3 +++
llvm/test/Analysis/DemandedBits/div_rem.ll | 22 +++++++++++-----------
2 files changed, 14 insertions(+), 11 deletions(-)
diff --git a/llvm/lib/Analysis/DemandedBits.cpp b/llvm/lib/Analysis/DemandedBits.cpp
index 1672d5aedccc9..d9e02fe108a67 100644
--- a/llvm/lib/Analysis/DemandedBits.cpp
+++ b/llvm/lib/Analysis/DemandedBits.cpp
@@ -262,6 +262,9 @@ void DemandedBits::determineLiveOperandBits(
AB = AOut.shl(Sh);
} else {
AB = AOut & APInt::getLowBitsSet(BitWidth, Sh);
+ if (IsSigned) {
+ AB.setSignBit();
+ }
}
} else if (IsDiv) { // Non power of 2 constant div
// x = q * C + r;
diff --git a/llvm/test/Analysis/DemandedBits/div_rem.ll b/llvm/test/Analysis/DemandedBits/div_rem.ll
index e4d76b370b0dd..1f8c958f20380 100644
--- a/llvm/test/Analysis/DemandedBits/div_rem.ll
+++ b/llvm/test/Analysis/DemandedBits/div_rem.ll
@@ -93,11 +93,11 @@ define i8 @test_udiv_const_amount_5(i32 %a) {
define i8 @test_udiv_const_amount_8(i32 %a) {
; CHECK-LABEL: 'test_udiv_const_amount_8'
+; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
+; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
; CHECK-DAG: DemandedBits: 0xff for %div = udiv i32 %a, 8
; CHECK-DAG: DemandedBits: 0x7f8 for %a in %div = udiv i32 %a, 8
; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %div = udiv i32 %a, 8
-; CHECK-DAG: DemandedBits: 0xff for %div.t = trunc i32 %div to i8
-; CHECK-DAG: DemandedBits: 0xff for %div in %div.t = trunc i32 %div to i8
;
%div = udiv i32 %a, 8
%div.t = trunc i32 %div to i8
@@ -135,10 +135,10 @@ define i8 @test_srem_zext_trunc_const_amount2(i8 %a) {
; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %ext, 2
-; CHECK-DAG: DemandedBits: 0xffffffff for %ext in %rem = srem i32 %ext, 2
+; CHECK-DAG: DemandedBits: 0x80000001 for %ext in %rem = srem i32 %ext, 2
; CHECK-DAG: DemandedBits: 0xffffffff for 2 in %rem = srem i32 %ext, 2
-; CHECK-DAG: DemandedBits: 0xffffffff for %ext = sext i8 %a to i32
-; CHECK-DAG: DemandedBits: 0xff for %a in %ext = sext i8 %a to i32
+; CHECK-DAG: DemandedBits: 0x80000001 for %ext = sext i8 %a to i32
+; CHECK-DAG: DemandedBits: 0x81 for %a in %ext = sext i8 %a to i32
;
%ext = sext i8 %a to i32
%rem = srem i32 %ext, 2
@@ -151,7 +151,7 @@ define i8 @test_srem_const_amount_4(i32 %a) {
; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 4
-; CHECK-DAG: DemandedBits: 0x3 for %a in %rem = srem i32 %a, 4
+; CHECK-DAG: DemandedBits: 0x80000003 for %a in %rem = srem i32 %a, 4
; CHECK-DAG: DemandedBits: 0xffffffff for 4 in %rem = srem i32 %a, 4
;
%rem = srem i32 %a, 4
@@ -174,11 +174,11 @@ define i8 @test_srem_const_amount_5(i32 %a) {
define i8 @test_srem_const_amount_8(i32 %a) {
; CHECK-LABEL: 'test_srem_const_amount_8'
-; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 8
-; CHECK-DAG: DemandedBits: 0x7 for %a in %rem = srem i32 %a, 8
-; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %rem = srem i32 %a, 8
; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
+; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0x80000007 for %a in %rem = srem i32 %a, 8
+; CHECK-DAG: DemandedBits: 0xffffffff for 8 in %rem = srem i32 %a, 8
;
%rem = srem i32 %a, 8
%rem.t = trunc i32 %rem to i8
@@ -200,11 +200,11 @@ define i8 @test_srem_const_amount_9(i32 %a) {
define i8 @test_srem(i32 %a, i32 %b) {
; CHECK-LABEL: 'test_srem'
+; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xff for %rem = srem i32 %a, %b
+; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
; CHECK-DAG: DemandedBits: 0xffffffff for %a in %rem = srem i32 %a, %b
; CHECK-DAG: DemandedBits: 0xffffffff for %b in %rem = srem i32 %a, %b
-; CHECK-DAG: DemandedBits: 0xff for %rem.t = trunc i32 %rem to i8
-; CHECK-DAG: DemandedBits: 0xff for %rem in %rem.t = trunc i32 %rem to i8
;
%rem = srem i32 %a, %b
%rem.t = trunc i32 %rem to i8
>From 216f44bcb2003466384d62ef9eaa85a1e59fdb9c Mon Sep 17 00:00:00 2001
From: Panagiotis Karouzakis <karouzakispan at gmail.com>
Date: Wed, 16 Jul 2025 10:40:37 +0300
Subject: [PATCH 6/6] removed risky demanded bits propagation with division
---
llvm/lib/Analysis/DemandedBits.cpp | 30 ------------------------------
1 file changed, 30 deletions(-)
diff --git a/llvm/lib/Analysis/DemandedBits.cpp b/llvm/lib/Analysis/DemandedBits.cpp
index d9e02fe108a67..c8e9e8be9791d 100644
--- a/llvm/lib/Analysis/DemandedBits.cpp
+++ b/llvm/lib/Analysis/DemandedBits.cpp
@@ -266,36 +266,6 @@ void DemandedBits::determineLiveOperandBits(
AB.setSignBit();
}
}
- } else if (IsDiv) { // Non power of 2 constant div
- // x = q * C + r;
- // q = x / C;
- // We think of it like grade school division in base 2.
- //
- // x = [ unused | window m-bits | ... | needed bits ]
- // ^ each step emits 1 quotient bit
- // |
- // |
- // C fits in m = ⌈log2 C⌉ bits
- // Each new quotient bit consumes the window of m low bits and
- // shifts one position left.
-
- // To produce the first LowQ quotient bits we slide the window
- // LowQ times --> need at most LowQ + m low bits of the dividend.
- // Need = LowQ + Ceil(log2(C)) (+1 sign bit for
- // sdiv). For example : Assume x = b7 b6 b5 b4 b3 b2 b1 b0.
- // LowQ = 4, C = 5 and ceil(log_2(C)) = 3.
- // step 0: b2 b1 b0, produces quotient q[0].
- // step 1: b3 b2 b1, produces quotient q[1].
- // step 2: b4 b3 b2, produces quotient q[2].
- // step 3: b5 b4 b3, produces quotient q[3].
- // k = LowQ - 1;
- // TopIndex = k + m-1 = 3 + 2 = 5;
- // The dividend bits b5...b0 are enough we don't care for b6 and b7.
- unsigned LowQ = AOut.getActiveBits();
- unsigned Need = LowQ + DivAmnt->ceilLogBase2();
- if (IsSigned)
- Need++;
- AB = APInt::getLowBitsSet(BitWidth, std::min(BitWidth, Need));
}
}
}
More information about the llvm-commits
mailing list