OV - exec - Function main.omp

	Function: main.omp_outlined.33	Module: exec	Source: stream.c:354-356	Coverage (incl. loops): 29.27% \| (excl. loops): 0.00%

/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c: 354 - 356

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

354: #pragma omp parallel for

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

356: 	    a[j] = b[j]+scalar*c[j];

0x203f70 PUSH	%RBP

0x203f71 MOV	%RSP,%RBP

0x203f74 PUSH	%R15

0x203f76 PUSH	%R14

0x203f78 PUSH	%R13

0x203f7a PUSH	%R12

0x203f7c PUSH	%RBX

0x203f7d SUB	$0x28,%RSP

0x203f81 MOV	(%RDI),%R14D

0x203f84 LEA	0x2f7d(%RIP),%RBX

0x203f8b MOV	%RDX,%R15

0x203f8e MOVQ	$0,-0x40(%RBP)

0x203f96 MOVQ	$0x3344ffff,-0x30(%RBP)

0x203f9e MOVQ	$0x1,-0x48(%RBP)

0x203fa6 MOVL	$0,-0x34(%RBP)

0x203fad SUB	$0x8,%RSP

0x203fb1 MOV	$0x1228,%RDI

0x203fbb LEA	-0x48(%RBP),%RAX

0x203fbf LEA	-0x40(%RBP),%R8

0x203fc3 LEA	-0x34(%RBP),%RCX

0x203fc7 LEA	-0x30(%RBP),%R9

0x203fcb MOV	%R14D,%ESI

0x203fce MOV	$0x22,%EDX

0x203fd3 ADD	%RBX,%RDI

0x203fd6 PUSH	$0x1

0x203fd8 PUSH	$0x1

0x203fda PUSH	%RAX

0x203fdb CALL	204d00 <@plt_start@+0x80>

0x203fe0 ADD	$0x20,%RSP

0x203fe4 MOV	-0x30(%RBP),%RCX

0x203fe8 MOV	-0x40(%RBP),%R8

0x203fec MOV	$0x3344ffff,%EAX

0x203ff1 CMP	$0x3344ffff,%RCX

0x203ff8 CMOVL	%RCX,%RAX

0x203ffc MOV	%RAX,-0x30(%RBP)

0x204000 CMP	%RAX,%R8

0x204003 JG	20418d

0x204009 VMOVSD	(%R15),%XMM0

0x20400e MOV	%RAX,%RDI

0x204011 SUB	%R8,%RDI

0x204014 MOV	$0x19a281278,%RCX

0x20401e MOV	$0x334501278,%RDX

0x204028 MOV	$0x1278,%RSI

0x204032 INC	%RDI

0x204035 CMP	$0x4,%RDI

0x204039 JAE	204043

0x20403b MOV	%R8,%R10

0x20403e JMP	204160

0x204043 CMP	$0x20,%RDI

0x204047 JAE	204051

0x204049 XOR	%R9D,%R9D

0x20404c JMP	204109

0x204051 LEA	(%RBX,%RCX,1),%R10

0x204055 LEA	(%RBX,%RDX,1),%R11

0x204059 LEA	(%RBX,%RSI,1),%R15

0x20405d VBROADCASTSD	%XMM0,%ZMM1

0x204063 MOV	%RDI,%R9

0x204066 AND	$-0x20,%R9

0x20406a XOR	%R12D,%R12D

0x20406d LEA	0xc0(%R10,%R8,8),%R10

0x204075 LEA	0xc0(%R11,%R8,8),%R11

0x20407d LEA	0xc0(%R15,%R8,8),%R15

0x204085 NOPW	%CS:(%RAX,%RAX,1)

(16) 0x204090 VMOVUPD	-0xc0(%R11,%R12,8),%ZMM2

(16) 0x204098 VMOVUPD	-0x80(%R11,%R12,8),%ZMM3

(16) 0x2040a0 VMOVUPD	-0x40(%R11,%R12,8),%ZMM4

(16) 0x2040a8 VMOVUPD	(%R11,%R12,8),%ZMM5

(16) 0x2040af VFMADD213PD	-0xc0(%R10,%R12,8),%ZMM1,%ZMM2

(16) 0x2040b7 VFMADD213PD	-0x80(%R10,%R12,8),%ZMM1,%ZMM3

(16) 0x2040bf VFMADD213PD	-0x40(%R10,%R12,8),%ZMM1,%ZMM4

(16) 0x2040c7 VFMADD213PD	(%R10,%R12,8),%ZMM1,%ZMM5

(16) 0x2040ce VMOVUPD	%ZMM2,-0xc0(%R15,%R12,8)

(16) 0x2040d6 VMOVUPD	%ZMM3,-0x80(%R15,%R12,8)

(16) 0x2040de VMOVUPD	%ZMM4,-0x40(%R15,%R12,8)

(16) 0x2040e6 VMOVUPD	%ZMM5,(%R15,%R12,8)

(16) 0x2040ed ADD	$0x20,%R12

(16) 0x2040f1 CMP	%R12,%R9

(16) 0x2040f4 JNE	204090

0x2040f6 CMP	%R9,%RDI

0x2040f9 JE	20418d

0x2040ff TEST	$0x1c,%DIL

0x204103 JE	2041b7

0x204109 MOV	%RDI,%R11

0x20410c LEA	(%RBX,%RCX,1),%R15

0x204110 LEA	(%RBX,%RDX,1),%R12

0x204114 AND	$-0x4,%R11

0x204118 LEA	(%RBX,%RSI,1),%R13

0x20411c VBROADCASTSD	%XMM0,%YMM1

0x204121 LEA	(%R8,%R11,1),%R10

0x204125 LEA	(%R15,%R8,8),%R15

0x204129 LEA	(%R12,%R8,8),%R12

0x20412d LEA	(%R13,%R8,8),%R8

0x204132 NOPW	%CS:(%RAX,%RAX,1)

(18) 0x204140 VMOVUPD	(%R12,%R9,8),%YMM2

(18) 0x204146 VFMADD213PD	(%R15,%R9,8),%YMM1,%YMM2

(18) 0x20414c VMOVUPD	%YMM2,(%R8,%R9,8)

(18) 0x204152 ADD	$0x4,%R9

(18) 0x204156 CMP	%R9,%R11

(18) 0x204159 JNE	204140

0x20415b CMP	%R11,%RDI

0x20415e JE	20418d

0x204160 DEC	%R10

0x204163 ADD	%RBX,%RCX

0x204166 ADD	%RBX,%RDX

0x204169 ADD	%RBX,%RSI

0x20416c NOPL	(%RAX)

(17) 0x204170 VMOVSD	0x8(%RDX,%R10,8),%XMM1

(17) 0x204177 VFMADD213SD	0x8(%RCX,%R10,8),%XMM0,%XMM1

(17) 0x20417e VMOVSD	%XMM1,0x8(%RSI,%R10,8)

(17) 0x204185 INC	%R10

(17) 0x204188 CMP	%RAX,%R10

(17) 0x20418b JL	204170

0x20418d MOV	$0x1240,%RAX

0x204197 MOV	%R14D,%ESI

0x20419a ADD	%RAX,%RBX

0x20419d MOV	%RBX,%RDI

0x2041a0 VZEROUPPER

0x2041a3 CALL	204d20 <@plt_start@+0xa0>

0x2041a8 ADD	$0x28,%RSP

0x2041ac POP	%RBX

0x2041ad POP	%R12

0x2041af POP	%R13

0x2041b1 POP	%R14

0x2041b3 POP	%R15

0x2041b5 POP	%RBP

0x2041b6 RET

0x2041b7 ADD	%R9,%R8

0x2041ba JMP	20403b

0x2041bf NOP

Coverage (%)	Name	Source Location	Module
►99.57+	__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:354-356
Module	exec

The function is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:354-356.

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 17.17 to 3.00 cycles (5.72x 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 17.17 to 1.25 cycles (13.71x 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/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:354
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:354

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.

LEA: 3 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355

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 28 bytes. The binary function is storing 36 bytes.

General properties

nb instructions	102
nb uops	103
loop length	434
used x86 registers	16
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	17.17 cycles
front end	17.17 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	13.50	13.50	13.50	13.50	6.00	4.67	4.67	4.67	0.00	1.00	1.00	0.00	0.00	0.00
cycles	13.50	13.50	13.50	13.50	6.00	4.67	4.67	4.67	0.00	1.00	1.00	0.00	0.00	0.00

Execution ports to units layout:

ALU0/BRU0: ALU, BRU
ALU1: ALU
ALU2: ALU
ALU3: ALU
BRU1: BRU
AGU0 (256 bits): store address, load
AGU1 (256 bits): store address, load
AGU2 (256 bits): store address, load
FP0 (256 bits): VPU, DIV/SQRT
FP1 (256 bits): VPU, DIV/SQRT
FP2 (256 bits): VPU
FP3 (256 bits): VPU
FP4 (256 bits): FP store data
FP5 (256 bits): FP store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	17.17
Dispatch	13.50
Overall L1	17.17

Vectorization ratios

INT

all	5%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	14%

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	5%
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	11%

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	14%

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	13%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 17.17 cycles. At this rate:

1% of peak load performance is reached (1.63 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (2.10 out of 64.00 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 17.17 to 13.50 cycles (1.27x speedup).

ASM code

In the binary file, the address of the function is: 203f70

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP1	FP2	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	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	0.17	N/A
PUSH %R15	1	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	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	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	1	0.50	N/A
SUB $0x28,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV (%RDI),%R14D	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	scal (6.3%)
LEA 0x2f7d(%RIP),%RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %RDX,%R15	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOVQ $0,-0x40(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVQ $0x3344ffff,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVQ $0x1,-0x48(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVL $0,-0x34(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
SUB $0x8,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV $0x1228,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
LEA -0x48(%RBP),%RAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x40(%RBP),%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x34(%RBP),%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x30(%RBP),%R9	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %R14D,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV $0x22,%EDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
PUSH $0x1	1	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	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALL 204d00 <@plt_start@+0x80>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0.50	N/A
ADD $0x20,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV -0x30(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	N/A
MOV -0x40(%RBP),%R8	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	N/A
MOV $0x3344ffff,%EAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
CMP $0x3344ffff,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMOVL %RCX,%RAX	1	0.50	0	0	0.50	0	0	0	0	0	0	1	0.50	N/A
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (12.5%)
CMP %RAX,%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JG 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
VMOVSD (%R15),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	1	0.50	scal (12.5%)
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
SUB %R8,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV $0x19a281278,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV $0x334501278,%RDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV $0x1278,%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
INC %RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
CMP $0x4,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 204043 <main.omp_outlined.33+0xd3>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
MOV %R8,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 204160 <main.omp_outlined.33+0x1f0>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
CMP $0x20,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 204051 <main.omp_outlined.33+0xe1>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
XOR %R9D,%R9D	1	0	0	0	0	0	0	0	0	0	0	0	0.25	N/A
JMP 204109 <main.omp_outlined.33+0x199>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
LEA (%RBX,%RCX,1),%R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RDX,1),%R11	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RSI,1),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VBROADCASTSD %XMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0.50	0.50	1	1	scal (12.5%)
MOV %RDI,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x20,%R9	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
XOR %R12D,%R12D	1	0	0	0	0	0	0	0	0	0	0	0	0.25	N/A
LEA 0xc0(%R10,%R8,8),%R10	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
LEA 0xc0(%R11,%R8,8),%R11	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
LEA 0xc0(%R15,%R8,8),%R15	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R9,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
TEST $0x1c,%DIL	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
JE 2041b7 <main.omp_outlined.33+0x247>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
MOV %RDI,%R11	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%RBX,%RCX,1),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RDX,1),%R12	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
AND $-0x4,%R11	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RSI,1),%R13	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VBROADCASTSD %XMM0,%YMM1	1	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
LEA (%R8,%R11,1),%R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R15,%R8,8),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R12,%R8,8),%R12	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R13,%R8,8),%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R11,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
DEC %R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
MOV $0x1240,%RAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV %R14D,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
ADD %RAX,%RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 204d20 <@plt_start@+0xa0>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0.50	N/A
ADD $0x28,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
RET	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50	N/A
ADD %R9,%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
JMP 20403b <main.omp_outlined.33+0xcb>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0.09	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:354-356
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 17.17 to 3.00 cycles (5.72x 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/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:354
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:354

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.

LEA: 3 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz16.benchmarkcenter.megware.com/177-374-0958/stream/build/stream/src/stream.c:355

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 28 bytes. The binary function is storing 36 bytes.

General properties

nb instructions	102
nb uops	103
loop length	434
used x86 registers	16
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	17.17 cycles
front end	17.17 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	13.50	13.50	13.50	13.50	6.00	4.67	4.67	4.67	0.00	1.00	1.00	0.00	0.00	0.00
cycles	13.50	13.50	13.50	13.50	6.00	4.67	4.67	4.67	0.00	1.00	1.00	0.00	0.00	0.00

Execution ports to units layout:

ALU0/BRU0: ALU, BRU
ALU1: ALU
ALU2: ALU
ALU3: ALU
BRU1: BRU
AGU0 (256 bits): store address, load
AGU1 (256 bits): store address, load
AGU2 (256 bits): store address, load
FP0 (256 bits): VPU, DIV/SQRT
FP1 (256 bits): VPU, DIV/SQRT
FP2 (256 bits): VPU
FP3 (256 bits): VPU
FP4 (256 bits): FP store data
FP5 (256 bits): FP store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	17.17
Dispatch	13.50
Overall L1	17.17

Vectorization ratios

INT

all	5%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	14%

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	5%
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	11%

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	14%

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	13%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 17.17 cycles. At this rate:

1% of peak load performance is reached (1.63 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (2.10 out of 64.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the function is: 203f70

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP1	FP2	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	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	0.17	N/A
PUSH %R15	1	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	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	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	1	0.50	N/A
SUB $0x28,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV (%RDI),%R14D	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	scal (6.3%)
LEA 0x2f7d(%RIP),%RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %RDX,%R15	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOVQ $0,-0x40(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVQ $0x3344ffff,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVQ $0x1,-0x48(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
MOVL $0,-0x34(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (6.3%)
SUB $0x8,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV $0x1228,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
LEA -0x48(%RBP),%RAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x40(%RBP),%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x34(%RBP),%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA -0x30(%RBP),%R9	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %R14D,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV $0x22,%EDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
PUSH $0x1	1	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	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALL 204d00 <@plt_start@+0x80>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0.50	N/A
ADD $0x20,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV -0x30(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	N/A
MOV -0x40(%RBP),%R8	1	0	0	0	0	0	0.33	0.33	0.33	0	0	3	0.33	N/A
MOV $0x3344ffff,%EAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
CMP $0x3344ffff,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMOVL %RCX,%RAX	1	0.50	0	0	0.50	0	0	0	0	0	0	1	0.50	N/A
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (12.5%)
CMP %RAX,%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JG 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
VMOVSD (%R15),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	1	0.50	scal (12.5%)
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
SUB %R8,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV $0x19a281278,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV $0x334501278,%RDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV $0x1278,%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
INC %RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
CMP $0x4,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 204043 <main.omp_outlined.33+0xd3>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
MOV %R8,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 204160 <main.omp_outlined.33+0x1f0>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
CMP $0x20,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 204051 <main.omp_outlined.33+0xe1>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
XOR %R9D,%R9D	1	0	0	0	0	0	0	0	0	0	0	0	0.25	N/A
JMP 204109 <main.omp_outlined.33+0x199>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
LEA (%RBX,%RCX,1),%R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RDX,1),%R11	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RSI,1),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VBROADCASTSD %XMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0.50	0.50	1	1	scal (12.5%)
MOV %RDI,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x20,%R9	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
XOR %R12D,%R12D	1	0	0	0	0	0	0	0	0	0	0	0	0.25	N/A
LEA 0xc0(%R10,%R8,8),%R10	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
LEA 0xc0(%R11,%R8,8),%R11	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
LEA 0xc0(%R15,%R8,8),%R15	2	0	0	0.50	0.50	0	0.33	0.33	0.33	0	0	2	0.50	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R9,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
TEST $0x1c,%DIL	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
JE 2041b7 <main.omp_outlined.33+0x247>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
MOV %RDI,%R11	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%RBX,%RCX,1),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RDX,1),%R12	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
AND $-0x4,%R11	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%RBX,%RSI,1),%R13	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VBROADCASTSD %XMM0,%YMM1	1	0	0	0	0	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
LEA (%R8,%R11,1),%R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R15,%R8,8),%R15	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R12,%R8,8),%R12	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
LEA (%R13,%R8,8),%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
CMP %R11,%RDI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 20418d <main.omp_outlined.33+0x21d>	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50-1	N/A
DEC %R10	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %RBX,%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
MOV $0x1240,%RAX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33	N/A
MOV %R14D,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
ADD %RAX,%RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 204d20 <@plt_start@+0xa0>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0.50	N/A
ADD $0x28,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
RET	1	0.50	0	0	0	0.50	0	0	0	0	0	1	0.50	N/A
ADD %R9,%R8	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
JMP 20403b <main.omp_outlined.33+0xcb>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	1	N/A
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A

Name	Coverage (%)	Time (s)
▼main.omp_outlined.33–	29.27	7.49
○Loop 16 - stream.c:356-356 - exec	29.27	7.29
○Loop 17 - stream.c:355-356 - exec	0.00	0.00
○Loop 18 - stream.c:356-356 - 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