ScaleSearch optimizes block floating point scales via fine-grained search to cut quantization error by 27% for NVFP4, improving PTQ by up to 15 points on MATH500 for Qwen3-8B and attention PPL by 0.77 on Llama 3.1 70B.
Title resolution pending
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
representative citing papers
FlexAttention supplies a compiler-driven interface that expresses common attention variants in a few lines of PyTorch and emits optimized kernels whose speed matches hand-written implementations.
Quest speeds up long-context LLM self-attention by up to 2.23x via query-dependent selection of top-K critical KV cache pages, cutting overall latency by 7.03x with negligible accuracy loss.
FCP shards sequences at block level with flexible P2P communication and bin-packing to achieve near-linear scaling up to 256 GPUs and 1.13x-2.21x higher attention MFU in foundation model pre-training.
citing papers explorer
-
Search Your Block Floating Point Scales!
ScaleSearch optimizes block floating point scales via fine-grained search to cut quantization error by 27% for NVFP4, improving PTQ by up to 15 points on MATH500 for Qwen3-8B and attention PPL by 0.77 on Llama 3.1 70B.
-
Flex Attention: A Programming Model for Generating Optimized Attention Kernels
FlexAttention supplies a compiler-driven interface that expresses common attention variants in a few lines of PyTorch and emits optimized kernels whose speed matches hand-written implementations.
-
Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
Quest speeds up long-context LLM self-attention by up to 2.23x via query-dependent selection of top-K critical KV cache pages, cutting overall latency by 7.03x with negligible accuracy loss.
-
Unleashing Scalable Context Parallelism for Foundation Models Pre-Training via FCP
FCP shards sequences at block level with flexible P2P communication and bin-packing to achieve near-linear scaling up to 256 GPUs and 1.13x-2.21x higher attention MFU in foundation model pre-training.
- Simply Stabilizing the Loop via Fully Looped Transformer