OV - exec - Function viscosity_kernel(int, int, int, int, clover::Buffer1D<double>&, clover::Buffer1D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&) [clone ._omp

	Function: viscosity_kernel(int, int, int, int, clover::Buffer1D<double>&, clover::Buffer1D<double>&, ...	Module: exec	Source: viscosity.cpp:36-64 [...]	Coverage (incl. loops): 2.42% \| (excl. loops): 0.00%

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h: 46 - 69

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

46:   T &operator[](size_t i) const { return data[i]; }

[...]

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

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/viscosity.cpp: 36 - 64

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

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

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

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

39:       double ugrad = (xvel0(i + 1, j + 0) + xvel0(i + 1, j + 1)) - (xvel0(i, j) + xvel0(i + 0, j + 1));

40:       double vgrad = (yvel0(i + 0, j + 1) + yvel0(i + 1, j + 1)) - (yvel0(i, j) + yvel0(i + 1, j + 0));

41:       double div = (celldx[i] * (ugrad) + celldy[j] * (vgrad));

42:       double strain2 = 0.5 * (xvel0(i + 0, j + 1) + xvel0(i + 1, j + 1) - xvel0(i, j) - xvel0(i + 1, j + 0)) / celldy[j] +

43:                        0.5 * (yvel0(i + 1, j + 0) + yvel0(i + 1, j + 1) - yvel0(i, j) - yvel0(i + 0, j + 1)) / celldx[i];

44:       double pgradx = (pressure(i + 1, j + 0) - pressure(i - 1, j + 0)) / (celldx[i] + celldx[i + 1]);

45:       double pgrady = (pressure(i + 0, j + 1) - pressure(i + 0, j - 1)) / (celldy[j] + celldy[j + 2]);

46:       double pgradx2 = pgradx * pgradx;

47:       double pgrady2 = pgrady * pgrady;

48:       double limiter = ((0.5 * (ugrad) / celldx[i]) * pgradx2 + (0.5 * (vgrad) / celldy[j]) * pgrady2 + strain2 * pgradx * pgrady) /

49:                        std::fmax(pgradx2 + pgrady2, g_small);

50:       if ((limiter > 0.0) || (div >= 0.0)) {

51:         viscosity(i, j) = 0.0;

52:       } else {

53:         double dirx = 1.0;

54:         if (pgradx < 0.0) dirx = -1.0;

55:         pgradx = dirx * std::fmax(g_small, std::fabs(pgradx));

56:         double diry = 1.0;

57:         if (pgradx < 0.0) diry = -1.0;

58:         pgrady = diry * std::fmax(g_small, std::fabs(pgrady));

59:         double pgrad = std::sqrt(pgradx * pgradx + pgrady * pgrady);

60:         double xgrad = std::fabs(celldx[i] * pgrad / pgradx);

61:         double ygrad = std::fabs(celldy[j] * pgrad / pgrady);

62:         double grad = std::fmin(xgrad, ygrad);

63:         double grad2 = grad * grad;

64:         viscosity(i, j) = 2.0 * density0(i, j) * grad2 * limiter * limiter;

0x4388a0 STP	X29, X30, [SP, #912]!

0x4388a4 ADD	X29, SP, #0

0x4388a8 STP	X19, X20, [SP, #16]

0x4388ac STP	X25, X26, [SP, #64]

0x4388b0 LDP	W25, W19, [X0, #64]

0x4388b4 LDP	W20, W3, [X0, #56]

0x4388b8 ADD	W25, W25, #1

0x4388bc ADD	W19, W19, #2

0x4388c0 CMP	W25, W19

0x4388c4 B.GE	438bac

0x4388c8 ADD	W20, W20, #1

0x4388cc STP	X21, X22, [SP, #32]

0x4388d0 ADD	W21, W3, #2

0x4388d4 SUB	W26, W19, W25

0x4388d8 CMP	W20, W21

0x4388dc B.GE	438c0c

0x4388e0 ORR	X22, XZR, X0

0x4388e4 STP	X23, X24, [SP, #48]

0x4388e8 SUB	W24, W21, W20

0x4388ec MADD	W26, W26, W24, WZR

0x4388f0 STP	X27, X28, [SP, #80]

0x4388f4 BL	403aa0

0x4388f8 ORR	W23, WZR, W0

0x4388fc BL	403980

0x438900 UDIV	W1, W26, W23

0x438904 ORR	W2, WZR, W0

0x438908 MSUB	W0, W1, W23, W26

0x43890c CMP	W2, W0

0x438910 B.CC	438c00

0x438914 MADD	W2, W1, W2, W0

0x438918 ADD	W27, W1, W2

0x43891c CMP	W2, W27

0x438920 B.CS	438ba0

0x438924 UDIV	W0, W2, W24

0x438928 ADRP	X3,

0x43892c FMOV	D31, #0.5000000

0x438930 LDR	D30, [X3, #3744]

0x438934 LDP	X3, X28, [X22]

0x438938 STR	X3, [SP, #104]

0x43893c LDP	X30, X26, [X22, #16]

0x438940 MSUB	W3, W0, W24, W2

0x438944 ADD	W0, W0, W25

0x438948 SBFM	X13, X0, #0, #31

0x43894c LDP	X23, X25, [X22, #32]

0x438950 ADD	W0, W3, W20

0x438954 SUB	W3, W21, W0

0x438958 LDR	X22, [X22, #48]

0x43895c HINT	#0

  (1012) 0x438960 CMP	W1, W3

  (1012) 0x438964 CSEL	W8, W1, W3, #9

  (1012) 0x438968 ADD	W16, W2, W8

  (1012) 0x43896c CMP	W2, W16

  (1012) 0x438970 B.CS	438bd4

  (1013) 0x438974 LDR	X15, [X26]

  (1013) 0x438978 SBFM	X1, X0, #0, #31

  (1013) 0x43897c SUB	W8, W8, #1

  (1013) 0x438980 ADD	X2, X1, #2

  (1013) 0x438984 ADD	X3, X1, #1

  (1013) 0x438988 LDR	X10, [X23]

  (1013) 0x43898c ADD	X8, X8, X2

  (1013) 0x438990 SUB	W11, W13, #1

  (1013) 0x438994 SBFM	X11, X11, #0, #31

  (1013) 0x438998 ADD	W0, W0, #1

  (1013) 0x43899c LDR	X9, [X26, #16]

  (1013) 0x4389a0 ADD	W12, W13, #2

  (1013) 0x4389a4 SBFM	X0, X0, #0, #31

  (1013) 0x4389a8 MADD	X6, X13, X15, XZR

  (1013) 0x4389ac ADD	X21, X13, #1

  (1013) 0x4389b0 LDR	X4, [X22]

  (1013) 0x4389b4 MADD	X11, X11, X15, XZR

  (1013) 0x4389b8 MADD	X10, X13, X10, XZR

  (1013) 0x4389bc ADD	X2, X3, X6

  (1013) 0x4389c0 LDR	X5, [X25]

  (1013) 0x4389c4 ADD	X2, X9, X2,LSL #3

  (1013) 0x4389c8 SUB	X11, X11, X6

  (1013) 0x4389cc LDR	X9, [X23, #16]

  (1013) 0x4389d0 UBFM	X10, X10, #61, #60

  (1013) 0x4389d4 SUB	X10, X10, #8

  (1013) 0x4389d8 MADD	X6, X13, X4, X4

  (1013) 0x4389dc LDR	X14, [X28, #8]

  (1013) 0x4389e0 ADD	X12, X14, W12,SXTW #3

  (1013) 0x4389e4 MADD	X18, X13, X5, XZR

  (1013) 0x4389e8 LDR	X17, [X22, #16]

  (1013) 0x4389ec SUB	X4, X6, X4

  (1013) 0x4389f0 ADD	X10, X9, X10

  (1013) 0x4389f4 LDR	X9, [SP, #104]

  (1013) 0x4389f8 ADD	X5, X5, X18

  (1013) 0x4389fc ADD	X14, X14, X13,LSL #3

  (1013) 0x438a00 LDR	X7, [X25, #16]

  (1013) 0x438a04 ADD	X4, X17, X4,LSL #3

  (1013) 0x438a08 ADD	X6, X17, X6,LSL #3

  (1013) 0x438a0c LDR	X9, [X9, #8]

  (1013) 0x438a10 ADD	X5, X7, X5,LSL #3

  (1013) 0x438a14 ADD	X7, X7, X18,LSL #3

  (1013) 0x438a18 HINT	#0

  (1013) 0x438a1c HINT	#0

  (1013) 0x438a20 LDR	D21, [X4, X1,LSL #3]

  (1013) 0x438a24 SUB	X17, X2, #8

  (1013) 0x438a28 LDR	D6, [X6, X0,LSL #3]

  (1013) 0x438a2c LDR	D27, [X7, X1,LSL #3]

  (1013) 0x438a30 LDR	D3, [X5, X1,LSL #3]

  (1013) 0x438a34 LDR	D28, [X5, X0,LSL #3]

  (1013) 0x438a38 LDR	D5, [X4, X0,LSL #3]

  (1013) 0x438a3c FADD	D2, D3, D27

  (1013) 0x438a40 LDR	D23, [X6, X1,LSL #3]

  (1013) 0x438a44 FADD	D3, D3, D28

  (1013) 0x438a48 LDR	D29, [X7, X0,LSL #3]

  (1013) 0x438a4c FADD	D18, D5, D21

  (1013) 0x438a50 FADD	D5, D5, D6

  (1013) 0x438a54 LDR	D17, [X9, X1,LSL #3]

  (1013) 0x438a58 LDR	D0, [X14]

  (1013) 0x438a5c FADD	D19, D6, D23

  (1013) 0x438a60 FSUB	D3, D3, S27

  (1013) 0x438a64 LDR	D7, [X17, X15,LSL #3]

  (1013) 0x438a68 FADD	D25, D28, D29

  (1013) 0x438a6c FSUB	D5, D5, S21

  (1013) 0x438a70 LDR	D6, [X17, X11,LSL #3]

  (1013) 0x438a74 MOVI	D21, #0

  (1013) 0x438a78 LDR	D20, [X12]

  (1013) 0x438a7c FSUB	D18, D19, S18

  (1013) 0x438a80 FSUB	D3, D3, S29

  (1013) 0x438a84 LDUR	D1, [X2, #496]

  (1013) 0x438a88 FSUB	D2, D25, S2

  (1013) 0x438a8c FSUB	D5, D5, S23

  (1013) 0x438a90 LDR	D4, [X2]

  (1013) 0x438a94 FSUB	D6, D7, S6

  (1013) 0x438a98 LDR	D22, [X9, X0,LSL #3]

  (1013) 0x438a9c FADD	D20, D0, D20

  (1013) 0x438aa0 FMUL	D3, D3, D31

  (1013) 0x438aa4 FMUL	D5, D5, D31

  (1013) 0x438aa8 FSUB	D1, D4, S1

  (1013) 0x438aac FMUL	D4, D18, D31

  (1013) 0x438ab0 FDIV	D20, D6, D20

  (1013) 0x438ab4 FADD	D22, D17, D22

  (1013) 0x438ab8 FDIV	D5, D5, D17

  (1013) 0x438abc FDIV	D3, D3, D0

  (1013) 0x438ac0 FMUL	D23, D20, D20

  (1013) 0x438ac4 FDIV	D22, D1, D22

  (1013) 0x438ac8 FDIV	D4, D4, D0

  (1013) 0x438acc FADD	D3, D5, D3

  (1013) 0x438ad0 FMUL	D5, D2, D31

  (1013) 0x438ad4 FMUL	D25, D22, D22

  (1013) 0x438ad8 FMUL	D3, D22, D3

  (1013) 0x438adc FDIV	D5, D5, D17

  (1013) 0x438ae0 FMUL	D4, D4, D23

  (1013) 0x438ae4 FADD	D23, D23, D25

  (1013) 0x438ae8 FMAXNM	D23, D23, D30

  (1013) 0x438aec FMADD	D4, D5, D25, D4

  (1013) 0x438af0 FMADD	D3, D3, D20, D4

  (1013) 0x438af4 FDIV	D23, D3, D23

  (1013) 0x438af8 FCMPE	D23, D21

  (1013) 0x438afc B.GT	438b80

  (1013) 0x438b00 FMUL	D2, D17, D2

  (1013) 0x438b04 FMADD	D2, D0, D18, D2

  (1013) 0x438b08 FCMPE	D2, D21

  (1013) 0x438b0c B.GE	438b80

    (1015) 0x438b10 FCMPE	D22, D21

    (1015) 0x438b14 FABS	D19, D22

    (1015) 0x438b18 FMAXNM	D19, D19, D30

    (1015) 0x438b1c B.MI	438bbc

    (1015) 0x438b20 FMOV	D26, #1.0000000

    (1015) 0x438b24 FABS	D20, D20

    (1015) 0x438b28 FMAXNM	D20, D20, D30

    (1015) 0x438b2c FMUL	D20, D20, D26

    (1015) 0x438b30 LDR	X18, [X30]

    (1015) 0x438b34 LDR	X17, [X30, #16]

    (1015) 0x438b38 MADD	X1, X13, X18, X1

    (1015) 0x438b3c FMUL	D24, D20, D20

    (1015) 0x438b40 LDR	D16, [X17, X1,LSL #3]

    (1015) 0x438b44 FMADD	D24, D19, D19, D24

    (1015) 0x438b48 FADD	D16, D16, D16

    (1015) 0x438b4c FSQRT	D24, D24

    (1015) 0x438b50 FMUL	D0, D24, D0

    (1015) 0x438b54 FMUL	D24, D24, D17

    (1015) 0x438b58 FDIV	D0, D0, D20

    (1015) 0x438b5c FDIV	D24, D24, D19

    (1015) 0x438b60 FABS	D0, D0

    (1015) 0x438b64 FABS	D24, D24

    (1015) 0x438b68 FMINNM	D24, D0, D24

    (1015) 0x438b6c FMUL	D24, D24, D24

    (1015) 0x438b70 FMUL	D24, D16, D24

    (1015) 0x438b74 FMUL	D24, D24, D23

    (1015) 0x438b78 FMUL	D21, D24, D23

    (1015) 0x438b7c HINT	#0

  (1013) 0x438b80 STR	D21, [X10, X0,LSL #3]

  (1013) 0x438b84 ADD	X0, X0, #1

  (1013) 0x438b88 ADD	X2, X2, #8

  (1013) 0x438b8c ORR	X1, XZR, X3

  (1013) 0x438b90 CMP	X0, X8

  (1013) 0x438b94 B.EQ	438be0

  (1013) 0x438b98 ADD	X3, X3, #1

  (1013) 0x438b9c B	438a20

(1011) 0x438ba0 LDP	X21, X22, [SP, #32]

(1011) 0x438ba4 LDP	X23, X24, [SP, #48]

(1011) 0x438ba8 LDP	X27, X28, [SP, #80]

(1011) 0x438bac LDP	X19, X20, [SP, #16]

(1011) 0x438bb0 LDP	X25, X26, [SP, #64]

(1011) 0x438bb4 LDP	X29, X30, [SP], #112

(1011) 0x438bb8 RET

  (1014) 0x438bbc FNEG	D1, D19

  (1014) 0x438bc0 FCMPE	D19, #0

  (1014) 0x438bc4 FMOV	D26, #-1.0000000

  (1014) 0x438bc8 FMOV	D19, D1

  (1014) 0x438bcc B.GT	438b24

  (1014) 0x438bd0 B	438b20

  (1012) 0x438bd4 ORR	W16, WZR, W2

  (1012) 0x438bd8 ADD	X21, X13, #1

  (1012) 0x438bdc HINT	#0

  (1012) 0x438be0 ORR	X13, XZR, X21

  (1012) 0x438be4 CMP	W19, W21

  (1012) 0x438be8 B.LE	438ba0

  (1012) 0x438bec SUB	W1, W27, W16

  (1012) 0x438bf0 ORR	W2, WZR, W16

  (1012) 0x438bf4 ORR	W3, WZR, W24

  (1012) 0x438bf8 ORR	W0, WZR, W20

  (1012) 0x438bfc B	438960

0x438c00 ADD	W1, W1, #1

0x438c04 MOVZ	W0, #0

0x438c08 B	438914

0x438c0c LDP	X21, X22, [SP, #32]

0x438c10 B	438bac

0x438c14 HINT	#0

0x438c18 HINT	#0

0x438c1c HINT	#0

Coverage (%)	Name	Source Location	Module
►98.95+	gomp_thread_start	team.c:130	libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.05+	GOMP_parallel	libgomp.h:980	libgomp.so.1.0.0
○	viscosity(global_variables&)	viscosity.cpp:36	exec
○	timestep(global_variables&, pa[...]	timestep.cpp:64	exec
○	hydro(global_variables&, paral[...]	basic_string.h:1076	exec
○	main	clover_leaf.cpp:209	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	clover_leaf.cpp:51	exec

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	viscosity.cpp:36-64
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69
/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/viscosity.cpp:36-64

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 52% of vector register length is used (average across all VPU instructions).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers).

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 1.00 cycles (25.00x speedup).

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 96 bytes. The binary function is storing 104 bytes.

General properties

nb instructions	56
loop length	224
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	4.00	4.00	5.00	5.08	5.00	5.00	6.00	4.92	0.25	0.25	0.25	0.25	5.17	4.83	5.00	3.50	3.50
cycles	4.00	4.00	5.00	5.08	5.00	5.00	6.00	4.92	0.25	0.25	0.25	0.25	5.17	4.83	5.00	3.50	3.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions

10.00-25.00

Cycles summary

Front-end	6.50
Overall L1	10.00-25.00

Vectorization ratios

INT

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

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:

4% of peak load performance is reached (3.84 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
8% of peak store performance is reached (4.16 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4388a0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #912]!	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD X29, SP, #0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
STP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
LDP W25, W19, [X0, #64]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDP W20, W3, [X0, #56]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD W25, W25, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
ADD W19, W19, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W25, W19	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 438bac <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x30c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W20, W20, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
STP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD W21, W3, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
SUB W26, W19, W25	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W20, W21	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 438c0c <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x36c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X22, XZR, X0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
SUB W24, W21, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MADD W26, W26, W24, WZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
STP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
BL 403aa0 <@plt_start@+0x650>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W23, WZR, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
BL 403980 <@plt_start@+0x530>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W1, W26, W23	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ORR W2, WZR, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MSUB W0, W1, W23, W26	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
CMP W2, W0	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.CC 438c00 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x360>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W2, W1, W2, W0	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
ADD W27, W1, W2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W2, W27	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.CS 438ba0 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x300>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W0, W2, W24	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ADRP X3, <43a928>	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
FMOV D31, #0.5000000	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (50.0%)
LDR D30, [X3, #3744]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
LDP X3, X28, [X22]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
STR X3, [SP, #104]	1	0	0	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 X30, X26, [X22, #16]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
MSUB W3, W0, W24, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
ADD W0, W0, W25	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SBFM X13, X0, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
LDP X23, X25, [X22, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
ADD W0, W3, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB W3, W21, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
LDR X22, [X22, #48]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
HINT #0																					N/A
ADD W1, W1, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
MOVZ W0, #0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
B 438914 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x74>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
B 438bac <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x30c>	1	0.50	0.50	0	0	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
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	viscosity.cpp:36-64
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69
/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/viscosity.cpp:36-64

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

Your function is not vectorized. Only 52% of vector register length is used (average across all VPU instructions).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers).

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 1.00 cycles (25.00x speedup).

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 96 bytes. The binary function is storing 104 bytes.

General properties

nb instructions	56
loop length	224
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	4.00	4.00	5.00	5.08	5.00	5.00	6.00	4.92	0.25	0.25	0.25	0.25	5.17	4.83	5.00	3.50	3.50
cycles	4.00	4.00	5.00	5.08	5.00	5.00	6.00	4.92	0.25	0.25	0.25	0.25	5.17	4.83	5.00	3.50	3.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions

10.00-25.00

Cycles summary

Front-end	6.50
Overall L1	10.00-25.00

Vectorization ratios

INT

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

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:

4% of peak load performance is reached (3.84 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
8% of peak store performance is reached (4.16 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4388a0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #912]!	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD X29, SP, #0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
STP X25, X26, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
LDP W25, W19, [X0, #64]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDP W20, W3, [X0, #56]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD W25, W25, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
ADD W19, W19, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W25, W19	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 438bac <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x30c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W20, W20, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
STP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD W21, W3, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
SUB W26, W19, W25	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W20, W21	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 438c0c <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x36c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X22, XZR, X0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
SUB W24, W21, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MADD W26, W26, W24, WZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
STP X27, X28, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
BL 403aa0 <@plt_start@+0x650>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W23, WZR, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
BL 403980 <@plt_start@+0x530>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W1, W26, W23	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ORR W2, WZR, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MSUB W0, W1, W23, W26	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
CMP W2, W0	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.CC 438c00 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x360>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W2, W1, W2, W0	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
ADD W27, W1, W2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
CMP W2, W27	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.CS 438ba0 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x300>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
UDIV W0, W2, W24	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ADRP X3, <43a928>	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
FMOV D31, #0.5000000	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (50.0%)
LDR D30, [X3, #3744]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
LDP X3, X28, [X22]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
STR X3, [SP, #104]	1	0	0	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 X30, X26, [X22, #16]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
MSUB W3, W0, W24, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
ADD W0, W0, W25	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SBFM X13, X0, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
LDP X23, X25, [X22, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
ADD W0, W3, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
SUB W3, W21, W0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
LDR X22, [X22, #48]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
HINT #0																					N/A
ADD W1, W1, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
MOVZ W0, #0	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	1	0.17	N/A
B 438914 <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x74>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
B 438bac <_Z16viscosity_kerneliiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_._omp_fn.0+0x30c>	1	0.50	0.50	0	0	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
HINT #0																					N/A

Name	Coverage (%)	Time (s)
▼viscosity_kernel(int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&) [clone ._omp_fn.0]–	2.42	1.87
▼Loop 1011 - viscosity.cpp:36-64 - exec–	0.00	0.00
○Loop 1013 - viscosity.cpp:39-51 - exec	2.42	1.85
○Loop 1012 - viscosity.cpp:38-40 - exec	0.00	0.00
▼Loop 1014 - viscosity.cpp:54-64 - exec–	0.00	0.00
○Loop 1015 - viscosity.cpp:54-64 - exec	0.00	0.01

Report Configuration

Code clean check

Workaround

Vectorization

Details

Expensive FP math instructions/calls

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Code clean check

Workaround

Vectorization

Details

Expensive FP math instructions/calls

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code