lux/consensus
2
0
Fork 0
mirror of https://github.com/luxfi/consensus.git synced 2026-09-24 00:26:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: clarify MLX supports Metal (Apple) & CUDA (NVIDIA) GPUs

Browse Source
This commit is contained in:
zeekay
2025-11-10 12:11:02 -08:00
parent 51ad974d76
commit f045fbb32e
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/content/docs/index.mdx
+3 -3
View File
@@ -10,7 +10,7 @@ A high-performance, multi-language consensus engine supporting Chain, DAG, and P
## Features
- **Multi-Language SDKs**: Native implementations in Go, C, Rust, Python, and C++
- **GPU Acceleration**: MLX-powered GPU support for 25-30x speedup on Apple Silicon
- **GPU Acceleration**: MLX-powered GPU support for Apple Silicon (Metal) and NVIDIA (CUDA) with 25-30x speedup
- **Quantum-Resistant**: Post-quantum cryptography with Dilithium and Kyber
- **AI-Powered**: Neural network and LLM-based consensus validation
- **High Performance**: Nanosecond latency, million ops/sec throughput
@@ -49,10 +49,10 @@ Lux Consensus implements multiple consensus families:
| Rust | 607 ns | - | - | - |
| Python + MLX | 16 μs | 60K votes/sec | 250 MB | **359x** |
*GPU benchmarks on Apple M1 Max (10-core CPU, 32-core GPU)*
*GPU benchmarks on Apple M1 Max (10-core CPU, 32-core GPU). Similar performance on NVIDIA GPUs with CUDA backend.*
## Next Steps
- [SDK Documentation](/docs/sdk) - Go, Python, Rust, C++, C
- [MLX GPU Acceleration](/docs/sdk/mlx) - 25-30x speedup on Apple Silicon
- [MLX GPU Acceleration](/docs/sdk/mlx) - Metal (Apple Silicon) & CUDA (NVIDIA) support
- [Benchmarks](/docs/benchmarks) - Detailed performance analysis