This is a chart with performance numbers in cycles/limb for many mpn (i.e., low-level) functions of GMP. A straight number without any special annotations means that the mpn function of that line is implemented for the CPU of that column either in the official repository or in a local repository of a maintainer. For annotated numbers, please see the table above.

To compare these numbers fairly, 32-bit machines should only be compared to 32-bit machines, and 64-bit machines should only be compared to 64-bit machines. A 64-bit machine performs twice the amount of work for many functions, but 4 times the work for multiply primitives, compared to 32-bit machines.

	AMD K7 32	Intel Nor 32	Intel Pres 32	Intel Doth 32	Intel Atom 32	AMD K8 64	AMD K10 64	AMD Bulld 64	AMD Bobc 64	Intel Noc 64	Intel Core2 64	Intel NHM 64	Intel SBR 64	Intel Atom 64	VIA Nano 64	PPC 74x7 32	PPC 970 64	IBM PWR5 64	IBM PWR6 64	IBM PWR7 64	Sun US3 64	Sun T1 64	Alpha 21264 64	Itanium 2 64	S/390 z990 64
add_n	1.64{1.5}	4	4.25	2.14	3	1.5	1.5		2.5	4	2	#2.25	1.61	3	3	4	2	2.25	2.63	2.4	4.5	17	2.125	1.25	3[2.5]
sub_n	1.64{1.5}	4	4.25	2.14	3	1.5	1.5		2.5	4	2	#2.25	1.61	3	3	4	2	2.25	2.63	2.4	4.5	17	2.125	1.25	3[2.5]
addlsh1_n	2.5	4.25	5		6	2	2{1.69}		2.875	5.8	3.1	2.75	2	4.875	3	5	3	2.9	3.5	3		21	[3.25]	1.5	4.75{3.87}
sublsh1_n	2.87				6.667	2.18	2.18{2}		3.25	5.8	3	3.1{2.5}	2.47{2.17}	5	3	5	3	2.9	3.5	3		21	(3.25)	1.5	5
rsblsh1_n					6	2	2{1.69}		2.875	13.75	3.1	2.75	2	4.875	3	[5]						21			4.75{3.87}
addlsh2_n					6	2.1	2		3.3	5.8	3.1	2.75	2	5.75	3	[5]	3	2.9	3.5	3		21		1.5
sublsh2_n					7					5.8	3	3.1	2.47		{4}	[5]	3	2.9	3.5	3		21		1.5
rsblsh2_n					6	2.1	2		3.3	13	3.1	2.75	2	5.75	3	[5]						21
addlsh_n						2.87	2.75		5.46	15	3	2.8	2.75	7.75{6}	4.7{4}									(1.75)
sublsh_n						{2.5-3.25}	{2.5-3.25}				{2.75}	{2.75}	{3}		{4.125}									(1.75)
rsblsh_n						2.87	2.75		5.46	15	3	2.8	2.75	7.75{6}	4.7{4}									(1.75)
add_n_sub_n						[2.5]	[2.5]										(3)						(3)	2.25
rsh1add_n		4.5	5.25			2	2{1.67}		3.25	5.63	3.2{2.67}	3.87[3.3]{2.5}	2.05	5.25	3	(5)	#4	3.5	4.5	3.45			(3.5)	1.5
rsh1sub_n						2	2{1.67}		3.25	5.63	3.2{2.67}	3.87[3.3]{2.5}	2.05	5.25	3	(5)	#4	3.5	4.5	3.45			(3.5)	1.5
mul_1	3.25	4	4.5	4.16{3.75}	7.5	2.5	2.5		5.5	12.6	4	3.75	2.9{2.5}	19.75{17}	4.25	6	7.25	7.25	13.5(8)	2.9	18.25	68	2.25	2{1.5}	22[20]
mul_1c	Y	Y	Y	Y	Y	Y	Y		Y	Y	Y	Y	Y	Y	Y		Y	Y	Y		N		N	[Y]
addmul_1	3.75	5{4}	5	5.21{4.75}	8	2.5	2.5		6.17	14.9	4.25	5{4}	4{3.25}	21.25{19}	5	9.5	8	8	12.25	3.77	17.3	74	3.5	2(1.75)	23[21]
submul_1	3.75	6	6.5	#5.5	8	2.5	2.5		6.17	14.9	4.25	5{4}	4{3.25}	21.25{19}	5	10.5	8.3	8.25	12.8	4.9	22.75	74	3.5	2.25(2)	24[21.75]
mul_2		(4)	(4)			2.25	2.25		5.625	[12.3]	4	3.83{3.67}	3.15	19.5	4.12		(4.75)	(4.75)	(5.5)				(3)	1.5
mul_3																								[1.333]
mul_4																								[1.25]
mul_5																								[1.2]
mul_6																								[1.167]
addmul_2		(4)	(4)			2.375	2.375		5.75	16[13.6]	4.375	4.33{3.75}	3.41	19.9	4.25		(4.75)	(4.75)	(5.5)	(3)	10.25		(3)	1.625	[20.7]
addmul_3																	(4)	(4)	(4)				(3)	{1.42}
addmul_4		(3)	(3)													(2)	"	"	"				(2.31)	{1.3125}
addmul_6																								(1.167)
mul_basecase	3.9[3.75]	4.6¹	5¹	5.3¹	8.9¹	2.5¹	2.5¹			15¹	4.5¹	4.3¹	3.45¹	20.5¹	4.5¹	(2)	8.38¹	8.3¹	13.4¹	4.02¹	(8)		(2.31)	(1+ε)	24.2¹
sqr_basecase	3.9[3.75]	5.3²	5.6²	6.0²	9.7²	#3.0²	#3.0²			#15.8²	#5.1²	#4.75²	3.73²	#21.8²	#4.75²		8.96²	8.67²	#18.5²	4.35²	(8)			(1+ε)	24.8¹
† sqr_diagonal																4							2.3
sqr_diag_addlsh1																								2
redc_1						2.5	2.5										*	*	*					*
redc_2						{2.375}	{2.375}																	*
lshift	1.2	1.75	2	1.75{1.46}	5	#2.35	2.35{1.5}		3.5	3.33	#1.27	1.375[1.25]	1.87	4.5	3.25{2}	2.25(1)	2.33	2.25	4	2.15	2.5	17	1.75	1	3
rshift	1.2	1.75	2	1.75{1.46}	5	#2.35	2.35{1.5}		3.5	3.33	#1.27	1.375	1.77	4.5	3.25{2}	2.25(1)	2.33	2.25	3.5	2.15	2.5	17	1.75	1	3
lshiftc	*	*	*	*	5.5	2.75	2.75(1.75)		4	4.15	1.5	1.75	2.25	5	3.5(2)	2.25	2.33	2.25	4	2.15	2.67	17	*	1.25	3.5
copyd	0.75-1	#2	#2	0.73{0.5}	1.75	1	1		1.75	2.8	1	1{0.5}	1{0.5}	2	2{1}	0.75	#1	1.13	1.9{1}	1.4	2.5	17	1	0.5	1.5
copyi	0.75-1	#1	#1.5	0.73{0.5}	1.75	1.25	1		1.75	2.8	1	1{0.5}	1{0.5}	2	2{1}	0.75	#1	1	2{1}	1.4	2.5	17	1	0.5	0.75
com	1					1.25	1.18		1.75	2.75	1.05	1.5	1.25	2.75	2	(0.75)	1.62	1.425	3.5	1.45			1.5	(0.5)
and_n	{1.5}				3	1.5	1.5\2		2.67	2.75	2	2	1.5	3.75	3	1.14	2	2	2.5	1.75			(1.75)	1	2.75
ior_n	{1.5}				3	1.5	1.5\2		2.67	2.75	2	2	1.5	3.75	3	1.14	2	2	2.5	1.75			(1.75)	1	2.75
xor_n	{1.5}				3	1.5	1.5\2		2.67	2.75	2	2	1.5	3.75	3	1.14	2	2	2.5	1.75			(1.75)	1	2.75
andn_n	{1.75}				3.5	1.5\2.5	1.5\2		2.5-2.75	3.35	2	2	1.75	3.75	3	1.14	2	2	2.5	1.75			(1.75)	1	3.25
iorn_n	{1.75}				3.5	1.5\2.5	1.5\2		2.5-2.75	3.35	2	2	1.75	3.75	3	1.39	2	2	2.5	1.75			(1.75)	1	3.25
xnor_n	{1.75}				3.5	1.5\2.5	1.5\2		2.5	3.35	2	2	1.75	3.75	3	1.39	2	2	2.5	1.75			(1.75)	1	3.25
nand_n	{1.75}				3.5	1.5\1.75	1.5\2		2.5	3.6	2	2	1.75	3.75	3	1.39	2	2	2.5	1.75			(2)	1	3.25
nior_n	{1.75}				3.5	1.5\1.75	1.5\2		2.5	3.6	2	2	1.75	3.75	3	1.14	2	2	2.5	1.75			(2)	1	3.25
† divrem_1 int	17[14]	32	34	24[19]	38[25-28]	13	13		17-18	44	24	19	15[14]	46	24	[21]	29	29	58(52)	25			[22]	30[22]
† divrem_1 frc	15[13]	30	32	17[15]	23[22]	12	12		16	42	19	18	12.4	36	22.6	[7]	19	19	41	14			[18]	30[22]
† pre_divrem_1	Y	Y	Y	Y		Y	Y		Y	Y	Y	Y	Y	Y	Y	*	*						*	Y
div_qr_1u_pi2						{9}	{9}			{34}	{13}	{11.5}	{9.5}		{14.5}
div_qr_1n_pi2						{7.5}	{7.5}			{31}	{13}	{10.5}	{7.5}		{13.5}		{22}	{23.5}	{38}				[16]
† divrem_2	22	63	70	29	44	18	18		27	68	34	30.25	21.3	73	33	29	40	37	62(55)	30.5			29	29
div_qr_2n_pi2						{13.5}	{13.5}			{47}	{23}	{18}	{13.5}	?	{21}
† dive_1	11	19	21	11	16-20	10	10		15	33	#13.25	14	8.5	36	18	[6-8]	16	16	46(39)	12			15	8
bdiv_qr_1_pi2						[8]	[8]			[24.7]	[13.4]	[12.7]	[7]		[15]
† mode1o	11	19	21	11	15	10	10		15	33	13	14.25	8.2	35	18	#8-10	16	16	35	12			15	8
diveby3																6
bdiv_dbm1c	3.5	13.5{7}	11	5	8	2.25	2.25		6.22	12.5	4	3.75	3.6	20	4	6.25	8.25	8.63	15	4.7			3	2	22{20}
mod_1_1p	7	16	18	10	17	6	6		9	26	12.5{10.5}	11{10.5}	8.4[8]	26	13		17	16	30	10.2				(9)
mod_1s_2p						4	4		8.61	19	8	6.5	4.5	28	7.65									(4.5)
mod_1s_3p						{3}	{3}		8	{16}	{5.41}	{4.5}	{3}		{5}
mod_1s_4p	4.75{4.25}	4	4.5	3.4	8.75	3{2.75}	3{2.75}		7.67	15.75	5	4[3.75]	3.25{2.5}	23	4.75{4.17}	[6.5]	9	9	13	3.5			3	(2.25)
mod_34lsub1	#1	1.25	1.25	#1.9	2.33	0.67	0.67		1.125	3.2	1.25	1.15	0.93	2.45	1.25	0.87	1.5	1.32	2.35	1			#1.67	1	2{1+ε}
gcd_1	Y					Y	Y			Y	Y	Y	Y	Y	Y								Y	Y
invert_limb	41					48	48		64	135	69	55	44	130	78	32	86	86	170	66			71	56	86
popcount	5(4)	3.9	4.25	#4.6	5.5	6	1.125		6.1	8	3.67{3}	1.25	1.5	10.75	6.5{5}	1.125	2.25						#1.5	1
hamdist	6(5)	{5.4}	{5.4}	6.08	8	7	2{1.5}		7.5	14.3{10}	8(4)	2{1.5}	2	17.5(12)	10.4(6)	(1.5)	(3)						#2.4	1
	AMD K7 32	Intel Nor 32	Intel Pres 32	Intel Doth 32	Intel Atom 32	AMD K8 64	AMD K10 64	AMD Bulld 64	AMD Bobc 64	Intel Noc 64	Intel Core2 64	Intel NHM 64	Intel SBR 64	Intel Atom 64	VIA Nano 64	PPC 74x7 32	PPC 970 64	IBM PWR5 64	IBM PWR6 64	IBM PWR7 64	Sun US3 64	Sun T1 64	Alpha 21264 64	Itanium 2 64	S/390 z990 64

¹ This value is for sizes around MUL_TOOM22_THRESHOLD, since mpn_mul_basecase is in most cases not used above that.
² This value is for sizes around SQR_TOOM2_THRESHOLD, since mpn_sqr_basecase is never used above that.
† Obsolete function that will be replaced in the next major GMP release.