StreamKL is the first fused GPU primitive for attention KL divergence that reduces memory from O(N_Q N_K) to O(1) via an online one-pass formulation and tile-wise recomputation.
hub
Flashattention-4: Algorithm and kernel pipelining co-design for asymmetric hardware scaling
17 Pith papers cite this work. Polarity classification is still indexing.
17
Pith papers citing it
hub tools
citation-role summary
background 1
baseline 1
method 1
other 1
citation-polarity summary
years
2026 17representative citing papers
CUDAHercules benchmark demonstrates that leading LLMs generate functional CUDA code but fail to recover expert-level optimization strategies needed for peak performance on Ampere, Hopper, and Blackwell GPUs.
RPA kernel for TPUs achieves 86% MBU in decode and 73% MFU in prefill on Llama 3 8B via tiling for ragged memory, fused pipelines, and specialized compilation for prefill/decode workloads.
Nautilus auto-compiles math-like tensor descriptions into optimized GPU kernels, delivering up to 42% higher throughput than prior compilers on transformer models across NVIDIA GPUs.
Vortex provides a programmable frontend and backend for sparse attention in LLM serving, delivering up to 3.46x throughput over full attention while preserving accuracy.
OSCAR achieves near-BF16 accuracy for 2-bit KV cache quantization by using offline spectral covariance-aware rotations aligned with attention, plus a custom deployable INT2 kernel compatible with paged serving.
PrismLLM constructs a sliced execution graph and uses hybrid emulation to faithfully reproduce performance and memory behavior of up to 8192-GPU LLM training runs on fewer than 1% of the original GPUs.
TLX introduces MIMW-based extensions to Triton that let developers orchestrate warp-group execution and asynchronous hardware features while preserving blocked programming productivity, with kernels deployed in large-scale training and inference.
GAPE augments RoPE with query- and key-dependent gates to stabilize attention and improve long-context performance in language models.
PRTS pretrains VLA models with contrastive goal-conditioned RL to embed goal-reachability probabilities from offline data, yielding SOTA results on robotic benchmarks especially for long-horizon and novel instructions.
Thin-film lithium niobate modulators implement electro-optic Softmax and Sigmoid alternatives for transformers that maintain competitive accuracy under 4-bit quantization and characterized noise up to 10 GBaud.
LPM 1.0 generates infinite-length, identity-stable, real-time audio-visual conversational performances for single characters using a distilled causal diffusion transformer and a new benchmark.
Jet-Long is a tuning-free bifocal RoPE method that dynamically sets remote group size from sequence length, recovering the base model within the pretrained window and beating prior zero-shot extenders on RULER, HELMET-RAG, and PG-19 with near-FA2 throughput.
Bernini is a framework that uses an MLLM planner to output semantic representations for a DiT renderer to generate or edit videos, reporting SOTA benchmark performance.
VFA optimizes Flash Attention by pre-computing global max approximations from key blocks and reordering traversal to reduce vector bottlenecks while preserving exact computation.
A survey-style reference book mapping the full agentic-AI stack from transformer internals to production deployment, with no new research result.
citing papers explorer
-
StreamKL: Fast and Memory-Efficient KL Divergence for Boosting Attention Distillation
StreamKL is the first fused GPU primitive for attention KL divergence that reduces memory from O(N_Q N_K) to O(1) via an online one-pass formulation and tile-wise recomputation.
-
CUDAHercules: Benchmarking Hardware-Aware Expert-level CUDA Optimization for LLMs
CUDAHercules benchmark demonstrates that leading LLMs generate functional CUDA code but fail to recover expert-level optimization strategies needed for peak performance on Ampere, Hopper, and Blackwell GPUs.
-
Ragged Paged Attention: A High-Performance and Flexible LLM Inference Kernel for TPU
RPA kernel for TPUs achieves 86% MBU in decode and 73% MFU in prefill on Llama 3 8B via tiling for ragged memory, fused pipelines, and specialized compilation for prefill/decode workloads.
-
Nautilus: An Auto-Scheduling Tensor Compiler for Efficient Tiled GPU Kernels
Nautilus auto-compiles math-like tensor descriptions into optimized GPU kernels, delivering up to 42% higher throughput than prior compilers on transformer models across NVIDIA GPUs.
-
Vortex: Efficient and Programmable Sparse Attention Serving for AI Agents
Vortex provides a programmable frontend and backend for sparse attention in LLM serving, delivering up to 3.46x throughput over full attention while preserving accuracy.
-
OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization
OSCAR achieves near-BF16 accuracy for 2-bit KV cache quantization by using offline spectral covariance-aware rotations aligned with attention, plus a custom deployable INT2 kernel compatible with paged serving.
-
A Few GPUs, A Whole Lotta Scale: Faithful LLM Training Emulation with PrismLLM
PrismLLM constructs a sliced execution graph and uses hybrid emulation to faithfully reproduce performance and memory behavior of up to 8192-GPU LLM training runs on fewer than 1% of the original GPUs.
-
TLX: Hardware-Native, Evolvable MIMW GPU Compiler for Large-scale Production Environments
TLX introduces MIMW-based extensions to Triton that let developers orchestrate warp-group execution and asynchronous hardware features while preserving blocked programming productivity, with kernels deployed in large-scale training and inference.
-
Remember to Forget: Gated Adaptive Positional Encoding
GAPE augments RoPE with query- and key-dependent gates to stabilize attention and improve long-context performance in language models.
-
PRTS: A Primitive Reasoning and Tasking System via Contrastive Representations
PRTS pretrains VLA models with contrastive goal-conditioned RL to embed goal-reachability probabilities from offline data, yielding SOTA results on robotic benchmarks especially for long-horizon and novel instructions.
-
Integrated electro-optic attention nonlinearities for transformers
Thin-film lithium niobate modulators implement electro-optic Softmax and Sigmoid alternatives for transformers that maintain competitive accuracy under 4-bit quantization and characterized noise up to 10 GBaud.
-
LPM 1.0: Video-based Character Performance Model
LPM 1.0 generates infinite-length, identity-stable, real-time audio-visual conversational performances for single characters using a distilled causal diffusion transformer and a new benchmark.
-
Jet-Long: Efficient Long-Context Extension with Dynamic Bifocal RoPE
Jet-Long is a tuning-free bifocal RoPE method that dynamically sets remote group size from sequence length, recovering the base model within the pretrained window and beating prior zero-shot extenders on RULER, HELMET-RAG, and PG-19 with near-FA2 throughput.
-
Bernini: Latent Semantic Planning for Video Diffusion
Bernini is a framework that uses an MLLM planner to output semantic representations for a DiT renderer to generate or edit videos, reporting SOTA benchmark performance.
-
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.
-
The Hitchhiker's Guide to Agentic AI: From Foundations to Systems
A survey-style reference book mapping the full agentic-AI stack from transformer internals to production deployment, with no new research result.
- Do Value Vectors in Deep Layers Need Context from the Residual Stream?