OV - exec - Loop 75

	Loop Id: 75	Module: exec	Source: advec_cell_kernel.f90:112-157 [...]	Coverage: 3.47%

0x226290 VADDSD	%XMM12,%XMM13,%XMM8

0x226295 INC	%RDI

0x226298 VMULSD	%XMM7,%XMM8,%XMM7

0x22629c VMOVSD	%XMM7,(%R9,%R11,8)    [5]

0x2262a2 INC	%R11

0x2262a5 CMP	%R11,0x10(%RSP)    [6]

0x2262aa JE	226220

0x2262b0 VMOVSD	(%R10,%R11,8),%XMM7    [2]

0x2262b6 VUCOMISD	%XMM7,%XMM0

0x2262ba JB	2262f0

0x2262bc MOV	0x38(%RSP),%RCX    [6]

0x2262c1 MOV	0x50(%RSP),%RAX    [6]

0x2262c6 MOV	0x28(%RSP),%RBX    [6]

0x2262cb MOV	%RDI,%R13

0x2262ce LEA	(%RCX,%RDI,1),%R15

0x2262d2 MOV	0x48(%RSP),%RCX    [6]

0x2262d7 LEA	0x1(%RAX,%R11,1),%EAX

0x2262dc CMP	%ECX,%EAX

0x2262de CMOVGE	%ECX,%EAX

0x2262e1 LEA	(%RBX,%R11,1),%RCX

0x2262e5 MOV	%EAX,%EDX

0x2262e7 JMP	226319

0x2262f0 MOV	0x28(%RSP),%RAX    [6]

0x2262f5 MOV	0x18(%RSP),%RCX    [6]

0x2262fa LEA	(%RAX,%R11,1),%R13

0x2262fe MOV	0x58(%RSP),%RAX    [6]

0x226303 LEA	(%RCX,%R11,1),%R15

0x226307 MOV	0xf0(%RSP),%RCX    [6]

0x22630f LEA	(%RCX,%R11,1),%EDX

0x226313 ADD	%R11D,%EAX

0x226316 MOV	%RDI,%RCX

0x226319 MOV	(%RSP),%RBX    [6]

0x22631d VANDPD	%XMM1,%XMM7,%XMM9

0x226321 MOVSXD	%EDX,%RDX

0x226324 SUB	0x20(%RSP),%RDX    [6]

0x226329 ADD	%R8,%R13

0x22632c CLTQ

0x22632e ADD	%R8,%RCX

0x226331 ADD	%R8,%RAX

0x226334 SUB	%R12,%R13

0x226337 SUB	%R12,%RCX

0x22633a SUB	%R12,%RAX

0x22633d VMOVSD	(%RBP,%R13,8),%XMM11    [3]

0x226344 VMOVSD	(%RBP,%RCX,8),%XMM13    [3]

0x22634a VSUBSD	(%RBP,%RAX,8),%XMM11,%XMM15    [3]

0x226350 VMOVSD	(%RBX,%R15,8),%XMM8    [9]

0x226356 MOV	0x30(%RSP),%RBX    [6]

0x22635b VSUBSD	%XMM11,%XMM13,%XMM14

0x226360 VXORPD	%XMM13,%XMM13,%XMM13

0x226365 VDIVSD	%XMM8,%XMM9,%XMM12

0x22636a VMOVSD	(%RBX,%R11,8),%XMM9    [8]

0x226370 MOV	0x8(%RSP),%RBX    [6]

0x226375 VMULSD	%XMM15,%XMM14,%XMM16

0x22637b VUCOMISD	%XMM13,%XMM16

0x226381 VXORPD	%XMM16,%XMM16,%XMM16

0x226387 VFMADD213SD	%XMM9,%XMM12,%XMM9

0x22638c VDIVSD	(%RBX,%RDX,8),%XMM9,%XMM10    [7]

0x226391 VSUBSD	%XMM12,%XMM2,%XMM9

0x226396 JBE	2263de

0x226398 VANDPD	%XMM3,%XMM15,%XMM15

0x22639c VANDPD	%XMM3,%XMM14,%XMM16

0x2263a2 VSUBSD	%XMM12,%XMM5,%XMM12

0x2263a7 VCMPSD	$0x1,%XMM14,%XMM0,%K1

0x2263ae VMULSD	%XMM10,%XMM15,%XMM17

0x2263b4 VMINSD	%XMM15,%XMM16,%XMM15

0x2263ba VFMADD231SD	%XMM9,%XMM16,%XMM17

0x2263c0 VXORPD	%XMM6,%XMM12,%XMM16

0x2263c6 VMOVSD	%XMM12,%XMM16,%XMM16{%K1}

0x2263cc VMULSD	%XMM4,%XMM17,%XMM17

0x2263d2 VMINSD	%XMM15,%XMM17,%XMM15

0x2263d8 VMULSD	%XMM16,%XMM15,%XMM16

0x2263de VADDSD	%XMM11,%XMM16,%XMM12

0x2263e4 VMULSD	%XMM7,%XMM12,%XMM7

0x2263e8 VMOVSD	%XMM7,(%R14,%R11,8)    [4]

0x2263ee VMOVSD	(%RSI,%R13,8),%XMM12    [1]

0x2263f4 VMOVSD	(%RSI,%RCX,8),%XMM14    [1]

0x2263f9 VSUBSD	(%RSI,%RAX,8),%XMM12,%XMM15    [1]

0x2263fe VSUBSD	%XMM12,%XMM14,%XMM14

0x226403 VMULSD	%XMM15,%XMM14,%XMM16

0x226409 VUCOMISD	%XMM13,%XMM16

0x22640f JBE	226290

0x226415 VMULSD	%XMM8,%XMM11,%XMM8

0x22641a VANDPD	%XMM3,%XMM7,%XMM13

0x22641e VANDPD	%XMM3,%XMM14,%XMM11

0x226422 VCMPSD	$0x1,%XMM14,%XMM0,%K1

0x226429 VDIVSD	%XMM8,%XMM13,%XMM8

0x22642e VANDPD	%XMM3,%XMM15,%XMM13

0x226432 VMULSD	%XMM10,%XMM13,%XMM10

0x226437 VFMADD213SD	%XMM10,%XMM11,%XMM9

0x22643c VMINSD	%XMM13,%XMM11,%XMM10

0x226441 VMULSD	%XMM4,%XMM9,%XMM9

0x226445 VMINSD	%XMM10,%XMM9,%XMM9

0x22644a VSUBSD	%XMM8,%XMM5,%XMM8

0x22644f VXORPD	%XMM6,%XMM8,%XMM10

0x226453 VMOVSD	%XMM8,%XMM10,%XMM10{%K1}

0x226459 VMULSD	%XMM10,%XMM9,%XMM13

0x22645e JMP	226290

/home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90: 112 - 157

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

112:           IF(vol_flux_x(j,k).GT.0.0)THEN

113:             upwind   =j-2

114:             donor    =j-1

115:             downwind =j

116:             dif      =donor

117:           ELSE

118:             upwind   =MIN(j+1,x_max+2)

[...]

124:           sigmat=ABS(vol_flux_x(j,k))/pre_vol(donor,k)

125:           sigma3=(1.0_8+sigmat)*(vertexdx(j)/vertexdx(dif))

126:           sigma4=2.0_8-sigmat

127:

128:           sigma=sigmat

129:           sigmav=sigmat

130:

131:           diffuw=density1(donor,k)-density1(upwind,k)

132:           diffdw=density1(downwind,k)-density1(donor,k)

133:           wind=1.0_8

134:           IF(diffdw.LE.0.0) wind=-1.0_8

135:           IF(diffuw*diffdw.GT.0.0)THEN

136:             limiter=(1.0_8-sigmav)*wind*MIN(ABS(diffuw),ABS(diffdw)&

137:               ,one_by_six*(sigma3*ABS(diffuw)+sigma4*ABS(diffdw)))

138:           ELSE

139:             limiter=0.0

140:           ENDIF

141:           mass_flux_x(j,k)=vol_flux_x(j,k)*(density1(donor,k)+limiter)

142:

143:           sigmam=ABS(mass_flux_x(j,k))/(density1(donor,k)*pre_vol(donor,k))

144:           diffuw=energy1(donor,k)-energy1(upwind,k)

145:           diffdw=energy1(downwind,k)-energy1(donor,k)

146:           wind=1.0_8

147:           IF(diffdw.LE.0.0) wind=-1.0_8

148:           IF(diffuw*diffdw.GT.0.0)THEN

149:             limiter=(1.0_8-sigmam)*wind*MIN(ABS(diffuw),ABS(diffdw)&

[...]

155:           ener_flux(j,k)=mass_flux_x(j,k)*(energy1(donor,k)+limiter)

156:

157:         ENDDO

Coverage (%)	Name	Source Location	Module

min	med	avg	max

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

min	med	avg	max

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

Path /

Metric	Value
CQA speedup if no scalar integer	1.05
CQA speedup if FP arith vectorized	3.05
CQA speedup if fully vectorized	8.42
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	NA
Bottlenecks	micro-operation queue,
Function	__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1
Source	advec_cell_kernel.f90:112-118,advec_cell_kernel.f90:124-149,advec_cell_kernel.f90:155-157
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	12.63
CQA cycles if no scalar integer	12.00
CQA cycles if FP arith vectorized	4.14
CQA cycles if fully vectorized	1.50
Front-end cycles	12.63
P0 cycles	4.50
P1 cycles	4.50
P2 cycles	4.50
P3 cycles	4.50
P4 cycles	4.50
P5 cycles	4.50
P6 cycles	6.50
P7 cycles	6.50
P8 cycles	6.50
P9 cycles	6.50
P10 cycles	11.25
P11 cycles	11.25
P12 cycles	11.25
P13 cycles	11.25
P14 cycles	2.50
P15 cycles	2.50
DIV/SQRT cycles	12.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	97.00
Nb uops	101.00
Nb loads	23.00
Nb stores	2.00
Nb stack references	12.00
FLOP/cycle	2.30
Nb FLOP add-sub	9.00
Nb FLOP mul	11.00
Nb FLOP fma	3.00
Nb FLOP div	3.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	15.84
Bytes prefetched	0.00
Bytes loaded	184.00
Bytes stored	16.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	16.95
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	0.00
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	41.67
Vector-efficiency ratio all	14.51
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	12.50
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	12.50
Vector-efficiency ratio div_sqrt	12.50
Vector-efficiency ratio other	17.45

Metric	Value
CQA speedup if no scalar integer	1.05
CQA speedup if FP arith vectorized	3.05
CQA speedup if fully vectorized	8.42
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	NA
Bottlenecks	micro-operation queue,
Function	__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1
Source	advec_cell_kernel.f90:112-118,advec_cell_kernel.f90:124-149,advec_cell_kernel.f90:155-157
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	12.63
CQA cycles if no scalar integer	12.00
CQA cycles if FP arith vectorized	4.14
CQA cycles if fully vectorized	1.50
Front-end cycles	12.63
P0 cycles	4.50
P1 cycles	4.50
P2 cycles	4.50
P3 cycles	4.50
P4 cycles	4.50
P5 cycles	4.50
P6 cycles	6.50
P7 cycles	6.50
P8 cycles	6.50
P9 cycles	6.50
P10 cycles	11.25
P11 cycles	11.25
P12 cycles	11.25
P13 cycles	11.25
P14 cycles	2.50
P15 cycles	2.50
DIV/SQRT cycles	12.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	97.00
Nb uops	101.00
Nb loads	23.00
Nb stores	2.00
Nb stack references	12.00
FLOP/cycle	2.30
Nb FLOP add-sub	9.00
Nb FLOP mul	11.00
Nb FLOP fma	3.00
Nb FLOP div	3.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	15.84
Bytes prefetched	0.00
Bytes loaded	184.00
Bytes stored	16.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	16.95
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	0.00
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	41.67
Vector-efficiency ratio all	14.51
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	12.50
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	12.50
Vector-efficiency ratio div_sqrt	12.50
Vector-efficiency ratio other	17.45

Path /

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

Function	__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1
Source file and lines	advec_cell_kernel.f90:112-157
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:112-118,124-149,155-157.

The related source loop is not unrolled or unrolled with no peel/tail loop.
Warnings:

Ignoring paths for analysis
Too many paths. Rerun with max-paths=8
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 8 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = max(0,x) (Fortran instrinsic procedure)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 12.63 to 12.00 cycles (1.05x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)
To reference allocatable arrays, use "allocatable" instead of "pointer" pointers or qualify them with the "contiguous" attribute (Fortran 2008)
For structures, limit to one indirection. For example, use a_b%c instead of a%b%c with a_b set to a%b before this loop

Vectorization

Your loop is probably not vectorized. Only 14% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your loop, you can lower the cost of an iteration from 12.63 to 1.50 cycles (8.42x 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 loop can use their full power.

Workaround

Try another compiler or update/tune your current one
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: Fortran storage order is column-major: do i do j a(i,j) = b(i,j) (slow, non stride 1) => do i do j a(j,i) = b(i,j) (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): do i a(i)%x = b(i)%x (slow, non stride 1) => do 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.

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

FMA

Detected 3 FMA (fused multiply-add) operations. Presence of both ADD/SUB and MUL operations.

Workaround

Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

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.

CMOVGE: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:118
LEA: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:118
VUCOMISD: 3 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:112
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:135
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:148

Conversion instructions

Detected expensive conversion instructions.

Details

CLTQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:131

Workaround

Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.

Type of elements and instruction set

35 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 29 FP arithmetical operations:

12: addition or subtraction (3 inside FMA instructions)
14: multiply (3 inside FMA instructions)
3: divide

The binary loop is loading 184 bytes (23 double precision FP elements). The binary loop is storing 16 bytes (2 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.14 FP operations per loaded or stored byte.

Unroll opportunity

Loop is potentially data access bound.

Workaround

Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually.

General properties

nb instructions	97
nb uops	101
loop length	460
used x86 registers	16
used mmx registers	0
used xmm registers	18
used ymm registers	0
used zmm registers	0
nb stack references	12
ADD-SUB / MUL ratio	0.82

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	12.63 cycles
front end	12.63 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	4.50	4.50	4.50	4.50	4.50	4.50	6.50	6.50	6.50	6.50	11.25	11.25	11.25	11.25	2.50	2.50
cycles	4.50	4.50	4.50	4.50	4.50	4.50	6.50	6.50	6.50	6.50	11.25	11.25	11.25	11.25	2.50	2.50

Execution ports to units layout:

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

Cycles executing div or sqrt instructions

12.00

Cycles summary

Front-end	12.63
Dispatch	11.25
DIV/SQRT	12.00
Overall L1	12.63

Vectorization ratios

INT

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)
other	0%

all	17%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	47%

INT+FP

all	16%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	41%

Vector efficiency ratios

INT

all	10%
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)
other	10%

all	14%
load	12%
store	12%
mul	12%
add-sub	12%
fma	12%
div/sqrt	12%
other	18%

INT+FP

all	14%
load	12%
store	12%
mul	12%
add-sub	12%
fma	12%
div/sqrt	12%
other	17%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 12.63 cycles. At this rate:

5% of peak load performance is reached (14.57 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (1.27 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck).

ASM code

In the binary file, the address of the loop is: 2262b0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
VADDSD %XMM12,%XMM13,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
INC %RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMULSD %XMM7,%XMM8,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMOVSD %XMM7,(%R9,%R11,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
INC %R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %R11,0x10(%RSP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 226220 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1130>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMOVSD (%R10,%R11,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VUCOMISD %XMM7,%XMM0	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JB 2262f0 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1200>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV 0x38(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x50(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x28(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV %RDI,%R13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%RCX,%RDI,1),%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x48(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA 0x1(%RAX,%R11,1),%EAX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	2	0.50	N/A
CMP %ECX,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (6.3%)
CMOVGE %ECX,%EAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
LEA (%RBX,%R11,1),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV %EAX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 226319 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1229>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A
MOV 0x28(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x18(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RAX,%R11,1),%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x58(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RCX,%R11,1),%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0xf0(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RCX,%R11,1),%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
ADD %R11D,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
MOV %RDI,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VANDPD %XMM1,%XMM7,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
MOVSXD %EDX,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB 0x20(%RSP),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %R8,%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CLTQ																				scal (12.5%)
ADD %R8,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD %R8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVSD (%RBP,%R13,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VMOVSD (%RBP,%RCX,8),%XMM13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VSUBSD (%RBP,%RAX,8),%XMM11,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMOVSD (%RBX,%R15,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VSUBSD %XMM11,%XMM13,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VXORPD %XMM13,%XMM13,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VDIVSD %XMM8,%XMM9,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VMOVSD (%RBX,%R11,8),%XMM9	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMULSD %XMM15,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VUCOMISD %XMM13,%XMM16	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
VXORPD %XMM16,%XMM16,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VFMADD213SD %XMM9,%XMM12,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VDIVSD (%RBX,%RDX,8),%XMM9,%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VSUBSD %XMM12,%XMM2,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
JBE 2263de <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x12ee>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VANDPD %XMM3,%XMM15,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VANDPD %XMM3,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VSUBSD %XMM12,%XMM5,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VCMPSD $0x1,%XMM14,%XMM0,%K1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	scal (12.5%)
VMULSD %XMM10,%XMM15,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM15,%XMM16,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VFMADD231SD %XMM9,%XMM16,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VXORPD %XMM6,%XMM12,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMOVSD %XMM12,%XMM16,%XMM16{%K1}	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMULSD %XMM4,%XMM17,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM15,%XMM17,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM16,%XMM15,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VADDSD %XMM11,%XMM16,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM7,%XMM12,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMOVSD %XMM7,(%R14,%R11,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
VMOVSD (%RSI,%R13,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VMOVSD (%RSI,%RCX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VSUBSD (%RSI,%RAX,8),%XMM12,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VSUBSD %XMM12,%XMM14,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM15,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VUCOMISD %XMM13,%XMM16	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JBE 226290 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x11a0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMULSD %XMM8,%XMM11,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VANDPD %XMM3,%XMM7,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VANDPD %XMM3,%XMM14,%XMM11	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VCMPSD $0x1,%XMM14,%XMM0,%K1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	scal (12.5%)
VDIVSD %XMM8,%XMM13,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VANDPD %XMM3,%XMM15,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMULSD %XMM10,%XMM13,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VFMADD213SD %XMM10,%XMM11,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VMINSD %XMM13,%XMM11,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM4,%XMM9,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM10,%XMM9,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VSUBSD %XMM8,%XMM5,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VXORPD %XMM6,%XMM8,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMOVSD %XMM8,%XMM10,%XMM10{%K1}	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMULSD %XMM10,%XMM9,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
JMP 226290 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x11a0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A

Function	__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1
Source file and lines	advec_cell_kernel.f90:112-157
Module	exec

Ignoring paths for analysis
Too many paths. Rerun with max-paths=8
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = max(0,x) (Fortran instrinsic procedure)

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)
To reference allocatable arrays, use "allocatable" instead of "pointer" pointers or qualify them with the "contiguous" attribute (Fortran 2008)
For structures, limit to one indirection. For example, use a_b%c instead of a%b%c with a_b set to a%b before this loop

Vectorization

Details

Workaround

Try another compiler or update/tune your current one
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: Fortran storage order is column-major: do i do j a(i,j) = b(i,j) (slow, non stride 1) => do i do j a(j,i) = b(i,j) (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): do i a(i)%x = b(i)%x (slow, non stride 1) => do 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.

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

FMA

Detected 3 FMA (fused multiply-add) operations. Presence of both ADD/SUB and MUL operations.

Workaround

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.

CMOVGE: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:118
LEA: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:118
VUCOMISD: 3 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:112
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:135
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:148

Conversion instructions

Detected expensive conversion instructions.

Details

CLTQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-049-0948/intel/CloverLeaf1.3-FC/build/CloverLeaf1.3-FC/CloverLeaf_ref/kernels/advec_cell_kernel.f90:131

Workaround

Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.

Type of elements and instruction set

35 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 29 FP arithmetical operations:

12: addition or subtraction (3 inside FMA instructions)
14: multiply (3 inside FMA instructions)
3: divide

The binary loop is loading 184 bytes (23 double precision FP elements). The binary loop is storing 16 bytes (2 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.14 FP operations per loaded or stored byte.

Unroll opportunity

Loop is potentially data access bound.

Workaround

Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually.

General properties

nb instructions	97
nb uops	101
loop length	460
used x86 registers	16
used mmx registers	0
used xmm registers	18
used ymm registers	0
used zmm registers	0
nb stack references	12
ADD-SUB / MUL ratio	0.82

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	12.63 cycles
front end	12.63 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	4.50	4.50	4.50	4.50	4.50	4.50	6.50	6.50	6.50	6.50	11.25	11.25	11.25	11.25	2.50	2.50
cycles	4.50	4.50	4.50	4.50	4.50	4.50	6.50	6.50	6.50	6.50	11.25	11.25	11.25	11.25	2.50	2.50

Execution ports to units layout:

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

Cycles executing div or sqrt instructions

12.00

Cycles summary

Front-end	12.63
Dispatch	11.25
DIV/SQRT	12.00
Overall L1	12.63

Vectorization ratios

INT

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)
other	0%

all	17%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	47%

INT+FP

all	16%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	41%

Vector efficiency ratios

INT

all	10%
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)
other	10%

all	14%
load	12%
store	12%
mul	12%
add-sub	12%
fma	12%
div/sqrt	12%
other	18%

INT+FP

all	14%
load	12%
store	12%
mul	12%
add-sub	12%
fma	12%
div/sqrt	12%
other	17%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 12.63 cycles. At this rate:

5% of peak load performance is reached (14.57 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (1.27 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck).

ASM code

In the binary file, the address of the loop is: 2262b0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
VADDSD %XMM12,%XMM13,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
INC %RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMULSD %XMM7,%XMM8,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMOVSD %XMM7,(%R9,%R11,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
INC %R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %R11,0x10(%RSP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JE 226220 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1130>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMOVSD (%R10,%R11,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VUCOMISD %XMM7,%XMM0	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JB 2262f0 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1200>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV 0x38(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x50(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x28(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV %RDI,%R13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%RCX,%RDI,1),%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x48(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA 0x1(%RAX,%R11,1),%EAX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	2	0.50	N/A
CMP %ECX,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (6.3%)
CMOVGE %ECX,%EAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
LEA (%RBX,%R11,1),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV %EAX,%EDX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 226319 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x1229>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A
MOV 0x28(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x18(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RAX,%R11,1),%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x58(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RCX,%R11,1),%R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0xf0(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA (%RCX,%R11,1),%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
ADD %R11D,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
MOV %RDI,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VANDPD %XMM1,%XMM7,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
MOVSXD %EDX,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB 0x20(%RSP),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD %R8,%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CLTQ																				scal (12.5%)
ADD %R8,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD %R8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB %R12,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVSD (%RBP,%R13,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VMOVSD (%RBP,%RCX,8),%XMM13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VSUBSD (%RBP,%RAX,8),%XMM11,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMOVSD (%RBX,%R15,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VSUBSD %XMM11,%XMM13,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VXORPD %XMM13,%XMM13,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VDIVSD %XMM8,%XMM9,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VMOVSD (%RBX,%R11,8),%XMM9	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RSP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMULSD %XMM15,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VUCOMISD %XMM13,%XMM16	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
VXORPD %XMM16,%XMM16,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VFMADD213SD %XMM9,%XMM12,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VDIVSD (%RBX,%RDX,8),%XMM9,%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VSUBSD %XMM12,%XMM2,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
JBE 2263de <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x12ee>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VANDPD %XMM3,%XMM15,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VANDPD %XMM3,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VSUBSD %XMM12,%XMM5,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VCMPSD $0x1,%XMM14,%XMM0,%K1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	scal (12.5%)
VMULSD %XMM10,%XMM15,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM15,%XMM16,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VFMADD231SD %XMM9,%XMM16,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VXORPD %XMM6,%XMM12,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMOVSD %XMM12,%XMM16,%XMM16{%K1}	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMULSD %XMM4,%XMM17,%XMM17	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM15,%XMM17,%XMM15	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM16,%XMM15,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VADDSD %XMM11,%XMM16,%XMM12	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM7,%XMM12,%XMM7	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMOVSD %XMM7,(%R14,%R11,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (12.5%)
VMOVSD (%RSI,%R13,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VMOVSD (%RSI,%RCX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VSUBSD (%RSI,%RAX,8),%XMM12,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VSUBSD %XMM12,%XMM14,%XMM14	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM15,%XMM14,%XMM16	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VUCOMISD %XMM13,%XMM16	2	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JBE 226290 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x11a0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMULSD %XMM8,%XMM11,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VANDPD %XMM3,%XMM7,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VANDPD %XMM3,%XMM14,%XMM11	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VCMPSD $0x1,%XMM14,%XMM0,%K1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	scal (12.5%)
VDIVSD %XMM8,%XMM13,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	13	4	scal (12.5%)
VANDPD %XMM3,%XMM15,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMULSD %XMM10,%XMM13,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VFMADD213SD %XMM10,%XMM11,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50	scal (12.5%)
VMINSD %XMM13,%XMM11,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VMULSD %XMM4,%XMM9,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
VMINSD %XMM10,%XMM9,%XMM9	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VSUBSD %XMM8,%XMM5,%XMM8	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (12.5%)
VXORPD %XMM6,%XMM8,%XMM10	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMOVSD %XMM8,%XMM10,%XMM10{%K1}	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMULSD %XMM10,%XMM9,%XMM13	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	scal (12.5%)
JMP 226290 <__nv_advec_cell_kernel_module_advec_cell_kernel__PARALLEL_F2L81_1_+0x11a0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Masked instructions

Details

Workaround

FMA

Workaround

Complex instructions

Details

Conversion instructions

Details

Workaround

Type of elements and instruction set

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

Arithmetic intensity

Unroll opportunity

Workaround

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

Masked instructions

Details

Workaround

FMA

Workaround

Complex instructions

Details

Conversion instructions

Details

Workaround

Type of elements and instruction set

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

Arithmetic intensity

Unroll opportunity

Workaround

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code