OV - exec - Function main.omp

	Function: main.omp_outlined.27	Module: exec	Source: stream.c:330-332	Coverage (incl. loops): 95.54% \| (excl. loops): 0.00%

/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c: 330 - 332

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

330: #pragma omp parallel for

331: 	for (j=0; j<STREAM_ARRAY_SIZE; j++)

332: 	    b[j] = scalar*c[j];

0x203710 PUSH	%RBP

0x203711 MOV	%RSP,%RBP

0x203714 PUSH	%R15

0x203716 PUSH	%R14

0x203718 PUSH	%RBX

0x203719 SUB	$0x28,%RSP

0x20371d MOV	(%RDI),%EBX

0x20371f LEA	-0x7(%RIP),%RAX

0x203726 MOV	$0x3049,%R15

0x203730 MOV	%RDX,%R14

0x203733 MOVQ	$0,-0x30(%RBP)

0x20373b MOVQ	$0x3344ffff,-0x20(%RBP)

0x203743 MOVQ	$0x1,-0x38(%RBP)

0x20374b MOVL	$0,-0x24(%RBP)

0x203752 ADD	%RAX,%R15

0x203755 SUB	$0x8,%RSP

0x203759 MOV	$-0x1210,%RDI

0x203763 LEA	-0x38(%RBP),%RAX

0x203767 MOV	$-0x1020,%R10

0x203771 LEA	-0x24(%RBP),%RCX

0x203775 LEA	-0x30(%RBP),%R8

0x203779 LEA	-0x20(%RBP),%R9

0x20377d MOV	%EBX,%ESI

0x20377f MOV	$0x22,%EDX

0x203784 ADD	%R15,%RDI

0x203787 PUSH	$0x1

0x203789 PUSH	$0x1

0x20378b PUSH	%RAX

0x20378c CALLQ	(%R15,%R10,1)

0x203790 ADD	$0x20,%RSP

0x203794 MOV	-0x20(%RBP),%RCX

0x203798 MOV	-0x30(%RBP),%RDI

0x20379c MOV	$0x3344ffff,%EAX

0x2037a1 CMP	$0x3344ffff,%RCX

0x2037a8 CMOVL	%RCX,%RAX

0x2037ac MOV	%RAX,-0x20(%RBP)

0x2037b0 CMP	%RAX,%RDI

0x2037b3 JG	2038e6

0x2037b9 VMOVSD	(%R14),%XMM0

0x2037be MOV	%RAX,%RSI

0x2037c1 SUB	%RDI,%RSI

0x2037c4 MOV	$0x334500028,%RCX

0x2037ce MOV	$0x19a280028,%RDX

0x2037d8 INC	%RSI

0x2037db CMP	$0x4,%RSI

0x2037df JAE	2037e9

0x2037e1 MOV	%RDI,%R9

0x2037e4 JMP	2038ba

0x2037e9 CMP	$0x20,%RSI

0x2037ed JAE	2037f7

0x2037ef XOR	%R8D,%R8D

0x2037f2 JMP	203876

0x2037f7 LEA	(%R15,%RCX,1),%R9

0x2037fb LEA	(%R15,%RDX,1),%R10

0x2037ff VBROADCASTSD	%XMM0,%ZMM1

0x203805 MOV	%RSI,%R8

0x203808 AND	$-0x20,%R8

0x20380c XOR	%R11D,%R11D

0x20380f LEA	0xc0(%R9,%RDI,8),%R9

0x203817 LEA	0xc0(%R10,%RDI,8),%R10

0x20381f NOP

(10) 0x203820 VMULPD	-0xc0(%R9,%R11,8),%ZMM1,%ZMM2

(10) 0x203828 VMULPD	-0x80(%R9,%R11,8),%ZMM1,%ZMM3

(10) 0x203830 VMULPD	-0x40(%R9,%R11,8),%ZMM1,%ZMM4

(10) 0x203838 VMULPD	(%R9,%R11,8),%ZMM1,%ZMM5

(10) 0x20383f VMOVUPD	%ZMM2,-0xc0(%R10,%R11,8)

(10) 0x203847 VMOVUPD	%ZMM3,-0x80(%R10,%R11,8)

(10) 0x20384f VMOVUPD	%ZMM4,-0x40(%R10,%R11,8)

(10) 0x203857 VMOVUPD	%ZMM5,(%R10,%R11,8)

(10) 0x20385e ADD	$0x20,%R11

(10) 0x203862 CMP	%R11,%R8

(10) 0x203865 JNE	203820

0x203867 CMP	%R8,%RSI

0x20386a JE	2038e6

0x20386c TEST	$0x1c,%SIL

0x203870 JE	203911

0x203876 MOV	%RSI,%R10

0x203879 LEA	(%R15,%RCX,1),%R11

0x20387d AND	$-0x4,%R10

0x203881 LEA	(%R15,%RDX,1),%R14

0x203885 VBROADCASTSD	%XMM0,%YMM1

0x20388a LEA	(%RDI,%R10,1),%R9

0x20388e LEA	(%R11,%RDI,8),%R11

0x203892 LEA	(%R14,%RDI,8),%RDI

0x203896 NOPW	%CS:(%RAX,%RAX,1)

(12) 0x2038a0 VMULPD	(%R11,%R8,8),%YMM1,%YMM2

(12) 0x2038a6 VMOVUPD	%YMM2,(%RDI,%R8,8)

(12) 0x2038ac ADD	$0x4,%R8

(12) 0x2038b0 CMP	%R8,%R10

(12) 0x2038b3 JNE	2038a0

0x2038b5 CMP	%R10,%RSI

0x2038b8 JE	2038e6

0x2038ba DEC	%R9

0x2038bd ADD	%R15,%RCX

0x2038c0 ADD	%R15,%RDX

0x2038c3 NOPW	%CS:(%RAX,%RAX,1)

(11) 0x2038d0 VMULSD	0x8(%RCX,%R9,8),%XMM0,%XMM1

(11) 0x2038d7 VMOVSD	%XMM1,0x8(%RDX,%R9,8)

(11) 0x2038de INC	%R9

(11) 0x2038e1 CMP	%RAX,%R9

(11) 0x2038e4 JL	2038d0

0x2038e6 MOV	$-0x11f8,%RDI

0x2038f0 MOV	$-0x1010,%RAX

0x2038fa MOV	%EBX,%ESI

0x2038fc ADD	%R15,%RDI

0x2038ff VZEROUPPER

0x203902 CALLQ	(%R15,%RAX,1)

0x203906 ADD	$0x28,%RSP

0x20390a POP	%RBX

0x20390b POP	%R14

0x20390d POP	%R15

0x20390f POP	%RBP

0x203910 RET

0x203911 ADD	%R8,%RDI

0x203914 JMP	2037e1

0x203919 NOPL	(%RAX)

Coverage (%)	Name	Source Location	Module
►99.83+	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so

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 /

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

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run orig_0

Source file and lines	stream.c:330-332
Module	exec

The function is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:330-332.

Warnings:
Ignoring paths for analysis

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 12.50 to 2.50 cycles (5.00x speedup).

Workaround

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

Vectorization

Your function is probably not vectorized. Only 11% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your function, you can lower the cost of an iteration from 12.50 to 0.94 cycles (13.33x speedup).

Details

Store and arithmetical 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 function can use their full power.

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function 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.

No data for this section

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.

unknown: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:330
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:330

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

CMOVL: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331
LEA: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 44 bytes. The binary function is storing 36 bytes.

General properties

nb instructions	95
nb uops	100
loop length	414
used x86 registers	14
used mmx registers	0
used xmm registers	1
used ymm registers	1
used zmm registers	1
nb stack references	4

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	12.50 cycles
front end	12.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	9.83	9.83	9.83	9.83	9.83	9.83	3.75	3.75	3.75	3.75	0.00	1.00	1.00	0.00	0.00	0.00
cycles	9.83	9.83	9.83	9.83	9.83	9.83	3.75	3.75	3.75	3.75	0.00	1.00	1.00	0.00	0.00	0.00

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

Cycles summary

Front-end	12.50
Dispatch	9.83
Overall L1	12.50

Vectorization ratios

INT

all	4%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	10%

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

INT+FP

all	4%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	8%

Vector efficiency ratios

INT

all	11%
load	6%
store	7%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

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

INT+FP

all	11%
load	9%
store	7%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 12.50 cycles. At this rate:

1% of peak load performance is reached (3.52 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.88 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 12.50 to 9.83 cycles (1.27x speedup).

ASM code

In the binary file, the address of the function is: 203710

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P11	P12	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV (%RDI),%EBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
LEA -0x7(%RIP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x3049,%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RDX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOVQ $0,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVQ $0x3344ffff,-0x20(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVQ $0x1,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVL $0,-0x24(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
ADD %RAX,%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB $0x8,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV $-0x1210,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x38(%RBP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $-0x1020,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x24(%RBP),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x30(%RBP),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x20(%RBP),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %EBX,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
MOV $0x22,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALLQ (%R15,%R10,1)	4	0	0	0	0.33	0.33	0.33	0.50	0.50	0.50	0.50	0	0	0	2.20	N/A
ADD $0x20,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV -0x20(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV -0x30(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV $0x3344ffff,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP $0x3344ffff,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVL %RCX,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
MOV %RAX,-0x20(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
CMP %RAX,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JG 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSD (%R14),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOV %RAX,%RSI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
SUB %RDI,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV $0x334500028,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x19a280028,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
INC %RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMP $0x4,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 2037e9 <main.omp_outlined.27+0xd9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RDI,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 2038ba <main.omp_outlined.27+0x1aa>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
CMP $0x20,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 2037f7 <main.omp_outlined.27+0xe7>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
XOR %R8D,%R8D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 203876 <main.omp_outlined.27+0x166>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
LEA (%R15,%RCX,1),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R15,%RDX,1),%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSD %XMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
MOV %RSI,%R8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %R11D,%R11D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA 0xc0(%R9,%RDI,8),%R9	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
LEA 0xc0(%R10,%RDI,8),%R10	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x1c,%SIL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JE 203911 <main.omp_outlined.27+0x201>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RSI,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%R15,%RCX,1),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
AND $-0x4,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R15,%RDX,1),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSD %XMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
LEA (%RDI,%R10,1),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R11,%RDI,8),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R14,%RDI,8),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
CMP %R10,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
DEC %R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV $-0x11f8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $-0x1010,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %EBX,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
ADD %R15,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALLQ (%R15,%RAX,1)	4	0	0	0	0.33	0.33	0.33	0.50	0.50	0.50	0.50	0	0	0	2.20	N/A
ADD $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50	N/A
ADD %R8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JMP 2037e1 <main.omp_outlined.27+0xd1>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run orig_0

Source file and lines	stream.c:330-332
Module	exec

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 12.50 to 2.50 cycles (5.00x speedup).

Workaround

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

Vectorization

Details

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function 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.

No data for this section

CALL instructions

Detected function call instructions.

Details

unknown: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:330
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:330

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

CMOVL: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331
LEA: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/177-365-4625/stream-scale/build/stream-scale/src/stream.c:331

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 44 bytes. The binary function is storing 36 bytes.

General properties

nb instructions	95
nb uops	100
loop length	414
used x86 registers	14
used mmx registers	0
used xmm registers	1
used ymm registers	1
used zmm registers	1
nb stack references	4

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	12.50 cycles
front end	12.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	9.83	9.83	9.83	9.83	9.83	9.83	3.75	3.75	3.75	3.75	0.00	1.00	1.00	0.00	0.00	0.00
cycles	9.83	9.83	9.83	9.83	9.83	9.83	3.75	3.75	3.75	3.75	0.00	1.00	1.00	0.00	0.00	0.00

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

Cycles summary

Front-end	12.50
Dispatch	9.83
Overall L1	12.50

Vectorization ratios

INT

all	4%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	10%

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

INT+FP

all	4%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	8%

Vector efficiency ratios

INT

all	11%
load	6%
store	7%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

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

INT+FP

all	11%
load	9%
store	7%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 12.50 cycles. At this rate:

1% of peak load performance is reached (3.52 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.88 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the function is: 203710

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P11	P12	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV (%RDI),%EBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
LEA -0x7(%RIP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x3049,%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RDX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOVQ $0,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVQ $0x3344ffff,-0x20(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVQ $0x1,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVL $0,-0x24(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
ADD %RAX,%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB $0x8,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV $-0x1210,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x38(%RBP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $-0x1020,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x24(%RBP),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x30(%RBP),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA -0x20(%RBP),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %EBX,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
MOV $0x22,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALLQ (%R15,%R10,1)	4	0	0	0	0.33	0.33	0.33	0.50	0.50	0.50	0.50	0	0	0	2.20	N/A
ADD $0x20,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV -0x20(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV -0x30(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV $0x3344ffff,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP $0x3344ffff,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVL %RCX,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
MOV %RAX,-0x20(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
CMP %RAX,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JG 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSD (%R14),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOV %RAX,%RSI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
SUB %RDI,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV $0x334500028,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x19a280028,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
INC %RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMP $0x4,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 2037e9 <main.omp_outlined.27+0xd9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RDI,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 2038ba <main.omp_outlined.27+0x1aa>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
CMP $0x20,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 2037f7 <main.omp_outlined.27+0xe7>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
XOR %R8D,%R8D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 203876 <main.omp_outlined.27+0x166>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
LEA (%R15,%RCX,1),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R15,%RDX,1),%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSD %XMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
MOV %RSI,%R8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %R11D,%R11D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA 0xc0(%R9,%RDI,8),%R9	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
LEA 0xc0(%R10,%RDI,8),%R10	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x1c,%SIL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JE 203911 <main.omp_outlined.27+0x201>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RSI,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%R15,%RCX,1),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
AND $-0x4,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R15,%RDX,1),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSD %XMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
LEA (%RDI,%R10,1),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R11,%RDI,8),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R14,%RDI,8),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
CMP %R10,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 2038e6 <main.omp_outlined.27+0x1d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
DEC %R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD %R15,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV $-0x11f8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $-0x1010,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %EBX,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
ADD %R15,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALLQ (%R15,%RAX,1)	4	0	0	0	0.33	0.33	0.33	0.50	0.50	0.50	0.50	0	0	0	2.20	N/A
ADD $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50	N/A
ADD %R8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JMP 2037e1 <main.omp_outlined.27+0xd1>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A

Name	Coverage (%)	Time (s)
▼main.omp_outlined.27–	95.54	4.18
○Loop 10 - stream.c:332-332 - exec	95.54	4.11
○Loop 11 - stream.c:331-332 - exec	0.00	0.00
○Loop 12 - stream.c:332-332 - exec	0.00	0.00

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

CALL instructions

Details

Complex instructions

Details

Type of elements and instruction set

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

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

CALL instructions

Details

Complex instructions

Details

Type of elements and instruction set

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code