This paper shows how to build a sparse tensor algebra compiler that is agnostic to tensor formats (data layouts). We develop an interface that describes formats in terms of their capabilities and properties, and show how to build a modular code generator where new formats can be added as plugins. We then describe six implementations of the interface that compose to form the dense, CSR/CSF, COO, DIA, ELL, and HASH tensor formats and countless variants thereof. With these implementations at hand, our code generator can generate code to compute any tensor algebra expression on any combination of the aforementioned formats.
To demonstrate our technique, we have implemented it in the taco tensor algebra compiler. Our modular code generator design makes it simple to add support for new tensor formats, and the performance of the generated code is competitive with hand-optimized implementations. Furthermore, by extending taco to support a wider range of formats specialized for different application and data characteristics, we can improve end-user application performance. For example, if input data is provided in the COO format, our technique allows computing a single matrix-vector multiplication directly with the data in COO, which is up to 3.6× faster than by first converting the data to CSR.
Wed 7 Nov Times are displayed in time zone: Guadalajara, Mexico City, Monterrey change
15:30 - 17:00
|Format Abstraction for Sparse Tensor Algebra Compilers|
|ShareJIT: JIT Code Cache Sharing across Processes and its Practical Implementation|
|Reconciling High-level Optimizations and Low-level Code in LLVM|
|An Empirical Study of the Effect of Source-level Loop Transformations on Compiler Stability|
Zhangxiaowen GongUniversity of Illinois at Urbana-Champaign, Zhi ChenUniversity of California, Irvine, Justin SzadayUniversity of Illinois at Urbana-Champaign, David WongIntel, Zehra SuraIBM Research, Neftali Watkinson, Saeed MalekiMicrosoft Research, David PaduaUniversity of Illinois at Urbana-Champaign, Alexander VeidenbaumUniversity of California, Irvine, Alexandru NicolauUniversity of California, Irvine, Josep TorrellasUniversity of Illinois at Urbana-ChampaignMedia Attached