Stories by limoce

Google drops Gemini 3 Pro image preview

DGX Spark may have only half the performance claimed

You Can Cool Chips with Lasers

Fully Integrated 2D Flash Chip Unveiled

Linux 6.18 UDP receive performance improved by 47%, under DDoS

OpenAI 2025 ICPC Submissions

Unveiling Silicon Art: Dieshots of Microchip Masterpieces

AetherCode: Evaluating LLMs' Ability to Win in Premier Programming Competitions

Operation Costs in CPU Clock Cycles (2016)

Principles and Methodologies for Serial Performance Optimization

The Koala Benchmarks for the Shell

SmallThinker: A Family of Efficient LLMs Natively Trained for Local Deployment

Step3 Technical Report [pdf]

FP8 is ~100 tflops faster when the kernel name has "cutlass" in it

Polaris: A Post-training recipe for scaling RL on Advanced Reasoning models

Overclocking LLM Reasoning: Monitoring and Controlling LLM Thinking Path Lengths

Neutrino: Probing-Based eBPF-Like GPU Kernel Profiling

Machine Learning Conferences Should Establish "Refutations and Critiques" Track

SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines

SepLLM: Accelerate LLMs by Compressing One Segment into One Separator

1
2
3
4
5
6
7