Security and Communication Networks

Research Article

Utilizing the Double-Precision Floating-Point Computing Power of GPUs for RSA Acceleration

Table 2

Computing power comparison among 32-bit integer, SPF, and DPF arithmetic instruction in GTX TITAN.


Instruction type	Parameter	Data format
Instruction type	Parameter	Integer	SPF	DPF

Add	Significant bits	32
	Instructions/SM/CLOCK	160	192	64
	Bits/SM/CLOCK	5120	4416	3328

Multiply(-add)	Significant bits	32
	Instructions/SM/CLOCK	32	192	64
	Bits/SM/CLOCK		27648	43264

and adder should be lower than (SPF) or (DPF) bits to avoid round-off problem. and multiplier should be lower than (SPF) or (DPF) bits to avoid round-off problem. the product of two 32-bit integers requires 2 multiplication instructions, “Bits/SM/CLOCK” for “Integer” is .