OV - exec - Loop 4

	Loop Id: 4	Module: exec	Source: main.c:140-147	Coverage: 0.16%

0x401a20 VMOVSS	0x40a1b0(%R12,%RBX,4),%XMM0    [2]

0x401a2a VMOVSS	0x46bc30(%R12,%RBX,4),%XMM1    [2]

0x401a34 VMOVSS	0x4cd6b0(%R12,%RBX,4),%XMM2    [2]

0x401a3e MOV	$0x40a1b0,%ESI

0x401a43 MOV	$0x46bc30,%EDX

0x401a48 MOV	$0x4cd6b0,%ECX

0x401a4d MOV	$0x52f130,%R8D

0x401a53 MOV	%R14D,%EDI

0x401a56 VMOVSS	0x45c6(%RIP),%XMM3    [3]

0x401a5e VMOVSS	0x45c2(%RIP),%XMM4    [3]

0x401a66 MOV	%R15,%R9

0x401a69 LEA	-0x48(%RBP),%RAX

0x401a6d PUSH	%RAX

0x401a6e LEA	-0x44(%RBP),%RAX

0x401a72 PUSH	%RAX

0x401a73 CALL	401b60 <Step10_orig>

0x401a78 ADD	$0x10,%RSP

0x401a7c VMOVSS	-0x40(%RBP),%XMM0    [1]

0x401a81 VMOVSS	0x45a3(%RIP),%XMM1    [3]

0x401a89 VFMADD213SS	0x590bb0(%R12,%RBX,4),%XMM1,%XMM0    [2]

0x401a93 VMOVSS	%XMM0,0x590bb0(%R12,%RBX,4)    [2]

0x401a9d VMOVSS	-0x44(%RBP),%XMM0    [1]

0x401aa2 VFMADD213SS	0x5f2630(%R12,%RBX,4),%XMM1,%XMM0    [2]

0x401aac VMOVSS	%XMM0,0x5f2630(%R12,%RBX,4)    [2]

0x401ab6 VMOVSS	-0x48(%RBP),%XMM0    [1]

0x401abb VFMADD213SS	0x6540b0(%R12,%RBX,4),%XMM1,%XMM0    [2]

0x401ac5 VMOVSS	%XMM0,0x6540b0(%R12,%RBX,4)    [2]

0x401acf INC	%RBX

0x401ad2 CMP	%RBX,%R13

0x401ad5 JNE	401a20

/home/eoseret/qaas_runs_GNR/173-814-2357/intel/HACCmk/build/HACCmk/src/main.c: 140 - 147

--------------------------------------------------------------------------------

140:       for ( i = 0; i < count; ++i)

141:       {

142:         Step10_orig( n, xx[i], yy[i], zz[i], fsrrmax2, mp_rsm2, xx, yy, zz, mass, &dx1, &dy1, &dz1 );

143:

144:         vx1[i] = vx1[i] + dx1 * fcoeff;

145:         vy1[i] = vy1[i] + dy1 * fcoeff;

146:         vz1[i] = vz1[i] + dz1 * fcoeff;

147:       }

Coverage (%)	Name	Source Location	Module
►100.00+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so

Path /

Metric	Value
CQA speedup if no scalar integer	1.38
CQA speedup if FP arith vectorized	1.38
CQA speedup if fully vectorized	15.48
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.38
Bottlenecks	micro-operation queue,
Function	main.extracted.8
Source	main.c:140-147
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	5.50
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	4.00
CQA cycles if fully vectorized	0.36
Front-end cycles	5.50
P0 cycles	1.70
P1 cycles	1.50
P2 cycles	4.00
P3 cycles	4.00
P4 cycles	3.00
P5 cycles	1.60
P6 cycles	1.60
P7 cycles	3.00
P8 cycles	3.00
P9 cycles	3.00
P10 cycles	1.60
P11 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	30.00
Nb uops	30.00
Nb loads	12.00
Nb stores	3.00
Nb stack references	3.00
FLOP/cycle	1.09
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	3.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	10.91
Bytes prefetched	0.00
Bytes loaded	48.00
Bytes stored	12.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	6.82
Vector-efficiency ratio load	6.25
Vector-efficiency ratio store	6.25
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	6.25
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	7.29

Metric	Value
CQA speedup if no scalar integer	1.38
CQA speedup if FP arith vectorized	1.38
CQA speedup if fully vectorized	15.48
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.38
Bottlenecks	micro-operation queue,
Function	main.extracted.8
Source	main.c:140-147
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	5.50
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	4.00
CQA cycles if fully vectorized	0.36
Front-end cycles	5.50
P0 cycles	1.70
P1 cycles	1.50
P2 cycles	4.00
P3 cycles	4.00
P4 cycles	3.00
P5 cycles	1.60
P6 cycles	1.60
P7 cycles	3.00
P8 cycles	3.00
P9 cycles	3.00
P10 cycles	1.60
P11 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	30.00
Nb uops	30.00
Nb loads	12.00
Nb stores	3.00
Nb stack references	3.00
FLOP/cycle	1.09
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	3.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	10.91
Bytes prefetched	0.00
Bytes loaded	48.00
Bytes stored	12.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	6.82
Vector-efficiency ratio load	6.25
Vector-efficiency ratio store	6.25
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	6.25
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	7.29

Path /

Average path: Display a virtual path defined by average values of all real paths

Function	main.extracted.8
Source file and lines	main.c:140-147
Module	exec

The loop is defined in /home/eoseret/qaas_runs_GNR/173-814-2357/intel/HACCmk/build/HACCmk/src/main.c:140-147.

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 5.50 to 4.00 cycles (1.37x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is not vectorized. Only 6% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your loop, you can lower the cost of an iteration from 5.50 to 0.36 cycles (15.48x speedup).

Details

All SSE/AVX instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

FMA

Detected 3 FMA (fused multiply-add) operations.

CALL instructions

Detected function call instructions.

Details

Calling (and then returning from) a function prevents many compiler optimizations (like vectorization), breaks control flow (which reduces pipeline performance) and executes extra instructions to save/restore the registers used inside it, which is very expensive (dozens of cycles). Consider to inline small functions.

Step10_orig: 1 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2357/intel/HACCmk/build/HACCmk/src/main.c:142

Type of elements and instruction set

3 SSE or AVX instructions are processing arithmetic or math operations on single precision FP elements in scalar mode (one at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 6 FP arithmetical operations:

3: addition or subtraction (all inside FMA instructions)
3: multiply (all inside FMA instructions)

The binary loop is loading 48 bytes (12 single precision FP elements). The binary loop is storing 12 bytes (3 single precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.10 FP operations per loaded or stored byte.

Unroll opportunity

Loop is data access bound.

Workaround

Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)

General properties

nb instructions	30
nb uops	30
loop length	187
used x86 registers	14
used mmx registers	0
used xmm registers	5
used ymm registers	0
used zmm registers	0
nb stack references	3

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	5.50 cycles
front end	5.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	1.70	1.50	4.00	4.00	3.00	1.60	1.60	3.00	3.00	3.00	1.60	4.00
cycles	1.70	1.50	4.00	4.00	3.00	1.60	1.60	3.00	3.00	3.00	1.60	4.00

Execution ports to units layout:

P0 (256 bits): VPU, BRU, ALU, DIV/SQRT
P1 (256 bits): ALU, VPU
P2 (512 bits): load
P3 (512 bits): load
P4 (512 bits): store data
P5 (512 bits): ALU, VPU
P6: ALU, BRU
P7: store address
P8: store address
P9 (512 bits): store data
P10: ALU
P11 (512 bits): load

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	5.50
Dispatch	4.00
Data deps.	1.00
Overall L1	5.50

Vectorization ratios

INT

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	0%

all	0%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	0%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	8%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	7%

all	6%
load	6%
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	6%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	6%
load	6%
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	6%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	7%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 5.50 cycles. At this rate:

4% of peak load performance is reached (8.73 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
1% of peak store performance is reached (2.18 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 5.50 to 4.00 cycles (1.38x speedup).

ASM code

In the binary file, the address of the loop is: 401a20

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput	Vectorization
VMOVSS 0x40a1b0(%R12,%RBX,4),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x46bc30(%R12,%RBX,4),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x4cd6b0(%R12,%RBX,4),%XMM2	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
MOV $0x40a1b0,%ESI	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x46bc30,%EDX	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x4cd6b0,%ECX	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x52f130,%R8D	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV %R14D,%EDI	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (6.3%)
VMOVSS 0x45c6(%RIP),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x45c2(%RIP),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
MOV %R15,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
LEA -0x48(%RBP),%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %RAX	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	5-12	0.50	N/A
LEA -0x44(%RBP),%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %RAX	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	5-12	0.50	N/A
CALL 401b60 <Step10_orig>	2	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	0	1	N/A
ADD $0x10,%RSP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVSS -0x40(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x45a3(%RIP),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x590bb0(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x590bb0(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
VMOVSS -0x44(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x5f2630(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x5f2630(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
VMOVSS -0x48(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x6540b0(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x6540b0(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
INC %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RBX,%R13	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	N/A
JNE 401a20 <main.extracted.8+0x90>	1	0.50	0	0	0	0	0	0.50	0	0	0	0	0	0	0.50	N/A

Function	main.extracted.8
Source file and lines	main.c:140-147
Module	exec

The loop is defined in /home/eoseret/qaas_runs_GNR/173-814-2357/intel/HACCmk/build/HACCmk/src/main.c:140-147.

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 5.50 to 4.00 cycles (1.37x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Details

All SSE/AVX instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

FMA

Detected 3 FMA (fused multiply-add) operations.

CALL instructions

Detected function call instructions.

Details

Step10_orig: 1 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2357/intel/HACCmk/build/HACCmk/src/main.c:142

Type of elements and instruction set

3 SSE or AVX instructions are processing arithmetic or math operations on single precision FP elements in scalar mode (one at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 6 FP arithmetical operations:

3: addition or subtraction (all inside FMA instructions)
3: multiply (all inside FMA instructions)

The binary loop is loading 48 bytes (12 single precision FP elements). The binary loop is storing 12 bytes (3 single precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.10 FP operations per loaded or stored byte.

Unroll opportunity

Loop is data access bound.

Workaround

General properties

nb instructions	30
nb uops	30
loop length	187
used x86 registers	14
used mmx registers	0
used xmm registers	5
used ymm registers	0
used zmm registers	0
nb stack references	3

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	5.50 cycles
front end	5.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	1.70	1.50	4.00	4.00	3.00	1.60	1.60	3.00	3.00	3.00	1.60	4.00
cycles	1.70	1.50	4.00	4.00	3.00	1.60	1.60	3.00	3.00	3.00	1.60	4.00

Execution ports to units layout:

P0 (256 bits): VPU, BRU, ALU, DIV/SQRT
P1 (256 bits): ALU, VPU
P2 (512 bits): load
P3 (512 bits): load
P4 (512 bits): store data
P5 (512 bits): ALU, VPU
P6: ALU, BRU
P7: store address
P8: store address
P9 (512 bits): store data
P10: ALU
P11 (512 bits): load

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	5.50
Dispatch	4.00
Data deps.	1.00
Overall L1	5.50

Vectorization ratios

INT

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	0%

all	0%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	0%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	8%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	7%

all	6%
load	6%
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	6%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	6%
load	6%
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	6%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	7%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 5.50 cycles. At this rate:

4% of peak load performance is reached (8.73 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
1% of peak store performance is reached (2.18 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the loop is: 401a20

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput	Vectorization
VMOVSS 0x40a1b0(%R12,%RBX,4),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x46bc30(%R12,%RBX,4),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x4cd6b0(%R12,%RBX,4),%XMM2	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
MOV $0x40a1b0,%ESI	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x46bc30,%EDX	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x4cd6b0,%ECX	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV $0x52f130,%R8D	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	scal (6.3%)
MOV %R14D,%EDI	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (6.3%)
VMOVSS 0x45c6(%RIP),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x45c2(%RIP),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
MOV %R15,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
LEA -0x48(%RBP),%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %RAX	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	5-12	0.50	N/A
LEA -0x44(%RBP),%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %RAX	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	5-12	0.50	N/A
CALL 401b60 <Step10_orig>	2	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	0	1	N/A
ADD $0x10,%RSP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVSS -0x40(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VMOVSS 0x45a3(%RIP),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x590bb0(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x590bb0(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
VMOVSS -0x44(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x5f2630(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x5f2630(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
VMOVSS -0x48(%RBP),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (6.3%)
VFMADD213SS 0x6540b0(%R12,%RBX,4),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	4	0.50	scal (6.3%)
VMOVSS %XMM0,0x6540b0(%R12,%RBX,4)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (6.3%)
INC %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RBX,%R13	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	N/A
JNE 401a20 <main.extracted.8+0x90>	1	0.50	0	0	0	0	0	0.50	0	0	0	0	0	0	0.50	N/A

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

FMA

CALL instructions

Details

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

Unroll opportunity

Workaround

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

FMA

CALL instructions

Details

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

Unroll opportunity

Workaround

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code