SIMD-accelerated computer vision on a $2 microcontroller
Web page · https://shraiwi.com/blog/simd-fast-esp32s3
First seen by Oriel on .
AI summary
A SIMD-accelerated FAST feature detector for the $2 ESP32-S3 reaches 11.2MP/s, roughly doubling reference performance and enabling realtime computer vision processing.
- The implementation uses the ESP32-S3’s 128-bit SIMD instructions to process sixteen eight-bit pixels at a time.
- The detector handles misaligned pixel blocks using register-slice extraction from concatenated registers.
- Because SIMD comparisons are signed and arithmetic saturates, unsigned pixels are remapped by XORing each value with 0x80.
- Testing improved throughput from 5.1MP/s to 11.2MP/s, with QVGA processing taking approximately 6ms.
- The work also produced basm, a basic register allocator, while developing assembly for the ESP32-S3.
#esp32-s3 #simd #computer vision #fast detector
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 226, latest observation ; +220 over observed period since
- Lobsters: 11, latest observation ; +9 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 65, latest observation ; +63 over observed period since
- Lobsters: 0, latest observation ; 0 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50236681 | https://shraiwi.com/blog/simd-fast-esp32s3/ Read discussion |
| Lobsters | qsmmke | https://shraiwi.com/blog/simd-fast-esp32s3/ Read discussion |