ESP32 vs Teensy 4.0
Xtensa LX6 (ESP32) at 240 MHz against Cortex-M7 (i.MX RT1062) at 600 MHz, on real hardware, same C kernels, clocks verified.
Teensy 4.0
Cortex-M7 (i.MX RT1062)
- Family
- Arm Cortex-M
- Board
- Teensy 4.0
- Clock
- 600 MHz
- float / double
- 32-bit / 64-bit
Typical gap
6.82×
Teensy 4.0 faster, geometric mean of 30 cells
Biggest win for ESP32
0.32×
uint64_t divide
Biggest win for Teensy 4.0
34.2×
double divide
Clock for clock
2.73×
Teensy 4.0 needs fewer cycles per op (2.50× clock gap)
Operation by operation
Every cell
How many times faster ESP32 is than Teensy 4.0, operations per second. Blue: ESP32 ahead. Orange: Teensy 4.0 ahead.
| Add | Subtract | Multiply | Divide | Modulo | |
|---|---|---|---|---|---|
| uint8_t | 1/6.77× 11.4 M/s · 77.4 M/s | 1/6.58× 11.4 M/s · 75.3 M/s | 1/6.59× 11.4 M/s · 75.3 M/s | 1/5.01× 10.4 M/s · 52.3 M/s | 1/4.63× 10 M/s · 46.3 M/s |
| uint16_t | 1/6.58× 11.4 M/s · 75.3 M/s | 1/6.56× 11.4 M/s · 75 M/s | 1/6.00× 11.4 M/s · 68.6 M/s | 1/4.45× 9.32 M/s · 41.5 M/s | 1/4.18× 8.97 M/s · 37.5 M/s |
| uint32_t | 1/6.89× 12 M/s · 82.8 M/s | 1/7.14× 12 M/s · 85.7 M/s | 1/5.71× 12 M/s · 68.6 M/s | 1/3.93× 10 M/s · 39.3 M/s | 1/3.63× 9.6 M/s · 34.8 M/s |
| uint64_t | 1/8.96× 7.05 M/s · 63.2 M/s | 1/9.10× 6.94 M/s · 63.2 M/s | 1/9.13× 4.79 M/s · 43.7 M/s | 1/3.17× 1.75 M/s · 5.54 M/s | 1/3.28× 1.69 M/s · 5.54 M/s |
| float | 1/6.48× 12 M/s · 77.7 M/s | 1/6.48× 12 M/s · 77.7 M/s | 1/6.48× 12 M/s · 77.7 M/s | 1/7.10× 3.8 M/s · 27 M/s | 1/3.85× 1.73 M/s · 6.67 M/s |
| double | 1/18.9× 2.96 M/s · 56 M/s | 1/20.4× 2.68 M/s · 54.7 M/s | 1/23.5× 1.86 M/s · 43.7 M/s | 1/34.2× 459 k/s · 15.7 M/s | 1/4.38× 1.11 M/s · 4.84 M/s |