OV - exec - Function advec_cell_kernel(int, int, 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>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&, clover::Buffer2D<double>&) [clone .extracted.27]

	Function: advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D<double>&, clover::Buffer1 ...	Module: exec	Source: advec_cell.cpp:65-110 [...]	Coverage (incl. loops): 3.46% \| (excl. loops): 0.00%

/home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp: 65 - 110

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

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

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

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

68:         ({

69:           int upwind, donor, downwind, dif;

70:           double sigmat, sigma3, sigma4, sigmav, sigmam, diffuw, diffdw, limiter, wind;

71:           if (vol_flux_x(i, j) > 0.0) {

72:             upwind = i - 2;

73:             donor = i - 1;

74:             downwind = i;

75:             dif = donor;

76:           } else {

77:             upwind = std::min(i + 1, x_max + 2);

78:             donor = i;

79:             downwind = i - 1;

80:             dif = upwind;

81:           }

82:           sigmat = std::fabs(vol_flux_x(i, j)) / pre_vol(donor, j);

83:           sigma3 = (1.0 + sigmat) * (vertexdx[i] / vertexdx[dif]);

84:           sigma4 = 2.0 - sigmat;

85:           sigmav = sigmat;

86:           diffuw = density1(donor, j) - density1(upwind, j);

87:           diffdw = density1(downwind, j) - density1(donor, j);

88:           wind = 1.0;

89:           if (diffdw <= 0.0) wind = -1.0;

90:           if (diffuw * diffdw > 0.0) {

91:             limiter = (1.0 - sigmav) * wind *

92:                       std::fmin(std::fmin(std::fabs(diffuw), std::fabs(diffdw)),

93:                                 one_by_six * (sigma3 * std::fabs(diffuw) + sigma4 * std::fabs(diffdw)));

94:           } else {

95:             limiter = 0.0;

96:           }

97:           mass_flux_x(i, j) = vol_flux_x(i, j) * (density1(donor, j) + limiter);

98:           sigmam = std::fabs(mass_flux_x(i, j)) / (density1(donor, j) * pre_vol(donor, j));

99:           diffuw = energy1(donor, j) - energy1(upwind, j);

100:           diffdw = energy1(downwind, j) - energy1(donor, j);

101:           wind = 1.0;

102:           if (diffdw <= 0.0) wind = -1.0;

103:           if (diffuw * diffdw > 0.0) {

104:             limiter = (1.0 - sigmam) * wind *

105:                       std::fmin(std::fmin(std::fabs(diffuw), std::fabs(diffdw)),

106:                                 one_by_six * (sigma3 * std::fabs(diffuw) + sigma4 * std::fabs(diffdw)));

107:           } else {

108:             limiter = 0.0;

109:           }

110:           ener_flux(i, j) = mass_flux_x(i, j) * (energy1(donor, j) + limiter);

/home/eoseret/qaas_runs_ZEN5/173-937-4090/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]; }

0x42f2e0 PUSH	%RBP

0x42f2e1 MOV	%RSP,%RBP

0x42f2e4 PUSH	%R15

0x42f2e6 PUSH	%R14

0x42f2e8 PUSH	%R13

0x42f2ea PUSH	%R12

0x42f2ec PUSH	%RBX

0x42f2ed SUB	$0x98,%RSP

0x42f2f4 MOV	%RCX,%RSI

0x42f2f7 MOV	%RDX,-0x30(%RBP)

0x42f2fb MOV	0x50(%RBP),%RAX

0x42f2ff MOV	0x40(%RBP),%R14

0x42f303 MOV	0x38(%RBP),%RCX

0x42f307 MOV	%RCX,-0x60(%RBP)

0x42f30b MOV	0x30(%RBP),%RCX

0x42f30f MOV	%RCX,-0xb0(%RBP)

0x42f316 MOV	0x28(%RBP),%RBX

0x42f31a MOV	0x20(%RBP),%R15

0x42f31e MOV	0x18(%RBP),%R12

0x42f322 MOV	0x10(%RBP),%R13

0x42f326 MOVL	$0,-0x50(%RBP)

0x42f32d TEST	%RAX,%RAX

0x42f330 JS	42f3bc

0x42f336 MOV	%RSI,-0x40(%RBP)

0x42f33a MOV	%R8,-0x48(%RBP)

0x42f33e MOV	%R9,-0x38(%RBP)

0x42f342 MOV	(%RDI),%ESI

0x42f344 MOVQ	$0,-0x78(%RBP)

0x42f34c MOV	%RAX,-0x70(%RBP)

0x42f350 MOVQ	$0x1,-0xb8(%RBP)

0x42f35b SUB	$0x8,%RSP

0x42f35f LEA	-0xb8(%RBP),%RAX

0x42f366 LEA	-0x50(%RBP),%RCX

0x42f36a LEA	-0x78(%RBP),%R8

0x42f36e LEA	-0x70(%RBP),%R9

0x42f372 MOV	$0x491700,%EDI

0x42f377 MOV	%ESI,-0x4c(%RBP)

0x42f37a MOV	$0x22,%EDX

0x42f37f PUSH	$0x1

0x42f381 PUSH	$0x1

0x42f383 PUSH	%RAX

0x42f384 CALL	404230 <__kmpc_for_static_init_8@plt>

0x42f389 ADD	$0x20,%RSP

0x42f38d MOV	-0x78(%RBP),%RCX

0x42f391 MOV	-0x70(%RBP),%RAX

0x42f395 MOV	%RAX,-0x58(%RBP)

0x42f399 CMP	%RAX,%RCX

0x42f39c JBE	42f3ce

0x42f39e MOV	$0x491720,%EDI

0x42f3a3 MOV	-0x4c(%RBP),%ESI

0x42f3a6 ADD	$0x98,%RSP

0x42f3ad POP	%RBX

0x42f3ae POP	%R12

0x42f3b0 POP	%R13

0x42f3b2 POP	%R14

0x42f3b4 POP	%R15

0x42f3b6 POP	%RBP

0x42f3b7 JMP	404050

0x42f3bc ADD	$0x98,%RSP

0x42f3c3 POP	%RBX

0x42f3c4 POP	%R12

0x42f3c6 POP	%R13

0x42f3c8 POP	%R14

0x42f3ca POP	%R15

0x42f3cc POP	%RBP

0x42f3cd RET

0x42f3ce SUB	-0x60(%RBP),%R14D

0x42f3d2 MOV	(%R12),%RAX

0x42f3d6 MOV	%RAX,-0xa8(%RBP)

0x42f3dd MOV	0x10(%R12),%RAX

0x42f3e2 MOV	%RAX,-0xa0(%RBP)

0x42f3e9 MOV	(%R15),%RAX

0x42f3ec MOV	(%RAX),%RDX

0x42f3ef MOV	%RDX,-0x98(%RBP)

0x42f3f6 MOV	0x10(%RAX),%RAX

0x42f3fa MOV	%RAX,-0x68(%RBP)

0x42f3fe MOV	-0x40(%RBP),%RAX

0x42f402 MOV	0x8(%RAX),%R15

0x42f406 MOV	-0x48(%RBP),%RAX

0x42f40a MOV	(%RAX),%RAX

0x42f40d MOV	(%RAX),%RDX

0x42f410 MOV	%RDX,-0x48(%RBP)

0x42f414 MOV	0x10(%RAX),%R12

0x42f418 MOV	(%R13),%RAX

0x42f41c MOV	%RAX,-0x40(%RBP)

0x42f420 MOV	0x10(%R13),%RAX

0x42f424 MOV	%RAX,-0x90(%RBP)

0x42f42b MOV	-0x38(%RBP),%RDX

0x42f42f MOV	(%RDX),%RAX

0x42f432 MOV	%RAX,-0x88(%RBP)

0x42f439 MOV	0x10(%RDX),%R10

0x42f43d MOV	(%RBX),%RAX

0x42f440 MOV	%RAX,-0x38(%RBP)

0x42f444 MOV	0x10(%RBX),%RAX

0x42f448 MOV	%RAX,-0x80(%RBP)

0x42f44c MOV	-0x30(%RBP),%RAX

0x42f450 ADD	$0x2,%EAX

0x42f453 MOV	%RAX,-0x30(%RBP)

0x42f457 INCQ	-0x58(%RBP)

0x42f45b MOVAPD	0x4466d(%RIP),%XMM1

0x42f463 MOVSD	0x445fd(%RIP),%XMM2

0x42f46b MOVSD	0x451ed(%RIP),%XMM3

0x42f473 MOVAPD	0x451d5(%RIP),%XMM4

0x42f47b MOVAPD	0x4464d(%RIP),%XMM5

0x42f483 MOVSD	0x451dd(%RIP),%XMM6

0x42f48b JMP	42f4b9

0x42f48d NOPL	(%RAX)

(283) 0x42f490 ADDSD	%XMM8,%XMM11

(283) 0x42f495 MULSD	%XMM10,%XMM11

(283) 0x42f49a IMUL	-0x38(%RBP),%RSI

(283) 0x42f49f ADD	%RBX,%RSI

(283) 0x42f4a2 MOV	-0x80(%RBP),%RAX

(283) 0x42f4a6 MOVSD	%XMM11,(%RAX,%RSI,8)

(283) 0x42f4ac INC	%RCX

(283) 0x42f4af CMP	%RCX,-0x58(%RBP)

(283) 0x42f4b3 JE	42f39e

(283) 0x42f4b9 MOV	%RCX,%R8

(283) 0x42f4bc SHR	$0x20,%R8

(283) 0x42f4c0 JE	42f4e0

(283) 0x42f4c2 MOV	%RCX,%RAX

(283) 0x42f4c5 XOR	%EDX,%EDX

(283) 0x42f4c7 DIV	%R14

(283) 0x42f4ca MOV	%RAX,%RSI

(283) 0x42f4cd TEST	%R8,%R8

(283) 0x42f4d0 JE	42f4ee

(283) 0x42f4d2 MOV	%RCX,%RAX

(283) 0x42f4d5 CQTO

(283) 0x42f4d7 IDIV	%R14

(283) 0x42f4da JMP	42f4f5

0x42f4dc NOPL	(%RAX)

(283) 0x42f4e0 MOV	%ECX,%EAX

(283) 0x42f4e2 XOR	%EDX,%EDX

(283) 0x42f4e4 DIV	%R14D

(283) 0x42f4e7 MOV	%EAX,%ESI

(283) 0x42f4e9 TEST	%R8,%R8

(283) 0x42f4ec JNE	42f4d2

(283) 0x42f4ee MOV	%ECX,%EAX

(283) 0x42f4f0 XOR	%EDX,%EDX

(283) 0x42f4f2 DIV	%R14D

(283) 0x42f4f5 ADD	-0xb0(%RBP),%RSI

(283) 0x42f4fc ADD	-0x60(%RBP),%EDX

(283) 0x42f4ff MOVSXD	%EDX,%RBX

(283) 0x42f502 MOVSXD	%ESI,%RSI

(283) 0x42f505 MOV	-0xa8(%RBP),%RAX

(283) 0x42f50c IMUL	%RSI,%RAX

(283) 0x42f510 ADD	%RBX,%RAX

(283) 0x42f513 MOV	-0xa0(%RBP),%RDI

(283) 0x42f51a MOVSD	(%RDI,%RAX,8),%XMM8

(283) 0x42f520 XORPD	%XMM0,%XMM0

(283) 0x42f524 UCOMISD	%XMM8,%XMM0

(283) 0x42f529 JAE	42f540

(283) 0x42f52b LEA	-0x2(%RDX),%EAX

(283) 0x42f52e DEC	%EDX

(283) 0x42f530 MOVSXD	%EDX,%R9

(283) 0x42f533 MOVSXD	%EAX,%RDX

(283) 0x42f536 MOV	%RBX,%R8

(283) 0x42f539 MOV	%R9,%RAX

(283) 0x42f53c JMP	42f55a

0x42f53e XCHG	%AX,%AX

(283) 0x42f540 LEA	0x1(%RDX),%EAX

(283) 0x42f543 MOV	-0x30(%RBP),%RDI

(283) 0x42f547 CMP	%EAX,%EDI

(283) 0x42f549 CMOVL	%EDI,%EAX

(283) 0x42f54c DEC	%EDX

(283) 0x42f54e MOVSXD	%EDX,%R8

(283) 0x42f551 MOVSXD	%EAX,%RDX

(283) 0x42f554 MOV	%RDX,%R9

(283) 0x42f557 MOV	%RBX,%RAX

(283) 0x42f55a MOVAPD	%XMM8,%XMM14

(283) 0x42f55f ANDPD	%XMM1,%XMM14

(283) 0x42f564 MOV	-0x98(%RBP),%R13

(283) 0x42f56b IMUL	%RSI,%R13

(283) 0x42f56f ADD	%RAX,%R13

(283) 0x42f572 MOV	-0x68(%RBP),%RDI

(283) 0x42f576 DIVSD	(%RDI,%R13,8),%XMM14

(283) 0x42f57c MOVAPD	%XMM14,%XMM7

(283) 0x42f581 ADDSD	%XMM2,%XMM7

(283) 0x42f585 MULSD	(%R15,%RBX,8),%XMM7

(283) 0x42f58b DIVSD	(%R15,%R9,8),%XMM7

(283) 0x42f591 MOV	-0x48(%RBP),%RDI

(283) 0x42f595 IMUL	%RSI,%RDI

(283) 0x42f599 LEA	(%RDI,%RAX,1),%R9

(283) 0x42f59d MOVSD	(%R12,%R9,8),%XMM12

(283) 0x42f5a3 LEA	(%RDI,%RDX,1),%R11

(283) 0x42f5a7 MOVAPD	%XMM12,%XMM15

(283) 0x42f5ac SUBSD	(%R12,%R11,8),%XMM15

(283) 0x42f5b2 ADD	%R8,%RDI

(283) 0x42f5b5 MOVSD	(%R12,%RDI,8),%XMM13

(283) 0x42f5bb MOVAPD	%XMM3,%XMM9

(283) 0x42f5c0 SUBSD	%XMM14,%XMM9

(283) 0x42f5c5 SUBSD	%XMM12,%XMM13

(283) 0x42f5ca MOVAPD	%XMM13,%XMM10

(283) 0x42f5cf MULSD	%XMM15,%XMM10

(283) 0x42f5d4 XORPD	%XMM11,%XMM11

(283) 0x42f5d9 UCOMISD	%XMM11,%XMM10

(283) 0x42f5de XORPD	%XMM10,%XMM10

(283) 0x42f5e3 JBE	42f64a

(283) 0x42f5e5 MOVAPD	%XMM2,%XMM10

(283) 0x42f5ea SUBSD	%XMM14,%XMM10

(283) 0x42f5ef XORPD	%XMM14,%XMM14

(283) 0x42f5f4 CMPSD	$0x1,%XMM13,%XMM14

(283) 0x42f5fa MOVAPD	%XMM14,%XMM0

(283) 0x42f5ff ANDPD	%XMM10,%XMM0

(283) 0x42f604 XORPD	%XMM4,%XMM10

(283) 0x42f609 ANDNPD	%XMM10,%XMM14

(283) 0x42f60e ORPD	%XMM0,%XMM14

(283) 0x42f613 ANDPD	%XMM5,%XMM15

(283) 0x42f618 ANDPD	%XMM5,%XMM13

(283) 0x42f61d MOVAPD	%XMM15,%XMM0

(283) 0x42f622 MINSD	%XMM13,%XMM0

(283) 0x42f627 MULSD	%XMM7,%XMM15

(283) 0x42f62c MULSD	%XMM9,%XMM13

(283) 0x42f631 ADDSD	%XMM15,%XMM13

(283) 0x42f636 MULSD	%XMM6,%XMM13

(283) 0x42f63b MINSD	%XMM0,%XMM13

(283) 0x42f640 MULSD	%XMM14,%XMM13

(283) 0x42f645 MOVAPD	%XMM13,%XMM10

(283) 0x42f64a ADDSD	%XMM12,%XMM10

(283) 0x42f64f MULSD	%XMM8,%XMM10

(283) 0x42f654 MOV	-0x40(%RBP),%RDI

(283) 0x42f658 IMUL	%RSI,%RDI

(283) 0x42f65c ADD	%RBX,%RDI

(283) 0x42f65f MOV	-0x90(%RBP),%R11

(283) 0x42f666 MOVSD	%XMM10,(%R11,%RDI,8)

(283) 0x42f66c MOV	-0x88(%RBP),%RDI

(283) 0x42f673 IMUL	%RSI,%RDI

(283) 0x42f677 ADD	%RDI,%RAX

(283) 0x42f67a MOVSD	(%R10,%RAX,8),%XMM8

(283) 0x42f680 ADD	%RDI,%RDX

(283) 0x42f683 MOVAPD	%XMM8,%XMM13

(283) 0x42f688 SUBSD	(%R10,%RDX,8),%XMM13

(283) 0x42f68e ADD	%R8,%RDI

(283) 0x42f691 MOVSD	(%R10,%RDI,8),%XMM12

(283) 0x42f697 SUBSD	%XMM8,%XMM12

(283) 0x42f69c MOVAPD	%XMM12,%XMM0

(283) 0x42f6a1 MULSD	%XMM13,%XMM0

(283) 0x42f6a6 UCOMISD	%XMM11,%XMM0

(283) 0x42f6ab JBE	42f490

(283) 0x42f6b1 MOVAPD	%XMM10,%XMM0

(283) 0x42f6b6 ANDPD	%XMM5,%XMM0

(283) 0x42f6ba MOV	-0x68(%RBP),%RAX

(283) 0x42f6be MOVSD	(%RAX,%R13,8),%XMM11

(283) 0x42f6c4 MULSD	(%R12,%R9,8),%XMM11

(283) 0x42f6ca DIVSD	%XMM11,%XMM0

(283) 0x42f6cf MOVAPD	%XMM2,%XMM11

(283) 0x42f6d4 SUBSD	%XMM0,%XMM11

(283) 0x42f6d9 XORPD	%XMM0,%XMM0

(283) 0x42f6dd CMPSD	$0x1,%XMM12,%XMM0

(283) 0x42f6e3 MOVAPD	%XMM0,%XMM14

(283) 0x42f6e8 ANDPD	%XMM11,%XMM14

(283) 0x42f6ed XORPD	%XMM4,%XMM11

(283) 0x42f6f2 ANDNPD	%XMM11,%XMM0

(283) 0x42f6f7 ORPD	%XMM14,%XMM0

(283) 0x42f6fc ANDPD	%XMM5,%XMM13

(283) 0x42f701 ANDPD	%XMM5,%XMM12

(283) 0x42f706 MOVAPD	%XMM13,%XMM11

(283) 0x42f70b MINSD	%XMM12,%XMM11

(283) 0x42f710 MULSD	%XMM7,%XMM13

(283) 0x42f715 MULSD	%XMM9,%XMM12

(283) 0x42f71a ADDSD	%XMM13,%XMM12

(283) 0x42f71f MULSD	%XMM6,%XMM12

(283) 0x42f724 MINSD	%XMM11,%XMM12

(283) 0x42f729 MULSD	%XMM0,%XMM12

(283) 0x42f72e MOVAPD	%XMM12,%XMM11

(283) 0x42f733 JMP	42f490

0x42f738 NOPL	(%RAX,%RAX,1)

Coverage (%)	Name	Source Location	Module
►46.95+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►40.62+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
►6.14+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
○	start_thread		libc.so.6
►3.09+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
►1.38+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	advec_cell_kernel(int, int, in[...]	advec_cell.cpp:65	exec
○	advec_cell_driver(global_varia[...]	advec_cell.cpp:232	exec
○	advection	advection.cpp:50	exec
○	hydro	basic_string.h:1024	exec
○	main	clover_leaf.cpp:208	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	definitions.cpp:0	exec
►1.37+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	advec_cell_kernel(int, int, in[...]	advec_cell.cpp:65	exec
○	advec_cell_driver(global_varia[...]	advec_cell.cpp:232	exec
○	advection	advection.cpp:81	exec
○	hydro	basic_string.h:1024	exec
○	main	clover_leaf.cpp:208	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	definitions.cpp:0	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 icx_10

Source file and lines	advec_cell.cpp:65-110
Module	exec

The function is defined in:

/home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:65-110
/home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69

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 15.75 to 1.75 cycles (9.00x 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 12% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your function, you can lower the cost of an iteration from 15.75 to 1.55 cycles (10.18x speedup).

Details

6% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

21% of SSE/AVX loads are used in vector version.
0% of SSE/AVX stores are used in vector version.
0% of SSE/AVX addition or subtraction instructions are used in vector version.
0% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized 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)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 15.75 to 13.38 cycles (1.18x speedup).

Workaround

Read less array elements
Write less array elements

No data for this section

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.

__kmpc_for_static_init_8@plt: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:65

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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 324 bytes. The binary function is storing 184 bytes.

General properties

nb instructions	110
nb uops	107
loop length	446
used x86 registers	15
used mmx registers	0
used xmm registers	6
used ymm registers	0
used zmm registers	0
nb stack references	27

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	13.38 cycles
front end	13.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	3.83	3.83	3.83	3.83	3.83	3.83	15.75	15.75	15.75	15.75	0.00	0.00	0.00	0.00	0.00	0.00
cycles	3.83	3.83	3.83	3.83	3.83	3.83	15.75	15.75	15.75	15.75	0.00	0.00	0.00	0.00	0.00	0.00

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

Cycles summary

Front-end	13.38
Dispatch	15.75
Overall L1	15.75

Vectorization ratios

INT

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

all	50%
load	50%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	6%
load	21%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	11%
load	10%
store	11%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	11%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

all	18%
load	18%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	12%
load	13%
store	11%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	11%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 15.75 cycles. At this rate:

8% of peak load performance is reached (20.57 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
9% of peak store performance is reached (11.68 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 42f2e0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
MOV %RCX,%RSI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %RDX,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x50(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x40(%RBP),%R14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV 0x38(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV %RCX,-0x60(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x30(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV %RCX,-0xb0(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x28(%RBP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x20(%RBP),%R15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x18(%RBP),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x10(%RBP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOVL $0,-0x50(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
TEST %RAX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
JS 42f3bc <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0xdc>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50-1	N/A
MOV %RSI,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV %R8,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV %R9,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV (%RDI),%ESI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (6.3%)
MOVQ $0,-0x78(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
MOV %RAX,-0x70(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOVQ $0x1,-0xb8(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
SUB $0x8,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
LEA -0xb8(%RBP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x50(%RBP),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x78(%RBP),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x70(%RBP),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
MOV $0x491700,%EDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV %ESI,-0x4c(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
MOV $0x22,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALL 404230 <__kmpc_for_static_init_8@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0.50	N/A
ADD $0x20,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
MOV -0x78(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV -0x70(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x58(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
CMP %RAX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
JBE 42f3ce <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0xee>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50-1	N/A
MOV $0x491720,%EDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV -0x4c(%RBP),%ESI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (6.3%)
ADD $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
JMP 404050 <__kmpc_for_static_fini@plt>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	1	N/A
ADD $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50	N/A
SUB -0x60(%RBP),%R14D	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	1	0.25	scal (6.3%)
MOV (%R12),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0xa8(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%R12),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0xa0(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV (%R15),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RDX,-0x98(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RAX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x68(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x40(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x8(%RAX),%R15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV -0x48(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RDX,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RAX),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%R13),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%R13),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x90(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x38(%RBP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RDX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x88(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RDX),%R10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV (%RBX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RBX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x80(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x30(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
ADD $0x2,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
INCQ -0x58(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	1	0.50	N/A
MOVAPD 0x4466d(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVSD 0x445fd(%RIP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
MOVSD 0x451ed(%RIP),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
MOVAPD 0x451d5(%RIP),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVAPD 0x4464d(%RIP),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVSD 0x451dd(%RIP),%XMM6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
JMP 42f4b9 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0x1d9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	1	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
XCHG %AX,%AX	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run icx_10

Source file and lines	advec_cell.cpp:65-110
Module	exec

The function is defined in:

/home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:65-110
/home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69

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 15.75 to 1.75 cycles (9.00x speedup).

Workaround

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

Vectorization

Details

6% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

21% of SSE/AVX loads are used in vector version.
0% of SSE/AVX stores are used in vector version.
0% of SSE/AVX addition or subtraction instructions are used in vector version.
0% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized 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)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 15.75 to 13.38 cycles (1.18x speedup).

Workaround

Read less array elements
Write less array elements

No data for this section

CALL instructions

Detected function call instructions.

Details

__kmpc_for_static_init_8@plt: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-937-4090/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:65

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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 324 bytes. The binary function is storing 184 bytes.

General properties

nb instructions	110
nb uops	107
loop length	446
used x86 registers	15
used mmx registers	0
used xmm registers	6
used ymm registers	0
used zmm registers	0
nb stack references	27

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	13.38 cycles
front end	13.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	3.83	3.83	3.83	3.83	3.83	3.83	15.75	15.75	15.75	15.75	0.00	0.00	0.00	0.00	0.00	0.00
cycles	3.83	3.83	3.83	3.83	3.83	3.83	15.75	15.75	15.75	15.75	0.00	0.00	0.00	0.00	0.00	0.00

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

Cycles summary

Front-end	13.38
Dispatch	15.75
Overall L1	15.75

Vectorization ratios

INT

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

all	50%
load	50%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	6%
load	21%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	11%
load	10%
store	11%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	11%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

all	18%
load	18%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	12%
load	13%
store	11%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	11%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 15.75 cycles. At this rate:

8% of peak load performance is reached (20.57 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
9% of peak store performance is reached (11.68 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 42f2e0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
MOV %RCX,%RSI	1	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %RDX,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x50(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x40(%RBP),%R14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV 0x38(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV %RCX,-0x60(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x30(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV %RCX,-0xb0(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x28(%RBP),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x20(%RBP),%R15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x18(%RBP),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x10(%RBP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOVL $0,-0x50(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
TEST %RAX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
JS 42f3bc <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0xdc>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50-1	N/A
MOV %RSI,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV %R8,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV %R9,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV (%RDI),%ESI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (6.3%)
MOVQ $0,-0x78(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
MOV %RAX,-0x70(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOVQ $0x1,-0xb8(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
SUB $0x8,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
LEA -0xb8(%RBP),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x50(%RBP),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x78(%RBP),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
LEA -0x70(%RBP),%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.25	N/A
MOV $0x491700,%EDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV %ESI,-0x4c(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (6.3%)
MOV $0x22,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH $0x1	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RAX	1	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
CALL 404230 <__kmpc_for_static_init_8@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0.50	N/A
ADD $0x20,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
MOV -0x78(%RBP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV -0x70(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x58(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
CMP %RAX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
JBE 42f3ce <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0xee>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50-1	N/A
MOV $0x491720,%EDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV -0x4c(%RBP),%ESI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (6.3%)
ADD $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
JMP 404050 <__kmpc_for_static_fini@plt>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	1	N/A
ADD $0x98,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	scal (12.5%)
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	1	0.33	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50	N/A
SUB -0x60(%RBP),%R14D	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	1	0.25	scal (6.3%)
MOV (%R12),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0xa8(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%R12),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0xa0(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV (%R15),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RDX,-0x98(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RAX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x68(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x40(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV 0x8(%RAX),%R15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV -0x48(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RAX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RDX,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RAX),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%R13),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%R13),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x90(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x38(%RBP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV (%RDX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x88(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RDX),%R10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	scal (12.5%)
MOV (%RBX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV 0x10(%RBX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
MOV %RAX,-0x80(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
MOV -0x30(%RBP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.25	N/A
ADD $0x2,%EAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	1	0.17	N/A
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	4	0.50	scal (12.5%)
INCQ -0x58(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	1	0.50	N/A
MOVAPD 0x4466d(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVSD 0x445fd(%RIP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
MOVSD 0x451ed(%RIP),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
MOVAPD 0x451d5(%RIP),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVAPD 0x4464d(%RIP),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	3	0.50	vect (25.0%)
MOVSD 0x451dd(%RIP),%XMM6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	1	0.50	scal (12.5%)
JMP 42f4b9 <_Z17advec_cell_kerneliiiiiiRN6clover8Buffer1DIdEES2_RNS_8Buffer2DIdEES5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_S5_.extracted.27+0x1d9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	1	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
XCHG %AX,%AX	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A

Name	Coverage (%)	Time (s)
▼advec_cell_kernel(int, int, int, int, int, int, clover::Buffer1D&, clover::Buffer1D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D–	3.46	1.29
○Loop 283 - advec_cell.cpp:65-110 - exec	3.46	1.29

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

CALL instructions

Details

Type of elements and instruction set

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

Execution units bottlenecks

Workaround

CALL instructions

Details

Type of elements and instruction set

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