OV - exec - Loop 158

	Loop Id: 158	Module: exec	Source: advec_cell.cpp:55-59 [...]	Coverage: 0.94%

0x421240 MOVQ	%XMM7,%RAX

0x421245 XOR	%EDX,%EDX

0x421247 DIV	%R9

0x42124a MOVQ	%RAX,%XMM1

0x42124f PSHUFD	$-0x12,%XMM7,%XMM3

0x421254 MOVQ	%XMM3,%RAX

0x421259 XOR	%EDX,%EDX

0x42125b DIV	%R12

0x42125e MOVQ	%RAX,%XMM3

0x421263 PUNPCKLQDQ	%XMM3,%XMM1

0x421267 MOVDQA	%XMM1,%XMM15

0x42126c PADDQ	-0xc0(%RBP),%XMM15    [8]

0x421275 PMULUDQ	-0xd0(%RBP),%XMM1    [8]

0x42127d MOVDQA	%XMM7,%XMM3

0x421281 PSUBQ	%XMM1,%XMM3

0x421285 PSHUFD	$-0x18,%XMM3,%XMM14

0x42128b PADDD	-0xb0(%RBP),%XMM14    [8]

0x421294 PXOR	%XMM4,%XMM4

0x421298 PSHUFD	$-0x18,%XMM15,%XMM10

0x42129e PSLLQ	$0x20,%XMM15

0x4212a4 PCMPGTD	%XMM14,%XMM4

0x4212a9 PSRAD	$0x1f,%XMM15

0x4212af PSHUFD	$-0x13,%XMM15,%XMM1

0x4212b5 PUNPCKLDQ	%XMM1,%XMM10

0x4212ba MOVDQA	%XMM14,%XMM1

0x4212bf MOVDQA	%XMM8,%XMM3

0x4212c4 PMULUDQ	%XMM10,%XMM3

0x4212c9 MOVDQA	%XMM9,%XMM0

0x4212ce PUNPCKLDQ	%XMM4,%XMM1

0x4212d2 PSRLQ	$0x20,%XMM15

0x4212d8 MOVDQA	%XMM8,%XMM4

0x4212dd PMULUDQ	%XMM15,%XMM4

0x4212e2 PMULUDQ	%XMM10,%XMM0

0x4212e7 PADDQ	%XMM0,%XMM4

0x4212eb PADDQ	%XMM1,%XMM3

0x4212ef PSUBD	0x7a2c8(%RIP),%XMM14    [9]

0x4212f8 PSLLQ	$0x20,%XMM4

0x4212fd PXOR	%XMM0,%XMM0

0x421301 PCMPGTD	%XMM14,%XMM0

0x421306 PUNPCKLDQ	%XMM0,%XMM14

0x42130b PADDQ	%XMM4,%XMM3

0x42130f MOVDQA	%XMM2,%XMM0

0x421313 MOVDQA	%XMM11,%XMM4

0x421318 PMULUDQ	%XMM15,%XMM4

0x42131d PMULUDQ	%XMM10,%XMM0

0x421322 PADDQ	%XMM0,%XMM4

0x421326 MOVDQA	%XMM11,%XMM0

0x42132b PMULUDQ	%XMM10,%XMM0

0x421330 PSLLQ	$0x20,%XMM4

0x421335 PADDQ	%XMM0,%XMM4

0x421339 PSLLQ	$0x3,%XMM3

0x42133e PADDQ	-0xa0(%RBP),%XMM3    [8]

0x421346 MOVQ	%XMM3,%RAX

0x42134b PSHUFD	$-0x12,%XMM3,%XMM0

0x421350 PADDQ	%XMM4,%XMM14

0x421355 PSLLQ	$0x3,%XMM14

0x42135b MOVQ	%XMM0,%RDX

0x421360 PADDQ	%XMM12,%XMM14

0x421365 MOVQ	%XMM14,%RBX

0x42136a PSHUFD	$-0x12,%XMM14,%XMM0

0x421370 MOVQ	%XMM0,%R11

0x421375 PADDQ	%XMM1,%XMM4

0x421379 PSLLQ	$0x3,%XMM4

0x42137e MOVDQA	%XMM13,%XMM0

0x421383 PMULUDQ	%XMM10,%XMM0

0x421388 MOVDQA	%XMM5,%XMM14

0x42138d PMULUDQ	%XMM15,%XMM14

0x421392 PADDQ	%XMM0,%XMM14

0x421397 MOVSD	(%RBX),%XMM3    [11]

0x42139b PADDQ	%XMM12,%XMM4

0x4213a0 MOVQ	%XMM4,%RBX

0x4213a5 PSHUFD	$-0x12,%XMM4,%XMM0

0x4213aa MOVQ	%XMM0,%RSI

0x4213af MOVDQA	%XMM5,%XMM0

0x4213b3 PMULUDQ	%XMM10,%XMM0

0x4213b8 PADDQ	%XMM1,%XMM0

0x4213bc PSLLQ	$0x20,%XMM14

0x4213c2 PADDQ	%XMM14,%XMM0

0x4213c7 MOVSD	(%RAX),%XMM4    [2]

0x4213cb MOVHPD	(%R11),%XMM3    [4]

0x4213d0 MOVHPD	(%RDX),%XMM4    [12]

0x4213d4 PSLLQ	$0x3,%XMM0

0x4213d9 ADDPD	%XMM4,%XMM3

0x4213dd MOVQ	%R15,%XMM4

0x4213e2 PSHUFD	$0x44,%XMM4,%XMM4

0x4213e7 PADDQ	%XMM0,%XMM4

0x4213eb MOVSD	(%RBX),%XMM0    [10]

0x4213ef MOVHPD	(%RSI),%XMM0    [7]

0x4213f3 MOVQ	%XMM4,%RSI

0x4213f8 PSHUFD	$-0x12,%XMM4,%XMM4

0x4213fd MOVQ	%XMM4,%R11

0x421402 MOVDQA	%XMM6,%XMM4

0x421406 PMULUDQ	%XMM10,%XMM4

0x42140b SUBPD	%XMM0,%XMM3

0x42140f MOVDQA	%XMM6,%XMM0

0x421413 PSRLQ	$0x20,%XMM0

0x421418 PMULUDQ	%XMM10,%XMM0

0x42141d PMULUDQ	%XMM6,%XMM15

0x421422 PADDQ	%XMM0,%XMM15

0x421427 PADDQ	%XMM1,%XMM4

0x42142b MOVLPD	%XMM3,(%RSI)    [6]

0x42142f MOVHPD	%XMM3,(%R11)    [3]

0x421434 PSLLQ	$0x20,%XMM15

0x42143a PADDQ	%XMM15,%XMM4

0x42143f MOVQ	%R14,%XMM0

0x421444 PSLLQ	$0x3,%XMM4

0x421449 PSHUFD	$0x44,%XMM0,%XMM0

0x42144e PADDQ	%XMM4,%XMM0

0x421452 MOVQ	(%RAX),%XMM1    [2]

0x421456 MOVQ	%XMM0,%RAX

0x42145b PSHUFD	$-0x12,%XMM0,%XMM0

0x421460 MOVQ	%XMM0,%RSI

0x421465 MOVQ	(%RDX),%XMM0    [12]

0x421469 MOVQ	%XMM1,(%RAX)    [1]

0x42146d MOVQ	%XMM0,(%RSI)    [5]

0x421471 PADDQ	0x79fd7(%RIP),%XMM7    [9]

0x421479 ADD	$0x2,%RDI

0x42147d CMP	%R13,%RDI

0x421480 JB	421240

/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp: 55 - 59

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

55:       for (int j = (y_min - 2 + 1); j < (y_max + 2 + 2); j++) {

56:         for (int i = (x_min - 2 + 1); i < (x_max + 2 + 2); i++) {

57:           pre_vol(i, j) = volume(i, j) + vol_flux_x(i + 1, j + 0) - vol_flux_x(i, j);

58:           post_vol(i, j) = volume(i, j);

59:         }

/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h: 69 - 69

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

69:   T &operator()(size_t i, size_t j) const { return data[i + j * sizeX]; }

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	1.06
CQA speedup if FP arith vectorized	1.16
CQA speedup if fully vectorized	3.10
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.48
Bottlenecks	P0, P1, P5,
Function	advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone .extracted.32]
Source	advec_cell.cpp:55-59,context.h:69-69
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	high
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	31.00
CQA cycles if no scalar integer	29.17
CQA cycles if FP arith vectorized	26.67
CQA cycles if fully vectorized	10.00
Front-end cycles	21.00
P0 cycles	31.00
P1 cycles	31.00
P2 cycles	4.67
P3 cycles	4.67
P4 cycles	2.00
P5 cycles	31.00
P6 cycles	1.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	0.00
P11 cycles	4.67
DIV/SQRT cycles	20.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	119.00
Nb uops	126.00
Nb loads	14.00
Nb stores	4.00
Nb stack references	4.00
FLOP/cycle	0.13
Nb FLOP add-sub	4.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	6.19
Bytes prefetched	0.00
Bytes loaded	160.00
Bytes stored	32.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	10.00
Stride indirect	0.00
Vectorization ratio all	72.81
Vectorization ratio load	42.86
Vectorization ratio store	0.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	72.58
Vector-efficiency ratio all	19.85
Vector-efficiency ratio load	16.96
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	12.50
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	12.50
Vector-efficiency ratio other	20.97

Metric	Value
CQA speedup if no scalar integer	1.06
CQA speedup if FP arith vectorized	1.16
CQA speedup if fully vectorized	3.10
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.48
Bottlenecks	P0, P1, P5,
Function	advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone .extracted.32]
Source	advec_cell.cpp:55-59,context.h:69-69
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	high
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	31.00
CQA cycles if no scalar integer	29.17
CQA cycles if FP arith vectorized	26.67
CQA cycles if fully vectorized	10.00
Front-end cycles	21.00
P0 cycles	31.00
P1 cycles	31.00
P2 cycles	4.67
P3 cycles	4.67
P4 cycles	2.00
P5 cycles	31.00
P6 cycles	1.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	0.00
P11 cycles	4.67
DIV/SQRT cycles	20.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	119.00
Nb uops	126.00
Nb loads	14.00
Nb stores	4.00
Nb stack references	4.00
FLOP/cycle	0.13
Nb FLOP add-sub	4.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	6.19
Bytes prefetched	0.00
Bytes loaded	160.00
Bytes stored	32.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	10.00
Stride indirect	0.00
Vectorization ratio all	72.81
Vectorization ratio load	42.86
Vectorization ratio store	0.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	72.58
Vector-efficiency ratio all	19.85
Vector-efficiency ratio load	16.96
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	12.50
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	12.50
Vector-efficiency ratio other	20.97

Path /

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

Function	advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone .extracted.32]
Source file and lines	advec_cell.cpp:55-59
Module	exec

The loop is defined in:

/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55-59
/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

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

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 31.00 to 29.17 cycles (1.06x speedup).

Workaround

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

Vectorization

Your loop is partially vectorized. Only 19% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 31.00 to 10.00 cycles (3.10x speedup).

Details

72% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

42% of SSE/AVX loads are used in vector version.
0% of SSE/AVX stores are used in vector version.
0% of SSE/AVX divide and square root instructions are used in vector version.
72% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully 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

Performance is limited by execution of FP multiply or FMA (fused multiply-add) operations (the FP multiply/FMA unit is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 31.00 to 21.00 cycles (1.48x speedup).

Workaround

Reduce the number of FP multiply/FMA instructions

No data for this section

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.

DIV: 2 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant unknown stride: 10 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

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

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVHPD: 4 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

2 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 4 FP arithmetical operations:

4: addition or subtraction

The binary loop is loading 160 bytes (20 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.02 FP operations per loaded or stored byte.

General properties

nb instructions	119
nb uops	126
loop length	582
used x86 registers	12
used mmx registers	0
used xmm registers	16
used ymm registers	0
used zmm registers	0
nb stack references	4

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	21.00 cycles
front end	21.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	31.00	31.00	4.67	4.67	2.00	31.00	1.00	2.00	2.00	2.00	0.00	4.67
cycles	31.00	31.00	4.67	4.67	2.00	31.00	1.00	2.00	2.00	2.00	0.00	4.67

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	20.00
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	21.00
Dispatch	31.00
DIV/SQRT	20.00
Data deps.	1.00
Overall L1	31.00

Vectorization ratios

INT

all	77%
load	75%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
other	70%

all	20%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	72%
load	42%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	0%
other	72%

Vector efficiency ratios

INT

all	20%
load	20%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
other	20%

all	15%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	19%
load	16%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	12%
other	20%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 31.00 cycles. At this rate:

2% of peak load performance is reached (5.16 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (1.03 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: 421240

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput	Vectorization
MOVQ %XMM7,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIV %R9	5	0	3	0	0	0	0	0	0	0	0	0	0	11-16	10	scal (12.5%)
MOVQ %RAX,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM7,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIV %R12	5	0	3	0	0	0	0	0	0	0	0	0	0	11-16	10	scal (12.5%)
MOVQ %RAX,%XMM3	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PUNPCKLQDQ %XMM3,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOVDQA %XMM1,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PADDQ -0xc0(%RBP),%XMM15	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
PMULUDQ -0xd0(%RBP),%XMM1	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	5	0.50	vect (12.5%)
MOVDQA %XMM7,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSUBQ %XMM1,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	0-1	0.33	vect (25.0%)
PSHUFD $-0x18,%XMM3,%XMM14	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDD -0xb0(%RBP),%XMM14	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
PXOR %XMM4,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PSHUFD $-0x18,%XMM15,%XMM10	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PSLLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PCMPGTD %XMM14,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
PSRAD $0x1f,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $-0x13,%XMM15,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PUNPCKLDQ %XMM1,%XMM10	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
MOVDQA %XMM14,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM8,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM9,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM4,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PSRLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM8,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM1,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSUBD 0x7a2c8(%RIP),%XMM14	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	0-1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PXOR %XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PCMPGTD %XMM14,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
PUNPCKLDQ %XMM0,%XMM14	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PADDQ %XMM4,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM11,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSLLQ $0x20,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ -0xa0(%RBP),%XMM3	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM3,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM4,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVQ %XMM0,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM12,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM14,%RBX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM14,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM1,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM13,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM5,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RBX),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
PADDQ %XMM12,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM4,%RBX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM4,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVDQA %XMM5,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM1,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM14,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVHPD (%R11),%XMM3	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVHPD (%RDX),%XMM4	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
PSLLQ $0x3,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
ADDPD %XMM4,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVQ %R15,%XMM4	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $0x44,%XMM4,%XMM4	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RBX),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVHPD (%RSI),%XMM0	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVQ %XMM4,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM4,%XMM4	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM4,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVDQA %XMM6,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
SUBPD %XMM0,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVDQA %XMM6,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSRLQ $0x20,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM6,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM15	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM1,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVLPD %XMM3,(%RSI)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	4-12	0.50	scal (12.5%)
MOVHPD %XMM3,(%R11)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	4-12	0.50	scal (12.5%)
PSLLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM15,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %R14,%XMM0	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSLLQ $0x3,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $0x44,%XMM0,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM4,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ (%RAX),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVQ %XMM0,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM0,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ (%RDX),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVQ %XMM1,(%RAX)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (12.5%)
MOVQ %XMM0,(%RSI)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (12.5%)
PADDQ 0x79fd7(%RIP),%XMM7	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
ADD $0x2,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %R13,%RDI	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	N/A
JB 421240 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.32+0x1d0>	1	0.50	0	0	0	0	0	0.50	0	0	0	0	0	0	0.50	N/A

Function	advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone .extracted.32]
Source file and lines	advec_cell.cpp:55-59
Module	exec

The loop is defined in:

/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55-59
/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

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

gain
potential
hint
expert

Code clean check

Workaround

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

Vectorization

Details

72% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

42% of SSE/AVX loads are used in vector version.
0% of SSE/AVX stores are used in vector version.
0% of SSE/AVX divide and square root instructions are used in vector version.
72% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully 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

Workaround

Reduce the number of FP multiply/FMA instructions

No data for this section

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.

DIV: 2 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:55

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant unknown stride: 10 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

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

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVHPD: 4 occurrences►
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:57

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

2 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 4 FP arithmetical operations:

4: addition or subtraction

The binary loop is loading 160 bytes (20 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.02 FP operations per loaded or stored byte.

General properties

nb instructions	119
nb uops	126
loop length	582
used x86 registers	12
used mmx registers	0
used xmm registers	16
used ymm registers	0
used zmm registers	0
nb stack references	4

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	21.00 cycles
front end	21.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	31.00	31.00	4.67	4.67	2.00	31.00	1.00	2.00	2.00	2.00	0.00	4.67
cycles	31.00	31.00	4.67	4.67	2.00	31.00	1.00	2.00	2.00	2.00	0.00	4.67

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	20.00
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	21.00
Dispatch	31.00
DIV/SQRT	20.00
Data deps.	1.00
Overall L1	31.00

Vectorization ratios

INT

all	77%
load	75%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
other	70%

all	20%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	72%
load	42%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	0%
other	72%

Vector efficiency ratios

INT

all	20%
load	20%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
other	20%

all	15%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	19%
load	16%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	12%
other	20%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 31.00 cycles. At this rate:

2% of peak load performance is reached (5.16 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (1.03 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: 421240

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput	Vectorization
MOVQ %XMM7,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIV %R9	5	0	3	0	0	0	0	0	0	0	0	0	0	11-16	10	scal (12.5%)
MOVQ %RAX,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM7,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIV %R12	5	0	3	0	0	0	0	0	0	0	0	0	0	11-16	10	scal (12.5%)
MOVQ %RAX,%XMM3	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PUNPCKLQDQ %XMM3,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOVDQA %XMM1,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PADDQ -0xc0(%RBP),%XMM15	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
PMULUDQ -0xd0(%RBP),%XMM1	1	0.50	0.50	0.33	0.33	0	0	0	0	0	0	0	0.33	5	0.50	vect (12.5%)
MOVDQA %XMM7,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSUBQ %XMM1,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	0-1	0.33	vect (25.0%)
PSHUFD $-0x18,%XMM3,%XMM14	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDD -0xb0(%RBP),%XMM14	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
PXOR %XMM4,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PSHUFD $-0x18,%XMM15,%XMM10	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PSLLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PCMPGTD %XMM14,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
PSRAD $0x1f,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $-0x13,%XMM15,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PUNPCKLDQ %XMM1,%XMM10	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
MOVDQA %XMM14,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM8,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM9,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM4,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PSRLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM8,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM1,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSUBD 0x7a2c8(%RIP),%XMM14	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	0-1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PXOR %XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PCMPGTD %XMM14,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
PUNPCKLDQ %XMM0,%XMM14	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PADDQ %XMM4,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM11,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSLLQ $0x20,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ -0xa0(%RBP),%XMM3	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM3,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM4,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVQ %XMM0,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM12,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM14,%RBX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM14,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM1,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM13,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM5,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM15,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM14	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RBX),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
PADDQ %XMM12,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM4,%RBX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM4,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVDQA %XMM5,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM1,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM14	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM14,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVHPD (%R11),%XMM3	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVHPD (%RDX),%XMM4	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
PSLLQ $0x3,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
ADDPD %XMM4,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVQ %R15,%XMM4	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $0x44,%XMM4,%XMM4	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM0,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RBX),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVHPD (%RSI),%XMM0	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVQ %XMM4,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM4,%XMM4	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM4,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVDQA %XMM6,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM10,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
SUBPD %XMM0,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVDQA %XMM6,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSRLQ $0x20,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PMULUDQ %XMM10,%XMM0	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM6,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM0,%XMM15	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM1,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVLPD %XMM3,(%RSI)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	4-12	0.50	scal (12.5%)
MOVHPD %XMM3,(%R11)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	4-12	0.50	scal (12.5%)
PSLLQ $0x20,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM15,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %R14,%XMM0	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSLLQ $0x3,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $0x44,%XMM0,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM4,%XMM0	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ (%RAX),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVQ %XMM0,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM0,%XMM0	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM0,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ (%RDX),%XMM0	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVQ %XMM1,(%RAX)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (12.5%)
MOVQ %XMM0,(%RSI)	1	0	0	0	0	0.50	0	0	0.50	0.50	0.50	0	0	1	0.50	scal (12.5%)
PADDQ 0x79fd7(%RIP),%XMM7	1	0.33	0.33	0.33	0.33	0	0.33	0	0	0	0	0	0.33	1	0.33	vect (25.0%)
ADD $0x2,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %R13,%RDI	1	0.20	0.20	0	0	0	0.20	0.20	0	0	0	0.20	0	1	0.20	N/A
JB 421240 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.32+0x1d0>	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

Workaround

Complex instructions

Details

Slow data structures access

Details

Workaround

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

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Complex instructions

Details

Slow data structures access

Details

Workaround

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