matt_d

10 ms·

421.
SuperInfer: SLO-Aware Rotary Scheduling and Memory Management for LLM Inference(supercomputing-system-ai-lab.github.io)
3 points by matt_d 5mo ago|0 comments
422.
KV cache is becoming the memory hierarchy of inference(touchdown-labs.com)
46 points by matt_d 5mo ago|8 comments
423.
Ada-MK: Adaptive MegaKernel Optimization via DAG-Based Search for LLM Inference(arxiv.org)
1 points by matt_d 5mo ago|0 comments
424.
How to Write to SSDs [pdf](vldb.org)
214 points by matt_d 5mo ago|32 comments
425.
Scalable GPU Acceleration of Scalar Functions in Analytical Databases [pdf](vldb.org)
2 points by matt_d 5mo ago|0 comments
426.
The agent principal-agent problem(crawshaw.io)
3 points by matt_d 5mo ago|0 comments
427.
FrontierSmith: Synthesizing Open-Ended Coding Problems at Scale(frontier-cs.org)
2 points by matt_d 5mo ago|0 comments
428.
Demystifying the Silence of Correctness Bugs in PyTorch Compiler(arxiv.org)
3 points by matt_d 5mo ago|0 comments
429.
Fork, Explore, Commit: OS Primitives for Agentic Exploration(arxiv.org)
4 points by matt_d 5mo ago|0 comments
430.
Systematically Auditing AI Agent Benchmarks with BenchJack(arxiv.org)
1 points by matt_d 5mo ago|0 comments
431.
mimalloc: A new, high-performance, scalable memory allocator for the modern era(microsoft.com)
25 points by matt_d 5mo ago|2 comments
432.
Let AI Agents Write Your Serving Stack with VibeServe(syfi.cs.washington.edu)
1 points by matt_d 5mo ago|0 comments
433.
TLX: Hardware-Native, Evolvable MIMW GPU Compiler for Large-Scale Production(arxiv.org)
1 points by matt_d 5mo ago|0 comments
434.
TorchLean: Verified Neural Networks in Lean(robertj1.com)
2 points by matt_d 5mo ago|0 comments
435.
Deterministic Fully-Static Whole-Binary Translation Without Heuristics(arxiv.org)
298 points by matt_d 5mo ago|65 comments
436.
Dynamic Persistent Tile Scheduling w/ Cluster Launch Control (CLC) on Blackwell(research.colfax-intl.com)
2 points by matt_d 5mo ago|0 comments
437.
VibeServe: Can AI Agents Build Bespoke LLM Serving Systems?(github.com)
2 points by matt_d 5mo ago|0 comments
438.
CCL-Bench 1.0: A Trace-Based Benchmark for LLM Infrastructure(arxiv.org)
3 points by matt_d 5mo ago|0 comments
439.
Microbenchmark-Driven Analytical Performance Modeling Across Modern GPUs(arxiv.org)
1 points by matt_d 5mo ago|0 comments
440.
PyTorch DevLog(docs.pytorch.org)
2 points by matt_d 5mo ago|0 comments
441.
VDCores: Resource Decoupled Programming and Execution for Asynchronous GPU(arxiv.org)
2 points by matt_d 5mo ago|0 comments
442.
Aurora: A Leverage-Aware Optimizer for Rectangular Matrices(blog.tilderesearch.com)
1 points by matt_d 5mo ago|0 comments
443.
The Two Abstractions of System Design: Hide or Reduce(muratbuffalo.blogspot.com)
1 points by matt_d 5mo ago|0 comments
444.
Practical Formal Verification for MLIR Programs(arxiv.org)
2 points by matt_d 5mo ago|0 comments
445.
Kerncap: Automated Kernel Extraction and Isolation for AMD GPUs(arxiv.org)
2 points by matt_d 5mo ago|0 comments
446.
Capsules: Compile-time lock discipline in OxCaml(kcsrk.info)
3 points by matt_d 5mo ago|0 comments
447.
Data Race Freedom in OxCaml(kcsrk.info)
2 points by matt_d 6mo ago|0 comments
448.
cuda-oxide: a custom rustc backend for compiling GPU kernels in pure Rust(github.com)
3 points by matt_d 6mo ago|0 comments
449.
A case study with Aeneas and jxl-rs(jonathan.protzenko.fr)
3 points by matt_d 6mo ago|0 comments
450.
Finite Functional Programming(arxiv.org)
3 points by matt_d 6mo ago|0 comments

More ›