| RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \ |
| RUN: %p/Inputs/jump_table_icp.s -o %t.o |
| RUN: link_fdata %p/Inputs/jump_table_icp.s %t.o %t.fdata --nmtool llvm-nm |
| RUN: llvm-strip --strip-unneeded %t.o |
| RUN: %clang %cflags -no-pie %t.o -o %t.exe -Wl,-q |
| |
| ## This test has commands that rely on shell capabilities that won't execute |
| ## correctly on Windows e.g. subshell execution |
| REQUIRES: shell |
| |
| RUN: (llvm-bolt %t.exe --data %t.fdata -o %t --relocs \ |
| RUN: --reorder-blocks=cache --split-functions --split-all-cold \ |
| RUN: --use-gnu-stack --dyno-stats --indirect-call-promotion=jump-tables \ |
| RUN: --print-icp -v=0 \ |
| RUN: --enable-bat --print-cache-metrics \ |
| RUN: --icp-jt-remaining-percent-threshold=10 \ |
| RUN: --icp-jt-total-percent-threshold=2 \ |
| RUN: --indirect-call-promotion-topn=1 \ |
| RUN: --icp-jump-tables-targets --align-functions-max-bytes=7 2>&1 && \ |
| RUN: llvm-objdump -d %t --print-imm-hex) | FileCheck %s |
| |
| BOLT-INFO: ICP total indirect callsites = 0 |
| BOLT-INFO: ICP total jump table callsites = 2 |
| BOLT-INFO: ICP total number of calls = 2137 |
| BOLT-INFO: ICP percentage of calls that are indirect = 0.0% |
| BOLT-INFO: ICP percentage of indirect calls that can be optimized = 0.0% |
| BOLT-INFO: ICP percentage of indirect calls that are optimized = 0.0% |
| BOLT-INFO: ICP percentage of jump table calls that can be optimized = 17.7% |
| BOLT-INFO: ICP percentage of jump table calls that are optimized = 100.0% |
| |
| CHECK: Binary Function "_Z3inci" after indirect-call-promotion |
| CHECK: .LBB{{.*}} (8 instructions, align : 1) |
| CHECK-NEXT: Entry Point |
| CHECK-NEXT: Exec Count : 1073 |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 189, count: 189), .LFT{{.*}} (mispreds: 0, count: 881) |
| |
| CHECK: .LFT{{.*}} (4 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 881 |
| CHECK: Predecessors: .LBB{{.*}} |
| CHECK: je {{.*}} # Offset: 28 |
| CHECK-NEXT: Successors: .Ltmp{{.*}} (mispreds: 138, count: 155), .Ltmp{{.*}} (mispreds: 0, count: 726) |
| |
| CHECK: .Ltmp{{.*}} (1 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 726 |
| CHECK: Predecessors: .LFT{{.*}} |
| CHECK: jmpq {{.*}} # Offset: 28 |
| CHECK-NEXT: Successors: .L{{.*}} (mispreds: 126, count: 157), .L{{.*}} (mispreds: 140, count: 156), .L{{.*}} (mispreds: 134, count: 152), .L{{.*}} (mispreds: 137, count: 150), .L{{.*}} (mispreds: 129, count: 148), .L{{.*}} (mispreds: 0, count: 0) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 167 |
| CHECK: Predecessors: .Ltmp{{.*}}, .LFT{{.*}} |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 156 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 157 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 148 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 150 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| |
| CHECK: Binary Function "_Z7inc_dupi" after indirect-call-promotion |
| CHECK: .LBB{{.*}} (8 instructions, align : 1) |
| CHECK-NEXT: Entry Point |
| CHECK-NEXT: Exec Count : 1064 |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 143, count: 144), .LFT{{.*}} (mispreds: 0, count: 777) |
| |
| CHECK: .LFT{{.*}} (4 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 777 |
| CHECK: Predecessors: .LBB{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 120, count: 138), .Ltmp{{.*}} (mispreds: 0, count: 639) |
| |
| CHECK: .Ltmp{{.*}} (1 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 639 |
| CHECK: Predecessors: .LFT{{.*}} |
| CHECK: Successors: .L{{.*}} (mispreds: 130, count: 137), .L{{.*}} (mispreds: 126, count: 136), .L{{.*}} (mispreds: 122, count: 130), .L{{.*}} (mispreds: 111, count: 130), .L{{.*}} (mispreds: 104, count: 114), .L{{.*}} (mispreds: 0, count: 0) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 137 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 0, count: 106) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 136 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 0, count: 113) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 130 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 0, count: 97) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 130 |
| CHECK: Predecessors: .Ltmp{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 0, count: 105) |
| |
| CHECK: .Ltmp{{.*}} (5 instructions, align : 1) |
| CHECK-NEXT: Exec Count : 140 |
| CHECK: Predecessors: .Ltmp{{.*}}, .LFT{{.*}} |
| CHECK: Successors: .Ltmp{{.*}} (mispreds: 0, count: 98) |
| |
| CHECK: <_Z3inci>: |
| CHECK: movq 0x{{.*}}(,%rax,8), %rax |
| CHECK-NEXT: cmpq $0x{{.*}}, %rax |
| CHECK-NEXT: je {{.*}} <_Z3inci+0x{{.*}}> |
| CHECK-NEXT: jmpq *%rax |
| |
| CHECK: <_Z7inc_dupi>: |
| CHECK: movq 0x{{.*}}(,%rax,8), %rax |
| CHECK-NEXT: cmpq $0x{{.*}}, %rax |
| CHECK-NEXT: je {{.*}} <_Z7inc_dupi+0x{{.*}}> |
| CHECK-NEXT: jmpq *%rax |
