x86 Instruction Set Reference


Compute Square Roots of Packed Double-Precision Floating-Point Values

Opcode Mnemonic Description
66 0F 51 /r SQRTPD xmm1, xmm2/m128 Compute the square root of the packed doubleprecision floating-point values in xmm2/m128 and store the results in xmm1.

Performs an SIMD computation of the square roots of the two packed double-precision floatingpoint values in the source operand (second operand) stores the packed double-precision floatingpoint results in the destination operand. The source operand can be an XMM register or a 128- bit memory location. The destination operand is an XMM register. See Figure 11-3 in the IA-32 Intel Architecture Software Developer's Manual, Volume 1 for an illustration of an SIMD double-precision floating-point operation.

Destination[0..63] = SquareRoot(Source[0..63]);
Destination[64..127] = SquareRoot(Source[64..127]);
Instruction Latency Throughput Execution Unit
SQRTPD xmm, xmm70/69/58+5770/69/114FP_DIV