QuanTaichi: a compiler for quantized simulations
<jats:p> High-resolution simulations can deliver great visual quality, but they are often limited by available memory, especially on GPUs. We present a compiler for physical simulation that can achieve both high performance and significantly reduced memory costs, by enabling flexi...
| Main Authors: | , , , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Association for Computing Machinery (ACM)
2022
|
| Online Access: | https://hdl.handle.net/1721.1/143474 |
| _version_ | 1811069477999083520 |
|---|---|
| author | Hu, Yuanming Liu, Jiafeng Yang, Xuanda Xu, Mingkuan Kuang, Ye Xu, Weiwei Dai, Qiang Freeman, William T Durand, Frédo |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Hu, Yuanming Liu, Jiafeng Yang, Xuanda Xu, Mingkuan Kuang, Ye Xu, Weiwei Dai, Qiang Freeman, William T Durand, Frédo |
| author_sort | Hu, Yuanming |
| collection | MIT |
| description | <jats:p>
High-resolution simulations can deliver great visual quality, but they are often limited by available memory, especially on GPUs. We present a compiler for physical simulation that can achieve both high performance and significantly reduced memory costs, by enabling flexible and aggressive
<jats:italic>quantization.</jats:italic>
Low-precision ("quantized") numerical data types are used and packed to represent simulation states, leading to reduced memory space and bandwidth consumption. Quantized simulation allows higher resolution simulation with less memory, which is especially attractive on GPUs. Implementing a quantized simulator that has high performance and packs the data tightly for aggressive storage reduction would be extremely labor-intensive and error-prone using a traditional programming language. To make the creation of quantized simulation practical, we have developed a new set of language abstractions and a compilation system. A suite of tailored domain-specific optimizations ensure quantized simulators often run as fast as the full-precision simulators, despite the overhead of encoding-decoding the packed quantized data types. Our programming language and compiler, based on
<jats:italic>Taichi</jats:italic>
, allow developers to effortlessly switch between different full-precision and quantized simulators, to explore the full design space of quantization schemes, and ultimately to achieve a good balance between space and precision. The creation of quantized simulation with our system has large benefits in terms of memory consumption and performance, on a variety of hardware, from mobile devices to workstations with high-end GPUs. We can simulate with levels of resolution that were previously only achievable on systems with much more memory, such as multiple GPUs. For example, on a
<jats:italic>single</jats:italic>
GPU, we can simulate a Game of Life with 20 billion cells (8× compression per pixel), an Eulerian fluid system with 421 million active voxels (1.6× compression per voxel), and a hybrid Eulerian-Lagrangian elastic object simulation with 235 million particles (1.7× compression per particle). At the same time, quantized simulations create physically plausible results. Our quantization techniques are
<jats:italic>complementary</jats:italic>
to existing acceleration approaches of physical simulation: they can be used in combination with these existing approaches, such as sparse data structures, for even higher scalability and performance.
</jats:p> |
| first_indexed | 2024-09-23T08:11:04Z |
| format | Article |
| id | mit-1721.1/143474 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T08:11:04Z |
| publishDate | 2022 |
| publisher | Association for Computing Machinery (ACM) |
| record_format | dspace |
| spelling | mit-1721.1/1434742023-03-29T20:05:10Z QuanTaichi: a compiler for quantized simulations Hu, Yuanming Liu, Jiafeng Yang, Xuanda Xu, Mingkuan Kuang, Ye Xu, Weiwei Dai, Qiang Freeman, William T Durand, Frédo Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory <jats:p> High-resolution simulations can deliver great visual quality, but they are often limited by available memory, especially on GPUs. We present a compiler for physical simulation that can achieve both high performance and significantly reduced memory costs, by enabling flexible and aggressive <jats:italic>quantization.</jats:italic> Low-precision ("quantized") numerical data types are used and packed to represent simulation states, leading to reduced memory space and bandwidth consumption. Quantized simulation allows higher resolution simulation with less memory, which is especially attractive on GPUs. Implementing a quantized simulator that has high performance and packs the data tightly for aggressive storage reduction would be extremely labor-intensive and error-prone using a traditional programming language. To make the creation of quantized simulation practical, we have developed a new set of language abstractions and a compilation system. A suite of tailored domain-specific optimizations ensure quantized simulators often run as fast as the full-precision simulators, despite the overhead of encoding-decoding the packed quantized data types. Our programming language and compiler, based on <jats:italic>Taichi</jats:italic> , allow developers to effortlessly switch between different full-precision and quantized simulators, to explore the full design space of quantization schemes, and ultimately to achieve a good balance between space and precision. The creation of quantized simulation with our system has large benefits in terms of memory consumption and performance, on a variety of hardware, from mobile devices to workstations with high-end GPUs. We can simulate with levels of resolution that were previously only achievable on systems with much more memory, such as multiple GPUs. For example, on a <jats:italic>single</jats:italic> GPU, we can simulate a Game of Life with 20 billion cells (8× compression per pixel), an Eulerian fluid system with 421 million active voxels (1.6× compression per voxel), and a hybrid Eulerian-Lagrangian elastic object simulation with 235 million particles (1.7× compression per particle). At the same time, quantized simulations create physically plausible results. Our quantization techniques are <jats:italic>complementary</jats:italic> to existing acceleration approaches of physical simulation: they can be used in combination with these existing approaches, such as sparse data structures, for even higher scalability and performance. </jats:p> 2022-06-17T17:57:51Z 2022-06-17T17:57:51Z 2021 2022-06-17T17:36:00Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/143474 Hu, Yuanming, Liu, Jiafeng, Yang, Xuanda, Xu, Mingkuan, Kuang, Ye et al. 2021. "QuanTaichi: a compiler for quantized simulations." ACM Transactions on Graphics, 40 (4). en 10.1145/3450626.3459671 ACM Transactions on Graphics Creative Commons Attribution 4.0 International license https://creativecommons.org/licenses/by/4.0/ application/pdf Association for Computing Machinery (ACM) ACM |
| spellingShingle | Hu, Yuanming Liu, Jiafeng Yang, Xuanda Xu, Mingkuan Kuang, Ye Xu, Weiwei Dai, Qiang Freeman, William T Durand, Frédo QuanTaichi: a compiler for quantized simulations |
| title | QuanTaichi: a compiler for quantized simulations |
| title_full | QuanTaichi: a compiler for quantized simulations |
| title_fullStr | QuanTaichi: a compiler for quantized simulations |
| title_full_unstemmed | QuanTaichi: a compiler for quantized simulations |
| title_short | QuanTaichi: a compiler for quantized simulations |
| title_sort | quantaichi a compiler for quantized simulations |
| url | https://hdl.handle.net/1721.1/143474 |
| work_keys_str_mv | AT huyuanming quantaichiacompilerforquantizedsimulations AT liujiafeng quantaichiacompilerforquantizedsimulations AT yangxuanda quantaichiacompilerforquantizedsimulations AT xumingkuan quantaichiacompilerforquantizedsimulations AT kuangye quantaichiacompilerforquantizedsimulations AT xuweiwei quantaichiacompilerforquantizedsimulations AT daiqiang quantaichiacompilerforquantizedsimulations AT freemanwilliamt quantaichiacompilerforquantizedsimulations AT durandfredo quantaichiacompilerforquantizedsimulations |