FlashSinkhorn delivers up to 32x forward and 161x end-to-end speedups for entropic OT on A100 GPUs via IO-aware Triton kernels that fuse log-domain updates and streaming transport application.
hub Canonical reference
Online normalizer calculation for softmax
Canonical reference. 100% of citing Pith papers cite this work as background.
23
Pith papers citing it
Background
100% of classified citations
abstract
The Softmax function is ubiquitous in machine learning, multiple previous works suggested faster alternatives for it. In this paper we propose a way to compute classical Softmax with fewer memory accesses and hypothesize that this reduction in memory accesses should improve Softmax performance on actual hardware. The benchmarks confirm this hypothesis: Softmax accelerates by up to 1.3x and Softmax+TopK combined and fused by up to 5x.
hub tools
citation-role summary
background 6
citation-polarity summary
roles
background 6polarities
background 6representative citing papers
MiTA makes attention scalable by gathering query-aware top-k key-value pairs through landmarks as deformable routed experts and compressing the N-width fast-weight MLP into a shared narrower expert.
QFlash implements end-to-end integer FlashAttention with integer-only softmax, delivering up to 8.69x speedup and 18.8% energy savings on ViT models while preserving accuracy under per-tensor quantization.
MMEE encodes dataflow decisions in matrix form for fast exhaustive search, delivering 40-69% lower latency and energy use than prior methods while running 64-343x faster.
Ring Attention uses blockwise computation and ring communication to let Transformers process sequences up to device-count times longer than prior memory-efficient methods.
FlashAttention reduces GPU high-bandwidth memory accesses in self-attention via tiling, delivering exact attention with lower IO complexity, 2-3x wall-clock speedups on models like GPT-2, and the ability to train on sequences up to 64K long.
S2O uses online permutation and importance-based early stopping to increase effective sparsity in attention, delivering 7.51x attention and 3.81x end-to-end speedups on Llama-3.1-8B at 128K context with preserved accuracy.
Flashlight is a compiler-native PyTorch framework that generates efficient fused kernels for arbitrary and data-dependent attention variants, supporting more cases than FlexAttention with competitive performance.
CCE- is a Triton kernel implementation of cross-entropy loss with negative sampling that reduces memory by more than 10x and accelerates training by up to 2x for large-catalog sequential recommenders.
Large-chunk online updates during inference let test-time training scale state capacity to 40% of model size and handle contexts up to 1M tokens without custom kernels.
BatchLLM achieves 1.3x-10.8x higher throughput than vLLM and SGLang for batched LLM inference with prefix sharing via global prefix identification, decoding-first reordering, and memory-centric token batching.
HyLo upcycles Transformer LLMs into hybrids with MLA and Mamba2/Gated DeltaNet blocks via staged training and distillation, extending context to 2M tokens and outperforming prior upcycled hybrids on long-context benchmarks.
Salca is a new ASIC accelerator that achieves 3.82× speedup and 74.19× energy efficiency over A100 for long-context attention via dual-compression dynamic sparse attention and pipelined hardware.
ELSA casts online softmax attention as a prefix scan over monoid (m,S,W) to deliver exact FP32 semantics, O(n) memory, O(log n) depth, and Tensor-Core independence as a drop-in kernel.
Fused compressed-domain int4 attention on Apple Silicon delivers 48x speedup and 3.2x KV cache compression for 128K-context 70B models while matching FP16 token predictions.
Argus generates GPU kernels achieving 99-104% of hand-optimized throughput on key LLM kernels by enforcing compile-time data-flow invariants via a tag-based DSL and an in-context RL planner.
AEGIS reduces inter-GPU communication by up to 81.3% in self-attention and reaches 96.62% scaling efficiency with 3.86x speedup on four GPUs for 2048-token encrypted Transformer inference.
FlashAttention-2 achieves roughly 2x speedup over FlashAttention by parallelizing attention across thread blocks and distributing work within blocks, reaching 50-73% of theoretical peak FLOPs/s on A100 GPUs.
Attention Residuals replaces fixed residual summation with input-dependent softmax attention over preceding layers, and a blocked variant is shown to improve uniformity and downstream performance in a 48B-parameter model pre-trained on 1.4T tokens.
CIMple delivers a 32 kb digital SRAM-based compute-in-memory accelerator for transformer self-attention that reaches 26.1 TOPS/W at 0.85 V in 28 nm with INT8 precision using dual-banked architecture and LUT-based split softmax.
VFA optimizes Flash Attention by pre-computing global max approximations from key blocks and reordering traversal to reduce vector bottlenecks while preserving exact computation.
citing papers explorer
-
FlashSinkhorn: IO-Aware Entropic Optimal Transport on GPU
FlashSinkhorn delivers up to 32x forward and 161x end-to-end speedups for entropic OT on A100 GPUs via IO-aware Triton kernels that fuse log-domain updates and streaming transport application.
-
Mixture-of-Top-k Attention: Efficient Attention via Scalable Fast Weights
MiTA makes attention scalable by gathering query-aware top-k key-value pairs through landmarks as deformable routed experts and compressing the N-width fast-weight MLP into a shared narrower expert.
-
QFlash: Bridging Quantization and Memory Efficiency in Vision Transformer Attention
QFlash implements end-to-end integer FlashAttention with integer-only softmax, delivering up to 8.69x speedup and 18.8% energy savings on ViT models while preserving accuracy under per-tensor quantization.
-
Fast Cross-Operator Optimization of Attention Dataflow
MMEE encodes dataflow decisions in matrix form for fast exhaustive search, delivering 40-69% lower latency and energy use than prior methods while running 64-343x faster.
-
Ring Attention with Blockwise Transformers for Near-Infinite Context
Ring Attention uses blockwise computation and ring communication to let Transformers process sequences up to device-count times longer than prior memory-efficient methods.
-
FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
FlashAttention reduces GPU high-bandwidth memory accesses in self-attention via tiling, delivering exact attention with lower IO complexity, 2-3x wall-clock speedups on models like GPT-2, and the ability to train on sequences up to 64K long.
-
S2O: Early Stopping for Sparse Attention via Online Permutation
S2O uses online permutation and importance-based early stopping to increase effective sparsity in attention, delivering 7.51x attention and 3.81x end-to-end speedups on Llama-3.1-8B at 128K context with preserved accuracy.
-
Flashlight: PyTorch Compiler Extensions to Accelerate Attention Variants
Flashlight is a compiler-native PyTorch framework that generates efficient fused kernels for arbitrary and data-dependent attention variants, supporting more cases than FlexAttention with competitive performance.
-
Faster and Memory-Efficient Training of Sequential Recommendation Models for Large Catalogs
CCE- is a Triton kernel implementation of cross-entropy loss with negative sampling that reduces memory by more than 10x and accelerates training by up to 2x for large-catalog sequential recommenders.
-
Test-Time Training Done Right
Large-chunk online updates during inference let test-time training scale state capacity to 40% of model size and handle contexts up to 1M tokens without custom kernels.
-
BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching
BatchLLM achieves 1.3x-10.8x higher throughput than vLLM and SGLang for batched LLM inference with prefix sharing via global prefix identification, decoding-first reordering, and memory-centric token batching.
-
Long-Context Aware Upcycling: A New Frontier for Hybrid LLM Scaling
HyLo upcycles Transformer LLMs into hybrids with MLA and Mamba2/Gated DeltaNet blocks via staged training and distillation, extending context to 2M tokens and outperforming prior upcycled hybrids on long-context benchmarks.
-
Salca: A Sparsity-Aware Hardware Accelerator for Efficient Long-Context Attention Decoding
Salca is a new ASIC accelerator that achieves 3.82× speedup and 74.19× energy efficiency over A100 for long-context attention via dual-compression dynamic sparse attention and pipelined hardware.
-
ELSA: Exact Linear-Scan Attention for Fast and Memory-Light Vision Transformers
ELSA casts online softmax attention as a prefix scan over monoid (m,S,W) to deliver exact FP32 semantics, O(n) memory, O(log n) depth, and Tensor-Core independence as a drop-in kernel.
-
Open-TQ-Metal: Fused Compressed-Domain Attention for Long-Context LLM Inference on Apple Silicon
Fused compressed-domain int4 attention on Apple Silicon delivers 48x speedup and 3.2x KV cache compression for 128K-context 70B models while matching FP16 token predictions.
-
ARGUS: Agentic GPU Optimization Guided by Data-Flow Invariants
Argus generates GPU kernels achieving 99-104% of hand-optimized throughput on key LLM kernels by enforcing compile-time data-flow invariants via a tag-based DSL and an in-context RL planner.
-
AEGIS: Scaling Long-Sequence Homomorphic Encrypted Transformer Inference via Hybrid Parallelism on Multi-GPU Systems
AEGIS reduces inter-GPU communication by up to 81.3% in self-attention and reaches 96.62% scaling efficiency with 3.86x speedup on four GPUs for 2048-token encrypted Transformer inference.
-
FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning
FlashAttention-2 achieves roughly 2x speedup over FlashAttention by parallelizing attention across thread blocks and distributing work within blocks, reaching 50-73% of theoretical peak FLOPs/s on A100 GPUs.
-
Attention Residuals
Attention Residuals replaces fixed residual summation with input-dependent softmax attention over preceding layers, and a blocked variant is shown to improve uniformity and downstream performance in a 48B-parameter model pre-trained on 1.4T tokens.
-
CIMple: Standard-cell SRAM-based CIM with LUT-based split softmax for attention acceleration
CIMple delivers a 32 kb digital SRAM-based compute-in-memory accelerator for transformer self-attention that reaches 26.1 TOPS/W at 0.85 V in 28 nm with INT8 precision using dual-banked architecture and LUT-based split softmax.
-
VFA: Relieving Vector Operations in Flash Attention with Global Maximum Pre-computation
VFA optimizes Flash Attention by pre-computing global max approximations from key blocks and reordering traversal to reduce vector bottlenecks while preserving exact computation.
- RAT+: Train Dense, Infer Sparse -- Recurrence Augmented Attention for Dilated Inference
- Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs