XIAO MG24 vs dsPIC33AK512
Cortex-M33 (EFR32MG24) at 39 MHz against dsPIC33A (dsPIC33AK512MPS506) at 200 MHz, on real hardware, same C kernels, clocks verified.
XIAO MG24
Cortex-M33 (EFR32MG24)
- Family
- Arm Cortex-M
- Board
- XIAO MG24 Sense
- Clock
- 39 MHz
- float / double
- 32-bit / 64-bit
dsPIC33AK512
dsPIC33A (dsPIC33AK512MPS506)
- Family
- PIC / dsPIC
- Board
- dsPIC33AK512MPS506 Curiosity Nano
- Clock
- 200 MHz
- float / double
- 32-bit / 32-bit
Typical gap
4.42×
dsPIC33AK512 faster, geometric mean of 30 cells
Biggest win for XIAO MG24
2.87×
uint64_t divide
Biggest win for dsPIC33AK512
23.0×
double subtract
Clock for clock
1.16×
XIAO MG24 needs fewer cycles per op (5.13× clock gap)
Operation by operation
Every cell
How many times faster XIAO MG24 is than dsPIC33AK512, operations per second. Blue: XIAO MG24 ahead. Orange: dsPIC33AK512 ahead.
| Add | Subtract | Multiply | Divide | Modulo | |
|---|---|---|---|---|---|
| uint8_t | 1/6.36× 2.71 M/s · 17.3 M/s | 1/6.36× 2.71 M/s · 17.3 M/s | 1/6.84× 2.52 M/s · 17.3 M/s | 1/3.79× 2.44 M/s · 9.26 M/s | 1/4.08× 1.92 M/s · 7.81 M/s |
| uint16_t | 1/4.36× 2.94 M/s · 12.8 M/s | 1/4.38× 2.93 M/s · 12.8 M/s | 1/5.42× 2.93 M/s · 15.9 M/s | 1/4.44× 2.09 M/s · 9.26 M/s | 1/4.42× 1.77 M/s · 7.81 M/s |
| uint32_t | 1/3.20× 3.19 M/s · 10.2 M/s | 1/3.20× 3.19 M/s · 10.2 M/s | 1/4.11× 2.93 M/s · 12.1 M/s | 1/3.41× 2.21 M/s · 7.52 M/s | 1/3.44× 1.96 M/s · 6.76 M/s |
| uint64_t | 1/4.12× 2.36 M/s · 9.71 M/s | 1/4.11× 2.36 M/s · 9.71 M/s | 1/3.12× 1.97 M/s · 6.14 M/s | 2.87× 285 k/s · 99.2 k/s | 2.85× 285 k/s · 99.8 k/s |
| float | 1/2.67× 3.18 M/s · 8.48 M/s | 1/2.67× 3.18 M/s · 8.48 M/s | 1/2.66× 3.18 M/s · 8.48 M/s | 1/4.10× 1.54 M/s · 6.33 M/s | 1/6.22× 318 k/s · 1.98 M/s |
| double | 1/22.9× 370 k/s · 8.48 M/s | 1/23.0× 368 k/s · 8.48 M/s | 1/16.8× 503 k/s · 8.48 M/s | 1/16.7× 379 k/s · 6.33 M/s | 1/12.0× 165 k/s · 1.98 M/s |