SOL-ExecBench
PulseAugur coverage of SOL-ExecBench — every cluster mentioning SOL-ExecBench across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
CompoundingKernel agent writes top-ranked GPU kernels, outperforming baselines
A new self-improving agent system called CompoundingKernel has achieved top rankings on the SOL-ExecBench leaderboards for GPU kernel generation. Starting with only PyTorch definitions, the system's generated kernels ar…
-
Baidu's ERNIE Evolve system achieves new AI research benchmarks
Baidu has introduced ERNIE Evolve, an automated research system designed to optimize AI training and inference code. This system autonomously explores research directions, modifies code, and conducts experiments, retain…
-
Infinigence AI tops GPU performance charts with 88 first-place rankings
Infinigence AI has achieved significant success on the NVIDIA Research SOL-ExecBench benchmark, securing 88 individual first-place rankings across various GPU operator optimization categories. The company's performance …
-
DuoDiscover system accelerates autonomous discovery with novel evaluation methods
Researchers have developed DuoDiscover, a novel search system designed to overcome the evaluation bottleneck in autonomous discovery for complex engineering tasks. The system employs a fleet of coding agents that simult…
-
New frameworks optimize GPU kernels for deep learning and HPC · 3 sources tracked
Three new research papers introduce advanced frameworks for optimizing GPU kernels, crucial for deep learning and high-performance computing. HIERA focuses on workload-aware planning across different implementation spac…
-
AI-generated CUDA kernels cause silent bugs in deep learning training
AI-generated CUDA kernels, intended to accelerate deep learning computations, have been found to introduce subtle and hard-to-detect bugs. These kernels, which passed NVIDIA's SOL-ExecBench benchmark, failed in real-wor…