circuitpython benchmark

Raspberry Pi Pico (RP2040), Adafruit Metro M7 (NXP IMXRT10XX) and Intel i7 CPU benchmark.

abstract

Script for Python and CircuitPython measures computation time, covering:

• int and float datatypes
• vectorized vs. looped operations
• arithmetic, algebraic and trigonometric operations.

results

datatype	operation	Raspberry Pi Pico t (s)		Adafruit Metro M7 t (s)	vs. Pico		Intel i7-6700HQ Laptop t (s)	vs. Pico
int	bitshift	31.614		6.710	4.7 x		0.164	192.5 x
int	modulo	12.077		2.404	5.0 x		0.147	82.0 x
int	bitwise-and	11.597		2.313	5.0 x		0.145	80.1 x
int	bitwise-or	11.596		2.314	5.0 x		0.152	76.5 x
int	bitwise-xor	11.598		2.312	5.0 x		0.152	76.4 x
int	add	19.694		4.051	4.9 x		0.152	129.2 x
float	add	13.319		2.629	5.1 x		0.127	105.1 x
array(np.float)	add	1.561		0.183	8.5 x		0.006	248.7 x
vec speedup		8.5 x		14.4 x			20.2 x
int	sub	11.597		2.313	5.0 x		0.147	79.1 x
float	sub	13.566		2.633	5.2 x		0.132	102.5 x
array(np.float)	sub	1.754		0.179	9.8 x		0.006	272.3 x
vec speedup		7.7 x		14.7 x			20.5 x
int	mul	34.808		6.549	5.3 x		0.144	241.6 x
float	mul	13.383		2.639	5.1 x		0.135	98.9 x
array(np.float)	mul	1.890		0.204	9.3 x		0.006	311.1 x
vec speedup		7.1 x		12.9 x			22.3 x
int	div	13.549		2.381	5.7 x		0.138	97.9 x
float	div	13.991		2.692	5.2 x		0.125	111.5 x
array(np.float)	div	2.097		0.238	8.8 x		0.006	340.9 x
vec speedup		6.7 x		11.3 x			20.4 x
int	exp	20.625		3.221	6.4 x		0.297	69.4 x
float	exp	20.457		3.547	5.8 x		0.285	71.7 x
array(np.float)	exp	6.627		0.510	13.0 x		0.012	548.1 x
vec speedup		3.1 x		7.0 x			23.6 x
int	sqr	17.391		3.241	5.4 x		0.288	60.4 x
float	sqr	17.134		3.301	5.2 x		0.276	62.2 x
array(np.float)	sqr	2.844		1.017	2.8 x		0.006	445.9 x
vec speedup		6.0 x		3.2 x			43.2 x
float	sin	23.491		3.550	6.6 x		0.281	83.7 x
array(np.float)	sin	6.638		0.605	11.0 x		0.015	451.7 x
vec speedup		3.5 x		5.9 x			19.1 x
float	cos	20.729		3.539	5.9 x		0.299	69.3 x
array(np.float)	cos	6.625		0.603	11.0 x		0.014	467.0 x
vec speedup		3.1 x		5.9 x			21.1 x
float	tan	21.281		3.682	5.8 x		0.287	74.0 x
array(np.float)	tan	7.151		0.768	9.3 x		0.025	290.2 x
vec speedup		3.0 x		4.8 x			11.7 x
float	log	23.290		3.514	6.6 x		0.331	70.3 x
array(np.float)	log	8.475		0.625	13.6 x		0.012	724.0 x
vec speedup		2.7 x		5.6 x			28.3 x
for(int)	matmul	9.812		2.049	4.8 x		0.274	35.9 x
M(np.int16)	matmul	1.146		0.045	25.5 x		0.001	1200.4 x
vec speedup		8.6 x		45.6 x			286.5 x
for(float)	matmul	11.020		2.632	4.2 x		0.253	43.6 x
M(np.float)	matmul	0.802		0.041	19.6 x		0.000	1914.1 x
vec speedup		13.7 x		64.2 x			603.6 x