llvm-exegesis Analysis Results

Triple: x86_64-unknown-linux-gnu

Cpu: bdver2

Sched Class CLFLUSH_CLFLUSHOPT contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
44
  • CLFLUSH (x6)
  • CLFLUSH (x6)
  • CLFLUSH (x6)
46.06
[45.51;46.99]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
1
  • 5
0.50
  • PdAGLU01: 1
  • PdLoad: 1
  • PdAGLU01: 1.00
  • PdLoad: 1.00

Sched Class LXADD8_LXADD16_LXADD32_LXADD64 contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
45
  • LXADD16 (x14)
  • LXADD16 (x14)
  • LXADD16 (x14)
45.09
[44.22;46.78]
51
  • LXADD32 (x14)
  • LXADD32 (x14)
  • LXADD32 (x14)
45.79
[43.96;46.72]
52
  • LXADD64 (x14)
  • LXADD64 (x14)
  • LXADD64 (x14)
45.83
[43.96;46.78]
46
  • LXADD8 (x18)
  • LXADD8 (x18)
  • LXADD8 (x18)
45.59
[44.68;47.18]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
4
  • 6
0.50
  • PdEX: 1
  • PdEX01: 1
  • PdAGLU01: 0.50
  • PdEX0: 0.75
  • PdEX1: 0.75

Sched Class WriteVecStoreNT contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
53
  • MMX_MOVNTQmr (x6)
  • MMX_MOVNTQmr (x6)
  • MMX_MOVNTQmr (x6)
91.49
[90.80;92.74]
54
  • VMOVNTDQmr (x6)
  • VMOVNTDQmr (x6)
  • VMOVNTDQmr (x6)
93.19
[90.80;97.63]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
1
  • 2
1.00
  • PdAGLU01: 1
  • PdFPSTO: 1
  • PdFPU: 1
  • PdFPU1: 1
  • PdStore: 1
  • PdAGLU01: 1.00
  • PdFPSTO: 1.00
  • PdFPU0: 0.25
  • PdFPU1: 1.25
  • PdFPU2: 0.25
  • PdFPU3: 0.25
  • PdStore: 1.00

Sched Class WriteFStoreNTX contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
56
  • VMOVNTPDmr (x6)
  • VMOVNTPDmr (x6)
92.62
[91.08;94.17]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
1
  • 3
1.00
  • PdAGLU01: 1
  • PdFPSTO: 1
  • PdFPU: 1
  • PdFPU1: 1
  • PdStore: 1
  • PdAGLU01: 1.00
  • PdFPSTO: 1.00
  • PdFPU0: 0.25
  • PdFPU1: 1.25
  • PdFPU2: 0.25
  • PdFPU3: 0.25
  • PdStore: 1.00

Sched Class WriteFStoreNT contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
47
  • MOVNTSD (x6)
  • MOVNTSD (x6)
  • MOVNTSD (x6)
91.68
[90.72;92.97]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
1
  • 3
1.00
  • PdAGLU01: 1
  • PdFPSTO: 1
  • PdFPU: 1
  • PdFPU1: 1
  • PdStore: 1
  • PdAGLU01: 1.00
  • PdFPSTO: 1.00
  • PdFPU0: 0.25
  • PdFPU1: 1.25
  • PdFPU2: 0.25
  • PdFPU3: 0.25
  • PdStore: 1.00

Sched Class WritePCmpIStrM contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
48
  • PCMPISTRMrr
  • PCMPISTRMrr
  • PCMPISTRMrr
4.71
[4.04;6.04]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
7
  • 6
1.00
  • PdEX: 1
  • PdEX0: 1
  • PdFPFMA: 2
  • PdFPU: 1
  • PdFPU1: 1
  • PdAGLU01: 0.50
  • PdEX0: 1.25
  • PdEX1: 0.25
  • PdFPFMA: 2.00
  • PdFPU0: 0.25
  • PdFPU1: 1.25
  • PdFPU2: 0.25
  • PdFPU3: 0.25

Sched Class WriteFStoreNTY contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
55
  • VMOVNTPDYmr (x6)
  • VMOVNTPDYmr (x6)
93.73
[93.26;94.21]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
4
  • 3
2.00
  • PdAGLU01: 2
  • PdFPSTO: 2
  • PdFPU: 2
  • PdFPU1: 2
  • PdStore: 2
  • PdAGLU01: 2.00
  • PdFPSTO: 2.00
  • PdFPU0: 0.50
  • PdFPU1: 2.50
  • PdFPU2: 0.50
  • PdFPU3: 0.50
  • PdStore: 2.00

Sched Class XADD16rm_XADD32rm_XADD64rm_XADD8rm_XCHG16rm_XCHG32rm_XCHG64rm_XCHG8rm contains instructions whose performance characteristics do not match that of LLVM:

ClusterIdOpcode/Configinverse_throughput
49
  • XCHG16rm (x14)
  • XCHG16rm (x14)
  • XCHG16rm (x14)
45.83
[44.28;46.66]
57
  • XCHG32rm (x14)
  • XCHG32rm (x14)
  • XCHG32rm (x14)
45.68
[44.21;46.43]
50
  • XCHG64rm (x14)
  • XCHG64rm (x14)
  • XCHG64rm (x14)
44.93
[44.10;46.55]

llvm SchedModel data:

ValidVariantNumMicroOpsLatencyRThroughputWriteProcResIdealized Resource Pressure
2
  • 5 (WriteResourceID 0)
  • 1 (WriteResourceID 0)
1.00
  • PdAGLU01: 2
  • PdEX01: 1
  • PdLoad: 1
  • PdStore: 1
  • PdAGLU01: 2.00
  • PdEX0: 0.50
  • PdEX1: 0.50
  • PdLoad: 1.00
  • PdStore: 1.00