# CVTPS2PD
# Convert Packed Single-Precision Floating-Point Values to Packed Double-Precision Floating-Point Values
Opcode | Mnemonic | Description |
---|---|---|
0F 5A /r | CVTPS2PD xmm1, xmm2/m64 | Convert two packed single-precision floating-point values in xmm2/m64 to two packed double-precision floating-point values in xmm1. |
# Description
Converts two packed single-precision floating-point values in the source operand (second operand) to two packed double-precision floating-point values in the destination operand (first operand). The source operand can be an XMM register or a 64-bit memory location. The destination operand is an XMM register. When the source operand is an XMM register, the packed single-precision floating-point values are contained in the low quadword of the register.
# Operation
Destination[0..63] = ConvertFloatToDouble(Source[0..31]);
Destination[64..127] = ConvertFloatToDouble(Source[32..63]);
1
2
3
2
3
# SIMD Floating-Point Exceptions
Invalid, Denormal.
# Protected Mode Exceptions
# Real-Address Mode Exceptions
# Virtual-8086 Mode Exceptions
Same exceptions as in Real Address Mode
#PF(fault-code) | For a page fault. |
#PF(fault-code) | For a page fault. |
Instruction | Latency | Throughput | Execution Unit |
---|---|---|---|
CPUID | 0F3n/0F2n/069n | 0F3n/0F2n/069n | 0F2n |
CVTPS2PD xmm, xmm | 3/2/2+1 | 2/3/- | FP_ADD MMX_SHFT MMX_ALU |