OV - exec - Loop 150

	Loop Id: 150	Module: exec	Source: advec_cell.cpp:147-151 [...]	Coverage: 0.97%

0x41f1b0 MOVQ	%XMM7,%RAX

0x41f1b5 XOR	%EDX,%EDX

0x41f1b7 DIVQ	-0x78(%RBP)    [9]

0x41f1bb PSHUFD	$-0x12,%XMM7,%XMM1

0x41f1c0 MOVQ	%XMM1,%RDX

0x41f1c5 MOVQ	%RAX,%XMM1

0x41f1ca MOV	%RDX,%RAX

0x41f1cd XOR	%EDX,%EDX

0x41f1cf DIVQ	-0x70(%RBP)    [9]

0x41f1d3 MOVQ	%RAX,%XMM3

0x41f1d8 PUNPCKLQDQ	%XMM3,%XMM1

0x41f1dc PSHUFD	$-0x18,%XMM1,%XMM15

0x41f1e2 PADDD	-0xc0(%RBP),%XMM15    [9]

0x41f1eb PMULUDQ	-0xd0(%RBP),%XMM1    [9]

0x41f1f3 MOVDQA	%XMM7,%XMM3

0x41f1f7 PSUBQ	%XMM1,%XMM3

0x41f1fb PADDQ	-0xb0(%RBP),%XMM3    [9]

0x41f203 PSLLQ	$0x20,%XMM3

0x41f208 PSHUFD	$-0x13,%XMM3,%XMM11

0x41f20e PXOR	%XMM8,%XMM8

0x41f213 MOVDQA	%XMM11,%XMM3

0x41f218 PCMPGTD	%XMM15,%XMM8

0x41f21d MOVDQA	%XMM15,%XMM1

0x41f222 PUNPCKLDQ	%XMM8,%XMM1

0x41f227 PSRAD	$0x1f,%XMM3

0x41f22c MOVDQA	%XMM9,%XMM4

0x41f231 PMULUDQ	%XMM1,%XMM4

0x41f235 MOVDQA	%XMM10,%XMM2

0x41f23a PUNPCKLDQ	%XMM3,%XMM11

0x41f23f PMULUDQ	%XMM1,%XMM2

0x41f243 PUNPCKLDQ	0x7b554(%RIP),%XMM8    [11]

0x41f24c MOVDQA	%XMM9,%XMM3

0x41f251 PMULUDQ	%XMM8,%XMM3

0x41f256 PADDQ	%XMM2,%XMM3

0x41f25a PSLLQ	$0x20,%XMM3

0x41f25f PADDQ	%XMM4,%XMM3

0x41f263 PADDQ	%XMM11,%XMM3

0x41f268 PSUBD	0x7c34f(%RIP),%XMM15    [11]

0x41f271 PXOR	%XMM2,%XMM2

0x41f275 PCMPGTD	%XMM15,%XMM2

0x41f27a PSLLQ	$0x3,%XMM3

0x41f27f PUNPCKLDQ	%XMM2,%XMM15

0x41f284 MOVDQA	%XMM12,%XMM4

0x41f289 PMULUDQ	%XMM15,%XMM4

0x41f28e PADDQ	-0xa0(%RBP),%XMM3    [9]

0x41f296 PMULUDQ	%XMM0,%XMM15

0x41f29b PSHUFD	$-0x2c,%XMM2,%XMM2

0x41f2a0 PMULUDQ	%XMM12,%XMM2

0x41f2a5 MOVQ	%XMM3,%RAX

0x41f2aa PADDQ	%XMM15,%XMM2

0x41f2af PSLLQ	$0x20,%XMM2

0x41f2b4 PADDQ	%XMM11,%XMM4

0x41f2b9 PSHUFD	$-0x12,%XMM3,%XMM3

0x41f2be PADDQ	%XMM2,%XMM4

0x41f2c2 PSLLQ	$0x3,%XMM4

0x41f2c7 PADDQ	%XMM13,%XMM4

0x41f2cc MOVQ	%XMM3,%RDX

0x41f2d1 MOVQ	%XMM4,%RSI

0x41f2d6 MOVDQA	%XMM0,%XMM2

0x41f2da PMULUDQ	%XMM1,%XMM2

0x41f2de PSHUFD	$-0x12,%XMM4,%XMM3

0x41f2e3 MOVDQA	%XMM12,%XMM4

0x41f2e8 PMULUDQ	%XMM8,%XMM4

0x41f2ed PADDQ	%XMM2,%XMM4

0x41f2f1 MOVSD	(%RAX),%XMM2    [2]

0x41f2f5 MOVDQA	%XMM12,%XMM15

0x41f2fa PMULUDQ	%XMM1,%XMM15

0x41f2ff PSLLQ	$0x20,%XMM4

0x41f304 MOVQ	%XMM3,%R12

0x41f309 PADDQ	%XMM15,%XMM4

0x41f30e PADDQ	%XMM11,%XMM4

0x41f313 MOVHPD	(%RDX),%XMM2    [12]

0x41f317 PSLLQ	$0x3,%XMM4

0x41f31c PADDQ	%XMM13,%XMM4

0x41f321 MOVQ	%XMM4,%R11

0x41f326 MOVSD	(%RSI),%XMM3    [8]

0x41f32a PSHUFD	$-0x12,%XMM4,%XMM4

0x41f32f MOVQ	%XMM4,%RSI

0x41f334 MOVHPD	(%R12),%XMM3    [10]

0x41f33a MOVSD	(%R11),%XMM4    [5]

0x41f33f MOVDQA	%XMM14,%XMM15

0x41f344 PMULUDQ	%XMM1,%XMM15

0x41f349 ADDPD	%XMM2,%XMM3

0x41f34d MOVDQA	%XMM5,%XMM2

0x41f351 PMULUDQ	%XMM8,%XMM2

0x41f356 MOVHPD	(%RSI),%XMM4    [7]

0x41f35a PADDQ	%XMM15,%XMM2

0x41f35f MOVDQA	%XMM5,%XMM15

0x41f364 PMULUDQ	%XMM1,%XMM15

0x41f369 PSLLQ	$0x20,%XMM2

0x41f36e PADDQ	%XMM15,%XMM2

0x41f373 PADDQ	%XMM11,%XMM2

0x41f378 PSLLQ	$0x3,%XMM2

0x41f37d SUBPD	%XMM4,%XMM3

0x41f381 MOVQ	%R15,%XMM4

0x41f386 PSHUFD	$0x44,%XMM4,%XMM4

0x41f38b PADDQ	%XMM2,%XMM4

0x41f38f MOVQ	%XMM4,%RSI

0x41f394 MOVLPD	%XMM3,(%RSI)    [6]

0x41f398 PSHUFD	$-0x12,%XMM4,%XMM2

0x41f39d MOVQ	%XMM2,%RSI

0x41f3a2 MOVHPD	%XMM3,(%RSI)    [4]

0x41f3a6 MOVDQA	%XMM6,%XMM2

0x41f3aa MOVDQA	%XMM6,%XMM3

0x41f3ae PSRLQ	$0x20,%XMM3

0x41f3b3 PMULUDQ	%XMM1,%XMM2

0x41f3b7 PMULUDQ	%XMM1,%XMM3

0x41f3bb PMULUDQ	%XMM6,%XMM8

0x41f3c0 PADDQ	%XMM3,%XMM8

0x41f3c5 PSLLQ	$0x20,%XMM8

0x41f3cb PADDQ	%XMM2,%XMM8

0x41f3d0 PADDQ	%XMM11,%XMM8

0x41f3d5 MOVQ	%R14,%XMM1

0x41f3da PSLLQ	$0x3,%XMM8

0x41f3e0 PSHUFD	$0x44,%XMM1,%XMM1

0x41f3e5 PADDQ	%XMM8,%XMM1

0x41f3ea MOVSD	(%RAX),%XMM2    [2]

0x41f3ee MOVQ	%XMM1,%RAX

0x41f3f3 PSHUFD	$-0x12,%XMM1,%XMM1

0x41f3f8 MOVQ	%XMM1,%RSI

0x41f3fd MOVQ	(%RDX),%XMM1    [12]

0x41f401 MOVSD	%XMM2,(%RAX)    [1]

0x41f405 MOVQ	%XMM1,(%RSI)    [3]

0x41f409 PADDQ	0x7c03f(%RIP),%XMM7    [11]

0x41f411 ADD	$0x2,%RDI

0x41f415 CMP	%R13,%RDI

0x41f418 JB	41f1b0

/home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp: 147 - 151

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

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

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

149:           pre_vol(i, j) = volume(i, j) + vol_flux_y(i + 0, j + 1) - vol_flux_y(i, j);

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

151:         }

/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	3.32
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.01
Bottlenecks	P0,
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.12]
Source	advec_cell.cpp:147-151,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	33.17
CQA cycles if no scalar integer	31.17
CQA cycles if FP arith vectorized	28.67
CQA cycles if fully vectorized	10.00
Front-end cycles	22.33
P0 cycles	33.17
P1 cycles	32.83
P2 cycles	5.67
P3 cycles	5.67
P4 cycles	2.00
P5 cycles	33.00
P6 cycles	1.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	0.00
P11 cycles	5.67
DIV/SQRT cycles	20.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	127.00
Nb uops	134.00
Nb loads	17.00
Nb stores	4.00
Nb stack references	6.00
FLOP/cycle	0.12
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	6.75
Bytes prefetched	0.00
Bytes loaded	192.00
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	74.38
Vectorization ratio load	41.18
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	73.44
Vector-efficiency ratio all	19.73
Vector-efficiency ratio load	16.18
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	20.90

Metric	Value
CQA speedup if no scalar integer	1.06
CQA speedup if FP arith vectorized	1.16
CQA speedup if fully vectorized	3.32
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.01
Bottlenecks	P0,
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.12]
Source	advec_cell.cpp:147-151,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	33.17
CQA cycles if no scalar integer	31.17
CQA cycles if FP arith vectorized	28.67
CQA cycles if fully vectorized	10.00
Front-end cycles	22.33
P0 cycles	33.17
P1 cycles	32.83
P2 cycles	5.67
P3 cycles	5.67
P4 cycles	2.00
P5 cycles	33.00
P6 cycles	1.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	0.00
P11 cycles	5.67
DIV/SQRT cycles	20.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	127.00
Nb uops	134.00
Nb loads	17.00
Nb stores	4.00
Nb stack references	6.00
FLOP/cycle	0.12
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	6.75
Bytes prefetched	0.00
Bytes loaded	192.00
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	74.38
Vectorization ratio load	41.18
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	73.44
Vector-efficiency ratio all	19.73
Vector-efficiency ratio load	16.18
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	20.90

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.12]
Source file and lines	advec_cell.cpp:147-151
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:147-151
/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 33.17 to 31.17 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 33.17 to 10.00 cycles (3.32x speedup).

Details

74% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

41% 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.
73% 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).

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:147
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:147

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:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149

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 192 bytes (24 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	127
nb uops	134
loop length	622
used x86 registers	10
used mmx registers	0
used xmm registers	16
used ymm registers	0
used zmm registers	0
nb stack references	6

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	22.33 cycles
front end	22.33 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	33.17	32.83	5.67	5.67	2.00	33.00	1.00	2.00	2.00	2.00	0.00	5.67
cycles	33.17	32.83	5.67	5.67	2.00	33.00	1.00	2.00	2.00	2.00	0.00	5.67

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	22.33
Dispatch	33.17
DIV/SQRT	20.00
Data deps.	1.00
Overall L1	33.17

Vectorization ratios

INT

all	80%
load	70%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
other	71%

all	16%
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	74%
load	41%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	0%
other	73%

Vector efficiency ratios

INT

all	20%
load	18%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
other	20%

all	14%
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 33.17 cycles. At this rate:

3% of peak load performance is reached (5.79 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.96 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: 41f1b0

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
DIVQ -0x78(%RBP)	5	0	3	0.33	0.33	0	0	0	0	0	0	0	0.33	11-16	10	scal (12.5%)
PSHUFD $-0x12,%XMM7,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM1,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ %RAX,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
MOV %RDX,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIVQ -0x70(%RBP)	5	0	3	0.33	0.33	0	0	0	0	0	0	0	0.33	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%)
PSHUFD $-0x18,%XMM1,%XMM15	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDD -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%)
PADDQ -0xb0(%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%)
PSLLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $-0x13,%XMM3,%XMM11	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PXOR %XMM8,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MOVDQA %XMM11,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PCMPGTD %XMM15,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
MOVDQA %XMM15,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM8,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PSRAD $0x1f,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM9,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM10,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM3,%XMM11	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PUNPCKLDQ 0x7b554(%RIP),%XMM8	1	0	0.50	0.33	0.33	0	0.50	0	0	0	0	0	0.33	1	0.50	vect (12.5%)
MOVDQA %XMM9,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM2,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
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%)
PADDQ %XMM11,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSUBD 0x7c34f(%RIP),%XMM15	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%)
PXOR %XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PCMPGTD %XMM15,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	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%)
PUNPCKLDQ %XMM2,%XMM15	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
MOVDQA %XMM12,%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%)
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%)
PMULUDQ %XMM0,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSHUFD $-0x2c,%XMM2,%XMM2	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PMULUDQ %XMM12,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM11,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSHUFD $-0x12,%XMM3,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM2,%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%)
PADDQ %XMM13,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM3,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	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%)
MOVDQA %XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSHUFD $-0x12,%XMM4,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVDQA %XMM12,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM2,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM2	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVDQA %XMM12,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	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%)
MOVQ %XMM3,%R12	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
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%)
PADDQ %XMM11,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVHPD (%RDX),%XMM2	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	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%)
PADDQ %XMM13,%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,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVSD (%RSI),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	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,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVHPD (%R12),%XMM3	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVSD (%R11),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVDQA %XMM14,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
ADDPD %XMM2,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVDQA %XMM5,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVHPD (%RSI),%XMM4	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM5,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSLLQ $0x20,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM11,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
SUBPD %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 %XMM2,%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,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
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%)
PSHUFD $-0x12,%XMM4,%XMM2	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM2,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVHPD %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%)
MOVDQA %XMM6,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM6,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSRLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM1,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM6,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM3,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM2,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM11,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %R14,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSLLQ $0x3,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $0x44,%XMM1,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM8,%XMM1	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM2	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	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM1,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM1,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ (%RDX),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVSD %XMM2,(%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 %XMM1,(%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 0x7c03f(%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 41f1b0 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.12+0x1c0>	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.12]
Source file and lines	advec_cell.cpp:147-151
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:147-151
/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

74% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

41% 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.
73% 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).

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:147
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:147

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:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149
- /home/eoseret/qaas_runs_GNR/173-814-2343/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:149

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 192 bytes (24 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	127
nb uops	134
loop length	622
used x86 registers	10
used mmx registers	0
used xmm registers	16
used ymm registers	0
used zmm registers	0
nb stack references	6

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	22.33 cycles
front end	22.33 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	33.17	32.83	5.67	5.67	2.00	33.00	1.00	2.00	2.00	2.00	0.00	5.67
cycles	33.17	32.83	5.67	5.67	2.00	33.00	1.00	2.00	2.00	2.00	0.00	5.67

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	22.33
Dispatch	33.17
DIV/SQRT	20.00
Data deps.	1.00
Overall L1	33.17

Vectorization ratios

INT

all	80%
load	70%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
other	71%

all	16%
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	74%
load	41%
store	0%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	0%
other	73%

Vector efficiency ratios

INT

all	20%
load	18%
store	12%
mul	12%
add-sub	25%
fma	NA (no fma vectorizable/vectorized instructions)
other	20%

all	14%
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 33.17 cycles. At this rate:

3% of peak load performance is reached (5.79 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.96 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: 41f1b0

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
DIVQ -0x78(%RBP)	5	0	3	0.33	0.33	0	0	0	0	0	0	0	0.33	11-16	10	scal (12.5%)
PSHUFD $-0x12,%XMM7,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM1,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ %RAX,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
MOV %RDX,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %EDX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
DIVQ -0x70(%RBP)	5	0	3	0.33	0.33	0	0	0	0	0	0	0	0.33	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%)
PSHUFD $-0x18,%XMM1,%XMM15	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDD -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%)
PADDQ -0xb0(%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%)
PSLLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $-0x13,%XMM3,%XMM11	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PXOR %XMM8,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MOVDQA %XMM11,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PCMPGTD %XMM15,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	vect (25.0%)
MOVDQA %XMM15,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM8,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PSRAD $0x1f,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
MOVDQA %XMM9,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVDQA %XMM10,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PUNPCKLDQ %XMM3,%XMM11	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PUNPCKLDQ 0x7b554(%RIP),%XMM8	1	0	0.50	0.33	0.33	0	0.50	0	0	0	0	0	0.33	1	0.50	vect (12.5%)
MOVDQA %XMM9,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM2,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
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%)
PADDQ %XMM11,%XMM3	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSUBD 0x7c34f(%RIP),%XMM15	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%)
PXOR %XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
PCMPGTD %XMM15,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0-1	0.50	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%)
PUNPCKLDQ %XMM2,%XMM15	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (12.5%)
MOVDQA %XMM12,%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%)
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%)
PMULUDQ %XMM0,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSHUFD $-0x2c,%XMM2,%XMM2	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PMULUDQ %XMM12,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVQ %XMM3,%RAX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM11,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSHUFD $-0x12,%XMM3,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM2,%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%)
PADDQ %XMM13,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %XMM3,%RDX	1	1	0	0	0	0	0	0	0	0	0	0	0	1	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%)
MOVDQA %XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSHUFD $-0x12,%XMM4,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVDQA %XMM12,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM4	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM2,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM2	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVDQA %XMM12,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	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%)
MOVQ %XMM3,%R12	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
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%)
PADDQ %XMM11,%XMM4	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVHPD (%RDX),%XMM2	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	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%)
PADDQ %XMM13,%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,%R11	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVSD (%RSI),%XMM3	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	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,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVHPD (%R12),%XMM3	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
MOVSD (%R11),%XMM4	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVDQA %XMM14,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
ADDPD %XMM2,%XMM3	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	2-3	0.50	vect (25.0%)
MOVDQA %XMM5,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM8,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
MOVHPD (%RSI),%XMM4	1	0	0	0.33	0.33	0	1	0	0	0	0	0	0.33	4-12	1	scal (12.5%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVDQA %XMM5,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PMULUDQ %XMM1,%XMM15	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PSLLQ $0x20,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM15,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM11,%XMM2	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x3,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
SUBPD %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 %XMM2,%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,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
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%)
PSHUFD $-0x12,%XMM4,%XMM2	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM2,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVHPD %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%)
MOVDQA %XMM6,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
MOVDQA %XMM6,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0-1	0.17	vect (25.0%)
PSRLQ $0x20,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PMULUDQ %XMM1,%XMM2	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM1,%XMM3	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PMULUDQ %XMM6,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	5	0.50	vect (12.5%)
PADDQ %XMM3,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PSLLQ $0x20,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PADDQ %XMM2,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
PADDQ %XMM11,%XMM8	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVQ %R14,%XMM1	1	0	0	0	0	0	1	0	0	0	0	0	0	1	1	scal (12.5%)
PSLLQ $0x3,%XMM8	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1-2	0.50	vect (25.0%)
PSHUFD $0x44,%XMM1,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
PADDQ %XMM8,%XMM1	1	0.33	0.33	0	0	0	0.33	0	0	0	0	0	0	1	0.33	vect (25.0%)
MOVSD (%RAX),%XMM2	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	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
PSHUFD $-0x12,%XMM1,%XMM1	1	0	0.50	0	0	0	0.50	0	0	0	0	0	0	1	0.50	vect (25.0%)
MOVQ %XMM1,%RSI	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	scal (12.5%)
MOVQ (%RDX),%XMM1	1	0	0	0.33	0.33	0	0	0	0	0	0	0	0.33	1	0.33	scal (12.5%)
MOVSD %XMM2,(%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 %XMM1,(%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 0x7c03f(%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 41f1b0 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.12+0x1c0>	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