VCVTPD2PS (XMM, YMM) - Latency

Operands

Operand 1 (w): Register (XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7, XMM8, XMM9, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15)
Operand 2 (r): Register (YMM0, YMM1, YMM2, YMM3, YMM4, YMM5, YMM6, YMM7, YMM8, YMM9, YMM10, YMM11, YMM12, YMM13, YMM14, YMM15)

Latency operand 2 → 1: 7

Latency operand 2 → 1: 7

Experiment 1

Instruction: VCVTPD2PS XMM1, YMM0
Chain instructions: VSHUFPD XMM0, XMM1, XMM1, 0;VSHUFPD XMM0, XMM0, XMM0, 0;VSHUFPD XMM0, XMM0, XMM0, 0;VSHUFPD XMM0, XMM0, XMM0, 0;VSHUFPD XMM0, XMM0, XMM0, 0
Chain latency: 5

Code:

   0:	c5 fd 5a c8          	vcvtpd2ps xmm1,ymm0
   4:	c5 f1 c6 c1 00       	vshufpd xmm0,xmm1,xmm1,0x0
   9:	c5 f9 c6 c0 00       	vshufpd xmm0,xmm0,xmm0,0x0
   e:	c5 f9 c6 c0 00       	vshufpd xmm0,xmm0,xmm0,0x0
  13:	c5 f9 c6 c0 00       	vshufpd xmm0,xmm0,xmm0,0x0
  18:	c5 f9 c6 c0 00       	vshufpd xmm0,xmm0,xmm0,0x0

Init:

MOV R15, 10000;
L: VADDPS YMM0, YMM1, YMM1;
VADDPS YMM0, YMM1, YMM1;
DEC R15;
JNZ L;
VZEROALL;
MOV RAX, 0x4000000040000000;
MOV [R14+0], RAX;
MOV [R14+8], RAX;
MOV [R14+16], RAX;
MOV [R14+24], RAX;
VMOVUPD YMM0, [R14]

Show nanoBench command
Results:
- Instructions retired: 6.0
- Core cycles: 12.0
- Reference cycles: 8.43
- UOPS_EXECUTED.THREAD: 7.0