OV - exec - Loop 244

0x42de70 MOVUPD	(%RAX,%RSI,8),%XMM6    [2]

0x42de75 MOVAPD	%XMM3,%XMM7

0x42de79 DIVPD	%XMM6,%XMM7

0x42de7d MOVAPD	%XMM6,%XMM8

0x42de82 MULPD	%XMM4,%XMM8

0x42de87 MOVUPD	(%RCX,%RSI,8),%XMM9    [4]

0x42de8d MULPD	%XMM8,%XMM9

0x42de92 MOVUPD	%XMM9,(%R11,%RSI,8)    [3]

0x42de98 MULPD	%XMM5,%XMM6

0x42de9c MULPD	%XMM7,%XMM7

0x42dea0 MULPD	%XMM9,%XMM6

0x42dea5 MULPD	%XMM7,%XMM6

0x42dea9 SQRTPD	%XMM6,%XMM6

0x42dead MOVUPD	%XMM6,(%R9,%RSI,8)    [1]

0x42deb3 ADD	$0x2,%RSI

0x42deb7 CMP	%RDX,%RSI

0x42deba JBE	42de70

/home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90: 50 - 56

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

50:         v=1.0_8/density(j,k)

51:         pressure(j,k)=(1.4_8-1.0_8)*density(j,k)*energy(j,k)

52:         pressurebyenergy=(1.4_8-1.0_8)*density(j,k)

53:         pressurebyvolume=-density(j,k)*pressure(j,k)

54:         sound_speed_squared=v*v*(pressure(j,k)*pressurebyenergy-pressurebyvolume)

55:         soundspeed(j,k)=SQRT(sound_speed_squared)

56:       ENDDO

Coverage (%)	Name	Source Location	Module
►54.43+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►34.57+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
►4.16+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
○	start_thread		libc.so.6
►2.43+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
►1.98+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ideal_gas_kernel	ideal_gas_kernel.f90:45	exec
○	ideal_gas	ideal_gas.f90:51	exec
○	timestep	timestep.f90:62	exec
○	hydro	hydro.f90:54	exec
○	main		exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start		exec
►1.90+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ideal_gas_kernel	ideal_gas_kernel.f90:45	exec
○	ideal_gas	ideal_gas.f90:51	exec
○	pdv	PdV.f90:126	exec
○	hydro	hydro.f90:56	exec
○	main		exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start		exec

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	4.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	3.13
Bottlenecks	P10, P11,
Function	ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2
Source	ideal_gas_kernel.f90:50-56
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	12.50
CQA cycles if no scalar integer	12.50
CQA cycles if FP arith vectorized	3.13
CQA cycles if fully vectorized	3.13
Front-end cycles	2.00
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	1.00
P7 cycles	1.00
P8 cycles	1.00
P9 cycles	1.00
P10 cycles	4.00
P11 cycles	4.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	1.00
P15 cycles	1.00
DIV/SQRT cycles	12.50
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	17.00
Nb uops	16.00
Nb loads	2.00
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	1.28
Nb FLOP add-sub	0.00
Nb FLOP mul	12.00
Nb FLOP fma	0.00
Nb FLOP div	2.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	2.00
Nb FLOP rsqrt	0.00
Bytes/cycle	5.12
Bytes prefetched	0.00
Bytes loaded	32.00
Bytes stored	32.00
Stride 0	0.00
Stride 1	4.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	100.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	25.00
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	25.00
Vector-efficiency ratio other	25.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	4.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	3.13
Bottlenecks	P10, P11,
Function	ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2
Source	ideal_gas_kernel.f90:50-56
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	12.50
CQA cycles if no scalar integer	12.50
CQA cycles if FP arith vectorized	3.13
CQA cycles if fully vectorized	3.13
Front-end cycles	2.00
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	1.00
P7 cycles	1.00
P8 cycles	1.00
P9 cycles	1.00
P10 cycles	4.00
P11 cycles	4.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	1.00
P15 cycles	1.00
DIV/SQRT cycles	12.50
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	17.00
Nb uops	16.00
Nb loads	2.00
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	1.28
Nb FLOP add-sub	0.00
Nb FLOP mul	12.00
Nb FLOP fma	0.00
Nb FLOP div	2.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	2.00
Nb FLOP rsqrt	0.00
Bytes/cycle	5.12
Bytes prefetched	0.00
Bytes loaded	32.00
Bytes stored	32.00
Stride 0	0.00
Stride 1	4.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	100.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	25.00
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	25.00
Vector-efficiency ratio other	25.00

Path /

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

Function	ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2
Source file and lines	ideal_gas_kernel.f90:50-56
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:50-56.

It is main loop of related source loop which is unrolled by 2 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).

By fully vectorizing your loop, you can lower the cost of an iteration from 12.50 to 3.12 cycles (4.00x speedup).

Details

All SSE/AVX instructions are used in vector version (process two or more data elements in vector registers). Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Pass to your compiler a micro-architecture specialization option:
- use axHost or xHost
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: compile with -align array64byte (remark: not affects arrays in COMMON blocks).
2. inform your compiler that your arrays are vector aligned: Append !DIR$ VECTOR ALIGNED to the loop if all accessed arrays are aligned, or !DIR$ ASSUME_ALIGNED FOO: 64 if only FOO is aligned.
Use the LOOP COUNT directive

Execution units bottlenecks

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 12.50 to 4.00 cycles (3.13x speedup).

Workaround

Reduce the number of division or square root instructions:
- If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with no-prec-div or Ofast
Check whether you really need double precision. If not, switch to single precision to speedup execution

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 12.50 to 3.00 cycles (4.17x speedup).

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVUPD: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:50
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:51
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:51
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:55

Workaround

Pass to your compiler a micro-architecture specialization option:
- use axHost or xHost
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: compile with -align array64byte (remark: not affects arrays in COMMON blocks).
2. inform your compiler that your arrays are vector aligned: Append !DIR$ VECTOR ALIGNED to the loop if all accessed arrays are aligned, or !DIR$ ASSUME_ALIGNED FOO: 64 if only FOO is aligned.

Type of elements and instruction set

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in vector mode (two at a time).

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

The binary loop is composed of 16 FP arithmetical operations:

12: multiply
2: divide
2: square root

The binary loop is loading 32 bytes (4 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.25 FP operations per loaded or stored byte.

General properties

nb instructions	17
nb uops	16
loop length	76
used x86 registers	6
used mmx registers	0
used xmm registers	7
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	2.00 cycles
front end	2.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	1.00	1.00	1.00	1.00	4.00	4.00	0.00	0.00	1.00	1.00
cycles	0.33	0.33	0.33	0.33	0.33	0.33	1.00	1.00	1.00	1.00	4.00	4.00	0.00	0.00	1.00	1.00

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	12.50
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.00
Dispatch	4.00
DIV/SQRT	12.50
Data deps.	1.00
Overall L1	12.50

Vectorization ratios

all	100%
load	100%
store	100%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	100%

Vector efficiency ratios

all	25%
load	25%
store	25%
mul	25%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	25%
other	25%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 12.50 cycles. At this rate:

1% of peak load performance is reached (2.56 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.56 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 42de70

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P14	P15	Latency	Recip. throughput	Vectorization
MOVUPD (%RAX,%RSI,8),%XMM6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.50	vect (25.0%)
MOVAPD %XMM3,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
DIVPD %XMM6,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	13	4	vect (25.0%)
MOVAPD %XMM6,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MULPD %XMM4,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MOVUPD (%RCX,%RSI,8),%XMM9	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.50	vect (25.0%)
MULPD %XMM8,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MOVUPD %XMM9,(%R11,%RSI,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
MULPD %XMM5,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM7,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM9,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM7,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
SQRTPD %XMM6,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	21	8.50	vect (25.0%)
MOVUPD %XMM6,(%R9,%RSI,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
ADD $0x2,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RDX,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JBE 42de70 <ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2+0x290>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Function	ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2
Source file and lines	ideal_gas_kernel.f90:50-56
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:50-56.

It is main loop of related source loop which is unrolled by 2 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).

By fully vectorizing your loop, you can lower the cost of an iteration from 12.50 to 3.12 cycles (4.00x speedup).

Details

All SSE/AVX instructions are used in vector version (process two or more data elements in vector registers). Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Pass to your compiler a micro-architecture specialization option:
- use axHost or xHost
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: compile with -align array64byte (remark: not affects arrays in COMMON blocks).
2. inform your compiler that your arrays are vector aligned: Append !DIR$ VECTOR ALIGNED to the loop if all accessed arrays are aligned, or !DIR$ ASSUME_ALIGNED FOO: 64 if only FOO is aligned.
Use the LOOP COUNT directive

Execution units bottlenecks

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 12.50 to 4.00 cycles (3.13x speedup).

Workaround

Reduce the number of division or square root instructions:
- If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with no-prec-div or Ofast
Check whether you really need double precision. If not, switch to single precision to speedup execution

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 12.50 to 3.00 cycles (4.17x speedup).

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVUPD: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:50
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:51
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:51
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/ideal_gas_kernel.f90:55

Workaround

Pass to your compiler a micro-architecture specialization option:
- use axHost or xHost
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: compile with -align array64byte (remark: not affects arrays in COMMON blocks).
2. inform your compiler that your arrays are vector aligned: Append !DIR$ VECTOR ALIGNED to the loop if all accessed arrays are aligned, or !DIR$ ASSUME_ALIGNED FOO: 64 if only FOO is aligned.

Type of elements and instruction set

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in vector mode (two at a time).

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

The binary loop is composed of 16 FP arithmetical operations:

12: multiply
2: divide
2: square root

The binary loop is loading 32 bytes (4 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.25 FP operations per loaded or stored byte.

General properties

nb instructions	17
nb uops	16
loop length	76
used x86 registers	6
used mmx registers	0
used xmm registers	7
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	2.00 cycles
front end	2.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	1.00	1.00	1.00	1.00	4.00	4.00	0.00	0.00	1.00	1.00
cycles	0.33	0.33	0.33	0.33	0.33	0.33	1.00	1.00	1.00	1.00	4.00	4.00	0.00	0.00	1.00	1.00

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	12.50
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.00
Dispatch	4.00
DIV/SQRT	12.50
Data deps.	1.00
Overall L1	12.50

Vectorization ratios

all	100%
load	100%
store	100%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	100%

Vector efficiency ratios

all	25%
load	25%
store	25%
mul	25%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	25%
other	25%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 12.50 cycles. At this rate:

1% of peak load performance is reached (2.56 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.56 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 42de70

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P14	P15	Latency	Recip. throughput	Vectorization
MOVUPD (%RAX,%RSI,8),%XMM6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.50	vect (25.0%)
MOVAPD %XMM3,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
DIVPD %XMM6,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	13	4	vect (25.0%)
MOVAPD %XMM6,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MULPD %XMM4,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MOVUPD (%RCX,%RSI,8),%XMM9	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.50	vect (25.0%)
MULPD %XMM8,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MOVUPD %XMM9,(%R11,%RSI,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
MULPD %XMM5,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM7,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM9,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
MULPD %XMM7,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50	vect (25.0%)
SQRTPD %XMM6,%XMM6	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	21	8.50	vect (25.0%)
MOVUPD %XMM6,(%R9,%RSI,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
ADD $0x2,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RDX,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JBE 42de70 <ideal_gas_kernel_module_mp_ideal_gas_kernel_.DIR.OMP.PARALLEL.2+0x290>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Report Configuration

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Expensive FP math instructions/calls

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Expensive FP math instructions/calls

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code