OV - exec - Function ideal_gas_kernel(int, int, int, int, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&) [clone ._omp

	Function: ideal_gas_kernel(int, int, int, int, clover::Buffer2D<double>&, clover::Buffer2D<double>&, ...	Module: exec	Source: ideal_gas.cpp:37-45 [...]	Coverage (incl. loops): 4.53% \| (excl. loops): 0.00%

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp: 37 - 45

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

37: #pragma omp parallel for simd collapse(2)

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

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

40:       double v = 1.0 / density(i, j);

41:       pressure(i, j) = (1.4 - 1.0) * density(i, j) * energy(i, j);

42:       double pressurebyenergy = (1.4 - 1.0) * density(i, j);

43:       double pressurebyvolume = -density(i, j) * pressure(i, j);

44:       double sound_speed_squared = v * v * (pressure(i, j) * pressurebyenergy - pressurebyvolume);

45:       soundspeed(i, j) = std::sqrt(sound_speed_squared);

/home/eoseret/qaas/qaas_runs/178-219-7589/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]; }

0x431000 STP	X29, X30, [SP, #896]!

0x431004 ADD	X29, SP, #0

0x431008 STP	X21, X22, [SP, #32]

0x43100c STP	X27, X28, [SP, #80]

0x431010 LDP	W21, W27, [X0, #40]

0x431014 LDP	W28, W4, [X0, #32]

0x431018 ADD	W21, W21, #1

0x43101c ADD	W27, W27, #2

0x431020 CMP	W21, W27

0x431024 B.GE	431198

0x431028 ADD	W28, W28, #1

0x43102c STP	X19, X20, [SP, #16]

0x431030 ADD	W20, W4, #2

0x431034 SUB	W22, W27, W21

0x431038 CMP	W28, W20

0x43103c B.GE	431210

0x431040 SUB	W3, W20, W28

0x431044 STP	X25, X26, [SP, #64]

0x431048 ORR	X26, XZR, X0

0x43104c MUL	W22, W22, W3

0x431050 STR	W3, [SP, #116]

0x431054 BL	410670

0x431058 ORR	W19, WZR, W0

0x43105c BL	410550

0x431060 UDIV	W1, W22, W19

0x431064 LDR	W3, [SP, #116]

0x431068 MSUB	W2, W1, W19, W22

0x43106c CMP	W0, W2

0x431070 B.CC	431204

0x431074 MADD	W2, W1, W0, W2

0x431078 ADD	W0, W1, W2

0x43107c CMP	W2, W0

0x431080 B.CS	4311a8

0x431084 UDIV	W4, W2, W3

0x431088 ADRP	X5,

0x43108c STP	D14, D15, [SP, #96]

0x431090 FMOV	D14, #1.0000000

0x431094 STP	X23, X24, [SP, #48]

0x431098 LDR	D15, [X5, #496]

0x43109c MSUB	W19, W4, W3, W2

0x4310a0 ADD	W4, W4, W21

0x4310a4 SBFM	X25, X4, #0, #31

0x4310a8 ADD	W19, W19, W28

0x4310ac SUB	W4, W20, W19

(258) 0x4310b0 CMP	W1, W4

(258) 0x4310b4 CSEL	W1, W1, W4, #9

(258) 0x4310b8 ADD	W4, W2, W1

(258) 0x4310bc CMP	W2, W4

(258) 0x4310c0 B.CS	4311e8

(258) 0x4310c4 SBFM	X19, X19, #0, #31

(258) 0x4310c8 SUB	W1, W1, #1

(258) 0x4310cc LDP	X5, X2, [X26]

(258) 0x4310d0 ADD	X20, X19, #1

(258) 0x4310d4 ADD	X22, X20, X1

(258) 0x4310d8 LDR	X1, [X26, #16]

(258) 0x4310dc LDR	X24, [X2]

(258) 0x4310e0 LDR	X21, [X5]

(258) 0x4310e4 LDR	X23, [X1]

(258) 0x4310e8 MUL	X24, X25, X24

(258) 0x4310ec LDR	X1, [X1, #16]

(258) 0x4310f0 MUL	X21, X25, X21

(258) 0x4310f4 LDR	X2, [X2, #16]

(258) 0x4310f8 MUL	X23, X25, X23

(258) 0x4310fc LDR	X5, [X5, #16]

(258) 0x431100 ADD	X23, X1, X23,LSL #3

(258) 0x431104 ADD	X24, X2, X24,LSL #3

(258) 0x431108 ADD	X21, X5, X21,LSL #3

  (259) 0x43110c LDR	D0, [X21, X19,LSL #3]

  (259) 0x431110 LDR	D31, [X24, X19,LSL #3]

  (259) 0x431114 FMUL	D30, D0, D15

  (259) 0x431118 FDIV	D0, D14, D0

  (259) 0x43111c FMUL	D30, D30, D31

  (259) 0x431120 STR	D30, [X23, X19,LSL #3]

  (259) 0x431124 LDR	D29, [X21, X19,LSL #3]

  (259) 0x431128 FMUL	D0, D0, D0

  (259) 0x43112c FMUL	D31, D29, D15

  (259) 0x431130 FMUL	D31, D31, D30

  (259) 0x431134 FMADD	D31, D30, D29, D31

  (259) 0x431138 FMUL	D0, D0, D31

  (259) 0x43113c FCMP	D0, #0

  (259) 0x431140 B.PL	4311c0

  (259) 0x431144 STP	W3, W0, [SP, #116]

  (259) 0x431148 STR	W4, [SP, #124]

  (259) 0x43114c BL	4106a0

  (259) 0x431150 LDR	X2, [X26, #24]

  (259) 0x431154 CMP	X20, X22

  (259) 0x431158 LDP	W3, W0, [SP, #116]

  (259) 0x43115c LDR	W4, [SP, #124]

  (259) 0x431160 LDR	X1, [X2]

  (259) 0x431164 LDR	X2, [X2, #16]

  (259) 0x431168 MADD	X1, X25, X1, X19

  (259) 0x43116c ORR	X19, XZR, X20

  (259) 0x431170 STR	D0, [X2, X1,LSL #3]

  (259) 0x431174 B.EQ	4311e4

  (259) 0x431178 HINT	#0

  (259) 0x43117c HINT	#0

  (259) 0x431180 ADD	X20, X20, #1

  (259) 0x431184 B	43110c

0x431188 LDP	D14, D15, [SP, #96]

0x43118c LDP	X19, X20, [SP, #16]

0x431190 LDP	X23, X24, [SP, #48]

0x431194 LDP	X25, X26, [SP, #64]

0x431198 LDP	X21, X22, [SP, #32]

0x43119c LDP	X27, X28, [SP, #80]

0x4311a0 LDP	X29, X30, [SP], #128

0x4311a4 RET

0x4311a8 LDP	X19, X20, [SP, #16]

0x4311ac LDP	X25, X26, [SP, #64]

0x4311b0 LDP	X21, X22, [SP, #32]

0x4311b4 LDP	X27, X28, [SP, #80]

0x4311b8 LDP	X29, X30, [SP], #128

0x4311bc RET

  (259) 0x4311c0 FSQRT	D0, D0

  (259) 0x4311c4 LDR	X2, [X26, #24]

  (259) 0x4311c8 LDR	X1, [X2]

  (259) 0x4311cc LDR	X2, [X2, #16]

  (259) 0x4311d0 MADD	X1, X25, X1, X19

  (259) 0x4311d4 ORR	X19, XZR, X20

  (259) 0x4311d8 STR	D0, [X2, X1,LSL #3]

  (259) 0x4311dc CMP	X22, X20

  (259) 0x4311e0 B.NE	431180

(258) 0x4311e4 ORR	W2, WZR, W4

(258) 0x4311e8 ADD	X25, X25, #1

(258) 0x4311ec CMP	W27, W25

(258) 0x4311f0 B.LE	431188

(258) 0x4311f4 SUB	W1, W0, W2

(258) 0x4311f8 ORR	W4, WZR, W3

(258) 0x4311fc ORR	W19, WZR, W28

(258) 0x431200 B	4310b0

0x431204 ADD	W1, W1, #1

0x431208 MOVZ	W2, #0

0x43120c B	431074

0x431210 LDP	X19, X20, [SP, #16]

0x431214 B	431198

0x431218 HINT	#0

0x43121c HINT	#0

Coverage (%)	Name	Module
►98.41+	omp_fulfill_event	libgomp.so.1.0.0
○	start_thread	libc.so.6
○	thread_start	libc.so.6

min	med	avg	max

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

min	med	avg	max

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

Path /

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

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run gcc_0

Source file and lines	ideal_gas.cpp:37-45
Module	exec

The function is defined in:

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:37-45
/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 25.00 to 12.50 cycles (2.00x speedup).

Workaround

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

Vectorization

Your function is not vectorized. Only 31% of vector register length is used (average across all VPU instructions). By fully vectorizing your function, you can lower the cost of an iteration from 25.00 to 15.62 cycles (1.60x speedup).

Details

4% of VPU instructions are used in vector version (process two or more data elements in vector registers):

5% of VPU loads are used in vector version.
12% of VPU stores are used in vector version.
0% of VPU addition or subtraction instructions are used in vector version.
0% of VPU multiply instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root instructions are used in vector version.
0% of VPU 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 function can use their full power.

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 25.00 to 8.33 cycles (3.00x speedup).

CALL instructions

Detected function call instructions.

Details

Calling (and then returning from) a function prevents many compiler optimizations (like vectorization), breaks control flow (which reduces pipeline performance) and executes extra instructions to save/restore the registers used inside it, which is very expensive (dozens of cycles). Consider to inline small functions.

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:38
- /home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:38

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 3 bytes. The binary function is storing 3 bytes.

General properties

nb instructions	65
nb uops	63
loop length	260
used w registers	12
used x registers	16
used b registers	0
used h registers	0
used s registers	1
used d registers	2
used q registers	0
used v registers	0
used z registers	0
nb stack references	22

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.88 cycles
front end	7.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	5.00	5.00	7.50	7.50	7.50	7.50	0.50	0.50	0.50	0.50	8.33	8.33	8.33	3.50	3.50
cycles	5.00	5.00	7.50	7.50	7.50	7.50	0.50	0.50	0.50	0.50	8.33	8.33	8.33	3.50	3.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions

10.00-25.00

Cycles summary

Front-end	7.88
Dispatch	8.33
DIV/SQRT	10.00-25.00
Overall L1	10.00-25.00

Vectorization ratios

INT

all	4%
load	5%
store	12%
mul	0%
add-sub	0%
fma	0%
other	0%

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

INT+FP

all	4%
load	5%
store	12%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	0%

Vector efficiency ratios

INT

all	31%
load	43%
store	45%
mul	12%
add-sub	13%
fma	12%
other	23%

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

INT+FP

all	31%
load	43%
store	45%
mul	12%
add-sub	13%
fma	12%
div/sqrt	12%
other	25%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 25.00 cycles. At this rate:

0% of peak load performance is reached (0.12 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.12 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 431000

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #896]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
STP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
STP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDP W21, W27, [X0, #40]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
LDP W28, W4, [X0, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
ADD W21, W21, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
ADD W27, W27, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W21, W27	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 431198 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x198>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W28, W28, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD W20, W4, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SUB W22, W27, W21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W28, W20	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 431210 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x210>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB W3, W20, W28	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
STP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ORR X26, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
MUL W22, W22, W3	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
STR W3, [SP, #116]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (12.5%)
BL 410670 <@plt_start@+0x650>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W19, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
BL 410550 <@plt_start@+0x530>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W1, W22, W19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
LDR W3, [SP, #116]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (12.5%)
MSUB W2, W1, W19, W22	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W0, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.CC 431204 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x204>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W2, W1, W0, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W0, W1, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
CMP W2, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.CS 4311a8 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x1a8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W4, W2, W3	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	scal (12.5%)
ADRP X5, <452088>	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
STP D14, D15, [SP, #96]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50	vect (50.0%)
FMOV D14, #1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (25.0%)
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR D15, [X5, #496]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (25.0%)
MSUB W19, W4, W3, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W4, W4, W21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SBFM X25, X4, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (100.0%)
ADD W19, W19, W28	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SUB W4, W20, W19	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDP D14, D15, [SP, #96]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	vect (50.0%)
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X29, X30, [SP], #128	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X29, X30, [SP], #128	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W1, W1, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOVZ W2, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
B 431074 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x74>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
B 431198 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x198>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
HINT #0																			N/A
HINT #0																			N/A

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run gcc_0

Source file and lines	ideal_gas.cpp:37-45
Module	exec

The function is defined in:

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:37-45
/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

Warnings:
Ignoring paths for analysis

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

4% of VPU instructions are used in vector version (process two or more data elements in vector registers):

5% of VPU loads are used in vector version.
12% of VPU stores are used in vector version.
0% of VPU addition or subtraction instructions are used in vector version.
0% of VPU multiply instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root instructions are used in vector version.
0% of VPU 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 function can use their full power.

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 25.00 to 8.33 cycles (3.00x speedup).

CALL instructions

Detected function call instructions.

Details

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:38
- /home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:38

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 3 bytes. The binary function is storing 3 bytes.

General properties

nb instructions	65
nb uops	63
loop length	260
used w registers	12
used x registers	16
used b registers	0
used h registers	0
used s registers	1
used d registers	2
used q registers	0
used v registers	0
used z registers	0
nb stack references	22

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.88 cycles
front end	7.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	5.00	5.00	7.50	7.50	7.50	7.50	0.50	0.50	0.50	0.50	8.33	8.33	8.33	3.50	3.50
cycles	5.00	5.00	7.50	7.50	7.50	7.50	0.50	0.50	0.50	0.50	8.33	8.33	8.33	3.50	3.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions

10.00-25.00

Cycles summary

Front-end	7.88
Dispatch	8.33
DIV/SQRT	10.00-25.00
Overall L1	10.00-25.00

Vectorization ratios

INT

all	4%
load	5%
store	12%
mul	0%
add-sub	0%
fma	0%
other	0%

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

INT+FP

all	4%
load	5%
store	12%
mul	0%
add-sub	0%
fma	0%
div/sqrt	0%
other	0%

Vector efficiency ratios

INT

all	31%
load	43%
store	45%
mul	12%
add-sub	13%
fma	12%
other	23%

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

INT+FP

all	31%
load	43%
store	45%
mul	12%
add-sub	13%
fma	12%
div/sqrt	12%
other	25%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 25.00 cycles. At this rate:

0% of peak load performance is reached (0.12 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.12 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 431000

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #896]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
STP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
STP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDP W21, W27, [X0, #40]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
LDP W28, W4, [X0, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
ADD W21, W21, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
ADD W27, W27, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W21, W27	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 431198 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x198>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W28, W28, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD W20, W4, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SUB W22, W27, W21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W28, W20	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 431210 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x210>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB W3, W20, W28	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
STP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ORR X26, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
MUL W22, W22, W3	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
STR W3, [SP, #116]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (12.5%)
BL 410670 <@plt_start@+0x650>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W19, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
BL 410550 <@plt_start@+0x530>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W1, W22, W19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
LDR W3, [SP, #116]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (12.5%)
MSUB W2, W1, W19, W22	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W0, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.CC 431204 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x204>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W2, W1, W0, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W0, W1, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
CMP W2, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.CS 4311a8 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x1a8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W4, W2, W3	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	scal (12.5%)
ADRP X5, <452088>	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
STP D14, D15, [SP, #96]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50	vect (50.0%)
FMOV D14, #1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (25.0%)
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR D15, [X5, #496]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (25.0%)
MSUB W19, W4, W3, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W4, W4, W21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SBFM X25, X4, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (100.0%)
ADD W19, W19, W28	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
SUB W4, W20, W19	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDP D14, D15, [SP, #96]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	vect (50.0%)
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X29, X30, [SP], #128	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
LDP X29, X30, [SP], #128	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W1, W1, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOVZ W2, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
B 431074 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x74>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
B 431198 <_Z16ideal_gas_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_._omp_fn.0+0x198>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
HINT #0																			N/A
HINT #0																			N/A

Name	Coverage (%)	Time (s)
▼ideal_gas_kernel(int, int, int, int, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone ._omp_fn.0]–	4.53	6.18
▼Loop 258 - ideal_gas.cpp:39-45 - exec–	0.01	0.02
○Loop 259 - ideal_gas.cpp:40-45 - exec	4.52	6.00

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Expensive FP math instructions/calls

CALL instructions

Details

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

Expensive FP math instructions/calls

CALL instructions

Details

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code