[llvm] 9a31efa - [NFC][X86][MCA] AMD Zen 3: add tests for CMP dependency breaking
Roman Lebedev via llvm-commits
llvm-commits at lists.llvm.org
Sun May 9 14:04:12 PDT 2021
Author: Roman Lebedev
Date: 2021-05-10T00:03:20+03:00
New Revision: 9a31efa2f51b8150935a9b90eac6ab4eaa613841
URL: https://github.com/llvm/llvm-project/commit/9a31efa2f51b8150935a9b90eac6ab4eaa613841
DIFF: https://github.com/llvm/llvm-project/commit/9a31efa2f51b8150935a9b90eac6ab4eaa613841.diff
LOG: [NFC][X86][MCA] AMD Zen 3: add tests for CMP dependency breaking
Added:
Modified:
llvm/test/tools/llvm-mca/X86/Znver3/dependency-breaking-gpr.s
Removed:
################################################################################
diff --git a/llvm/test/tools/llvm-mca/X86/Znver3/dependency-breaking-gpr.s b/llvm/test/tools/llvm-mca/X86/Znver3/dependency-breaking-gpr.s
index d74a52d6fa88..667d169cc8df 100644
--- a/llvm/test/tools/llvm-mca/X86/Znver3/dependency-breaking-gpr.s
+++ b/llvm/test/tools/llvm-mca/X86/Znver3/dependency-breaking-gpr.s
@@ -11,6 +11,18 @@ sbbq %rax, %rax
mulxq %rax, %rax, %rax
# LLVM-MCA-END
+# LLVM-MCA-BEGIN
+mulxl %eax, %eax, %eax
+cmpl %eax, %eax
+cmovael %eax, %ecx
+# LLVM-MCA-END
+
+# LLVM-MCA-BEGIN
+mulxq %rax, %rax, %rax
+cmpq %rax, %rax
+cmovaeq %rax, %rcx
+# LLVM-MCA-END
+
# CHECK: [0] Code Region
# CHECK: Iterations: 1000
@@ -164,3 +176,169 @@ mulxq %rax, %rax, %rax
# CHECK-NEXT: 0. 2 1.5 0.5 1.5 sbbq %rax, %rax
# CHECK-NEXT: 1. 2 2.5 0.0 0.0 mulxq %rax, %rax, %rax
# CHECK-NEXT: 2 2.0 0.3 0.8 <total>
+
+# CHECK: [2] Code Region
+
+# CHECK: Iterations: 1000
+# CHECK-NEXT: Instructions: 3000
+# CHECK-NEXT: Total Cycles: 4005
+# CHECK-NEXT: Total uOps: 4000
+
+# CHECK: Dispatch Width: 6
+# CHECK-NEXT: uOps Per Cycle: 1.00
+# CHECK-NEXT: IPC: 0.75
+# CHECK-NEXT: Block RThroughput: 1.0
+
+# CHECK: Instruction Info:
+# CHECK-NEXT: [1]: #uOps
+# CHECK-NEXT: [2]: Latency
+# CHECK-NEXT: [3]: RThroughput
+# CHECK-NEXT: [4]: MayLoad
+# CHECK-NEXT: [5]: MayStore
+# CHECK-NEXT: [6]: HasSideEffects (U)
+
+# CHECK: [1] [2] [3] [4] [5] [6] Instructions:
+# CHECK-NEXT: 2 4 1.00 mulxl %eax, %eax, %eax
+# CHECK-NEXT: 1 1 0.25 cmpl %eax, %eax
+# CHECK-NEXT: 1 1 0.50 cmovael %eax, %ecx
+
+# CHECK: Resources:
+# CHECK-NEXT: [0] - Zn3AGU0
+# CHECK-NEXT: [1] - Zn3AGU1
+# CHECK-NEXT: [2] - Zn3AGU2
+# CHECK-NEXT: [3] - Zn3ALU0
+# CHECK-NEXT: [4] - Zn3ALU1
+# CHECK-NEXT: [5] - Zn3ALU2
+# CHECK-NEXT: [6] - Zn3ALU3
+# CHECK-NEXT: [7] - Zn3BRU1
+# CHECK-NEXT: [8] - Zn3FPP0
+# CHECK-NEXT: [9] - Zn3FPP1
+# CHECK-NEXT: [10] - Zn3FPP2
+# CHECK-NEXT: [11] - Zn3FPP3
+# CHECK-NEXT: [12.0] - Zn3FPP45
+# CHECK-NEXT: [12.1] - Zn3FPP45
+# CHECK-NEXT: [13] - Zn3FPSt
+# CHECK-NEXT: [14.0] - Zn3LSU
+# CHECK-NEXT: [14.1] - Zn3LSU
+# CHECK-NEXT: [14.2] - Zn3LSU
+# CHECK-NEXT: [15.0] - Zn3Load
+# CHECK-NEXT: [15.1] - Zn3Load
+# CHECK-NEXT: [15.2] - Zn3Load
+# CHECK-NEXT: [16.0] - Zn3Store
+# CHECK-NEXT: [16.1] - Zn3Store
+
+# CHECK: Resource pressure per iteration:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12.0] [12.1] [13] [14.0] [14.1] [14.2] [15.0] [15.1] [15.2] [16.0] [16.1]
+# CHECK-NEXT: - - - 0.67 1.00 0.67 0.67 - - - - - - - - - - - - - - - -
+
+# CHECK: Resource pressure by instruction:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12.0] [12.1] [13] [14.0] [14.1] [14.2] [15.0] [15.1] [15.2] [16.0] [16.1] Instructions:
+# CHECK-NEXT: - - - - 1.00 - - - - - - - - - - - - - - - - - - mulxl %eax, %eax, %eax
+# CHECK-NEXT: - - - - - 0.67 0.33 - - - - - - - - - - - - - - - - cmpl %eax, %eax
+# CHECK-NEXT: - - - 0.67 - - 0.33 - - - - - - - - - - - - - - - - cmovael %eax, %ecx
+
+# CHECK: Timeline view:
+# CHECK-NEXT: 012
+# CHECK-NEXT: Index 0123456789
+
+# CHECK: [0,0] DeeeeER . . mulxl %eax, %eax, %eax
+# CHECK-NEXT: [0,1] D====eER . . cmpl %eax, %eax
+# CHECK-NEXT: [0,2] D=====eER . . cmovael %eax, %ecx
+# CHECK-NEXT: [1,0] D====eeeeER . mulxl %eax, %eax, %eax
+# CHECK-NEXT: [1,1] .D=======eER. cmpl %eax, %eax
+# CHECK-NEXT: [1,2] .D========eER cmovael %eax, %ecx
+
+# CHECK: Average Wait times (based on the timeline view):
+# CHECK-NEXT: [0]: Executions
+# CHECK-NEXT: [1]: Average time spent waiting in a scheduler's queue
+# CHECK-NEXT: [2]: Average time spent waiting in a scheduler's queue while ready
+# CHECK-NEXT: [3]: Average time elapsed from WB until retire stage
+
+# CHECK: [0] [1] [2] [3]
+# CHECK-NEXT: 0. 2 3.0 0.5 0.0 mulxl %eax, %eax, %eax
+# CHECK-NEXT: 1. 2 6.5 0.0 0.0 cmpl %eax, %eax
+# CHECK-NEXT: 2. 2 7.5 0.0 0.0 cmovael %eax, %ecx
+# CHECK-NEXT: 2 5.7 0.2 0.0 <total>
+
+# CHECK: [3] Code Region
+
+# CHECK: Iterations: 1000
+# CHECK-NEXT: Instructions: 3000
+# CHECK-NEXT: Total Cycles: 4005
+# CHECK-NEXT: Total uOps: 4000
+
+# CHECK: Dispatch Width: 6
+# CHECK-NEXT: uOps Per Cycle: 1.00
+# CHECK-NEXT: IPC: 0.75
+# CHECK-NEXT: Block RThroughput: 1.0
+
+# CHECK: Instruction Info:
+# CHECK-NEXT: [1]: #uOps
+# CHECK-NEXT: [2]: Latency
+# CHECK-NEXT: [3]: RThroughput
+# CHECK-NEXT: [4]: MayLoad
+# CHECK-NEXT: [5]: MayStore
+# CHECK-NEXT: [6]: HasSideEffects (U)
+
+# CHECK: [1] [2] [3] [4] [5] [6] Instructions:
+# CHECK-NEXT: 2 4 1.00 mulxq %rax, %rax, %rax
+# CHECK-NEXT: 1 1 0.25 cmpq %rax, %rax
+# CHECK-NEXT: 1 1 0.50 cmovaeq %rax, %rcx
+
+# CHECK: Resources:
+# CHECK-NEXT: [0] - Zn3AGU0
+# CHECK-NEXT: [1] - Zn3AGU1
+# CHECK-NEXT: [2] - Zn3AGU2
+# CHECK-NEXT: [3] - Zn3ALU0
+# CHECK-NEXT: [4] - Zn3ALU1
+# CHECK-NEXT: [5] - Zn3ALU2
+# CHECK-NEXT: [6] - Zn3ALU3
+# CHECK-NEXT: [7] - Zn3BRU1
+# CHECK-NEXT: [8] - Zn3FPP0
+# CHECK-NEXT: [9] - Zn3FPP1
+# CHECK-NEXT: [10] - Zn3FPP2
+# CHECK-NEXT: [11] - Zn3FPP3
+# CHECK-NEXT: [12.0] - Zn3FPP45
+# CHECK-NEXT: [12.1] - Zn3FPP45
+# CHECK-NEXT: [13] - Zn3FPSt
+# CHECK-NEXT: [14.0] - Zn3LSU
+# CHECK-NEXT: [14.1] - Zn3LSU
+# CHECK-NEXT: [14.2] - Zn3LSU
+# CHECK-NEXT: [15.0] - Zn3Load
+# CHECK-NEXT: [15.1] - Zn3Load
+# CHECK-NEXT: [15.2] - Zn3Load
+# CHECK-NEXT: [16.0] - Zn3Store
+# CHECK-NEXT: [16.1] - Zn3Store
+
+# CHECK: Resource pressure per iteration:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12.0] [12.1] [13] [14.0] [14.1] [14.2] [15.0] [15.1] [15.2] [16.0] [16.1]
+# CHECK-NEXT: - - - 0.67 1.00 0.67 0.67 - - - - - - - - - - - - - - - -
+
+# CHECK: Resource pressure by instruction:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12.0] [12.1] [13] [14.0] [14.1] [14.2] [15.0] [15.1] [15.2] [16.0] [16.1] Instructions:
+# CHECK-NEXT: - - - - 1.00 - - - - - - - - - - - - - - - - - - mulxq %rax, %rax, %rax
+# CHECK-NEXT: - - - - - 0.67 0.33 - - - - - - - - - - - - - - - - cmpq %rax, %rax
+# CHECK-NEXT: - - - 0.67 - - 0.33 - - - - - - - - - - - - - - - - cmovaeq %rax, %rcx
+
+# CHECK: Timeline view:
+# CHECK-NEXT: 012
+# CHECK-NEXT: Index 0123456789
+
+# CHECK: [0,0] DeeeeER . . mulxq %rax, %rax, %rax
+# CHECK-NEXT: [0,1] D====eER . . cmpq %rax, %rax
+# CHECK-NEXT: [0,2] D=====eER . . cmovaeq %rax, %rcx
+# CHECK-NEXT: [1,0] D====eeeeER . mulxq %rax, %rax, %rax
+# CHECK-NEXT: [1,1] .D=======eER. cmpq %rax, %rax
+# CHECK-NEXT: [1,2] .D========eER cmovaeq %rax, %rcx
+
+# CHECK: Average Wait times (based on the timeline view):
+# CHECK-NEXT: [0]: Executions
+# CHECK-NEXT: [1]: Average time spent waiting in a scheduler's queue
+# CHECK-NEXT: [2]: Average time spent waiting in a scheduler's queue while ready
+# CHECK-NEXT: [3]: Average time elapsed from WB until retire stage
+
+# CHECK: [0] [1] [2] [3]
+# CHECK-NEXT: 0. 2 3.0 0.5 0.0 mulxq %rax, %rax, %rax
+# CHECK-NEXT: 1. 2 6.5 0.0 0.0 cmpq %rax, %rax
+# CHECK-NEXT: 2. 2 7.5 0.0 0.0 cmovaeq %rax, %rcx
+# CHECK-NEXT: 2 5.7 0.2 0.0 <total>
More information about the llvm-commits
mailing list