A Perl-based command-line tool that analyzes and optimizes NVIDIA SASS (Shader Assembly) code inside CUBIN files. It uses extracted latency tables to model instruction dependencies and reschedule code for fewer stalls. The tool addresses limitations in ptxas compiler output and is aimed at advanced GPU performance tuning.
In the Developer Tools space, Optimization of SASS Stall Counts takes a focused approach. Automatically optimizing instruction scheduling in compiled CUDA binaries to reduce stalls where compiler heuristics fall short. Optimization of SASS Stall Counts is an open-source project aimed at CUDA developers and GPU performance engineers. Optimization of SASS Stall Counts is open source under the LGPL-2.1 license. It runs on the web and the command line.
Behind Optimization of SASS Stall Counts is windows deep internals, and it first shipped in 2025. Development happens publicly on GitHub with 89 stars and 202 commits in the last 90 days. Key capabilities include instruction scheduling, stall count optimization, and SASS analysis.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Same category — not a similarity match