Tag: GPU acceleration
Browse our exclusive articles!
StreamServe: Low-Latency LLM Serving with Adaptive Flows
StreamServe boosts LLM serving efficiency with adaptive speculative decoding and metric-aware routing, cutting latency by up to 18x on multi-GPU setups.
Optimizing Prompt Compression for Faster LLM Inference
Explore how prompt compression reduces latency and memory use in LLMs while maintaining quality for faster, cost-effective AI inference.
GaloisSAT: Hybrid GPU-CPU Boolean SAT Solver Breakthrough
Discover GaloisSAT, a hybrid GPU-CPU SAT solver using differentiable programming for faster Boolean satisfiability solving and improved performance.
Differentiable Power-Flow Optimization for Scalable Grids
Discover Differentiable Power-Flow (DPF), a scalable, GPU-accelerated method for efficient power grid optimization and renewable energy management.
SonicMoE: Boost MoE Efficiency with IO & Tile Optimizations
Discover how SonicMoE enhances Mixture of Experts models with IO and tile-aware optimizations, cutting memory use and boosting GPU training speed.