Sign up to save your podcasts
Or
Share CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
🤗 Upvotes: 35 | cs.AR, cs.AI, cs.CL, cs.LG, cs.PL
Authors:
Ahmed Heakl, Sarim Hashmi, Gustavo Bertolo Stahl, Seung Hun Eddie Han, Salman Khan, Abdulrahman Mahmoud
Title:
CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
Arxiv:
http://arxiv.org/abs/2505.16968v3
Abstract:
We introduce CASS, the first large-scale dataset and model suite for cross-architecture GPU code transpilation, targeting both source-level (CUDA <--> HIP) and assembly-level (Nvidia SASS <--> AMD RDNA3) translation. The dataset comprises 70k verified code pairs across host and device, addressing a critical gap in low-level GPU code portability. Leveraging this resource, we train the CASS family of domain-specific language models, achieving 95% source translation accuracy and 37.5% assembly translation accuracy, substantially outperforming commercial baselines such as GPT-4o, Claude, and Hipify. Our generated code matches native performance in over 85% of test cases, preserving runtime and memory behavior. To support rigorous evaluation, we introduce CASS-Bench, a curated benchmark spanning 16 GPU domains with ground-truth execution. All data, models, and evaluation tools are released as open source to foster progress in GPU compiler tooling, binary compatibility, and LLM-guided hardware translation.
View all episodes
By Jingwen Liang, Gengyu Wang🤗 Upvotes: 35 | cs.AR, cs.AI, cs.CL, cs.LG, cs.PL
Authors:
Ahmed Heakl, Sarim Hashmi, Gustavo Bertolo Stahl, Seung Hun Eddie Han, Salman Khan, Abdulrahman Mahmoud
Title:
CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
Arxiv:
http://arxiv.org/abs/2505.16968v3
Abstract:
We introduce CASS, the first large-scale dataset and model suite for cross-architecture GPU code transpilation, targeting both source-level (CUDA <--> HIP) and assembly-level (Nvidia SASS <--> AMD RDNA3) translation. The dataset comprises 70k verified code pairs across host and device, addressing a critical gap in low-level GPU code portability. Leveraging this resource, we train the CASS family of domain-specific language models, achieving 95% source translation accuracy and 37.5% assembly translation accuracy, substantially outperforming commercial baselines such as GPT-4o, Claude, and Hipify. Our generated code matches native performance in over 85% of test cases, preserving runtime and memory behavior. To support rigorous evaluation, we introduce CASS-Bench, a curated benchmark spanning 16 GPU domains with ground-truth execution. All data, models, and evaluation tools are released as open source to foster progress in GPU compiler tooling, binary compatibility, and LLM-guided hardware translation.