[llvm] [PPC] adjust cost for extract i64 from vector on P9 and above (PR #82963)

[Chen Zheng via llvm-commits]

https://github.com/chenzheng1030 created https://github.com/llvm/llvm-project/pull/82963

https://godbolt.org/z/Ma347Tx1W

>From 024a7781da6d24fc73dea7fb956528a08187938c Mon Sep 17 00:00:00 2001
From: Chen Zheng <czhengsz at cn.ibm.com>
Date: Mon, 26 Feb 2024 02:37:28 -0500
Subject: [PATCH] [PPC] adjust cost for extract i64 from vector on P9 and above

---
 llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp     | 10 ++++------
 .../CostModel/PowerPC/insert_extract-inseltpoison.ll   |  4 ++--
 llvm/test/Analysis/CostModel/PowerPC/insert_extract.ll |  4 ++--
 3 files changed, 8 insertions(+), 10 deletions(-)

diff --git a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
index 48274e8835381f..7adf1adcc64768 100644
--- a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
@@ -716,12 +716,10 @@ InstructionCost PPCTTIImpl::getVectorInstrCost(unsigned Opcode, Type *Val,
       } else if (ISD == ISD::EXTRACT_VECTOR_ELT) {
         // It's an extract.  Maybe we can do a cheap move-from VSR.
         unsigned EltSize = Val->getScalarSizeInBits();
-        if (EltSize == 64) {
-          // FIXME: no need to worry about endian, P9 has both mfvsrd/mfvsrld.
-          unsigned MfvsrdIndex = ST->isLittleEndian() ? 1 : 0;
-          if (Index == MfvsrdIndex)
-            return 1;
-        } else if (EltSize == 32) {
+        // P9 has both mfvsrd and mfvsrld for 64 bit integer.
+        if (EltSize == 64 && Index != -1U)
+          return 1;
+        else if (EltSize == 32) {
           unsigned MfvsrwzIndex = ST->isLittleEndian() ? 2 : 1;
           if (Index == MfvsrwzIndex)
             return 1;
diff --git a/llvm/test/Analysis/CostModel/PowerPC/insert_extract-inseltpoison.ll b/llvm/test/Analysis/CostModel/PowerPC/insert_extract-inseltpoison.ll
index e85e6f0ab08961..9894fb95f69230 100644
--- a/llvm/test/Analysis/CostModel/PowerPC/insert_extract-inseltpoison.ll
+++ b/llvm/test/Analysis/CostModel/PowerPC/insert_extract-inseltpoison.ll
@@ -147,11 +147,11 @@ define void @vexti64(<2 x i64> %p1) {
 ;
 ; CHECK-P9BE-LABEL: 'vexti64'
 ; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
-; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
+; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
 ; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; CHECK-P9LE-LABEL: 'vexti64'
-; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
+; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
 ; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
 ; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
diff --git a/llvm/test/Analysis/CostModel/PowerPC/insert_extract.ll b/llvm/test/Analysis/CostModel/PowerPC/insert_extract.ll
index fda6a4ae51fc2c..0c2ed5cb96b970 100644
--- a/llvm/test/Analysis/CostModel/PowerPC/insert_extract.ll
+++ b/llvm/test/Analysis/CostModel/PowerPC/insert_extract.ll
@@ -172,11 +172,11 @@ define void @vexti64(<2 x i64> %p1) {
 ;
 ; CHECK-P9BE-LABEL: 'vexti64'
 ; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
-; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
+; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
 ; CHECK-P9BE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; CHECK-P9LE-LABEL: 'vexti64'
-; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
+; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i1 = extractelement <2 x i64> %p1, i32 0
 ; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %i2 = extractelement <2 x i64> %p1, i32 1
 ; CHECK-P9LE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;