Perfect Stream Fusion for Information Flow Processing
Project/Area Number |
21K11821
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Review Section |
Basic Section 60050:Software-related
|
Research Institution | Tohoku University |
Principal Investigator |
Kiselyov Oleg 東北大学, 情報科学研究科, 助教 (50754602)
|
Project Period (FY) |
2021-04-01 – 2025-03-31
|
Project Status |
Granted (Fiscal Year 2022)
|
Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2024: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2023: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2022: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2021: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
|
Keywords | stream processing / info flow processing / sensor fusion / complex event processing / DSL / Stream processing / Info flow processing |
Outline of Research at the Start |
Stream processing is transforming, correlating or reducing possibly unbounded sequences of data, in small space. The widening deployment of sensors (Internet-of-Things, IoT) and the need to analyze and react to the collected data bring stream processing into ordinary life: environment monitoring, traffic management, etc.
Earlier we have developed an approach for stream processing of vast but already available data. It is easy to use and it delivers the highest performance. The goal is to develop the similar approach for sensor data, to correlate and react to continuously-flowing events.
|
Outline of Annual Research Achievements |
Last year we have finally released our stream fusion library, Strymonas 2: Highest-performance Stream Processing: generating high-performance OCaml and C code from declarative stream pipelines (available on Github https://strymonas.github.io/) and presented it at an international (ACM SIGPLAN) workshop.
We have further improved the performance of our FM radio application, based on strymonas, and verified that it is fast enough for live FM radio reception, using the standard SDR (Software Defined Radio) hardware (HackRF).
Crucial to the high performance stream processing is efficient and reliable C code generation. It is implemented in strymonas and described in the "Generating C" paper and several presentations made on this topic. There are many subtleties in generating C, many of which relate to mutable variables. Solving the problem once and for all, without imposing any ad hoc restrictions, required a new theoretical treatment of assignment. The results were presented at several meetings.
|
Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
立てた目標を達成した。結果は、国際会議で発表しました.
|
Strategy for Future Research Activity |
In 2023,the student I'm advising (who is now M2) are planning to write out and present our results about FM Radio and the comparison with GNU Radio at an international meeting (we wrote one short paper last year for an international meeting, but the meeting was cancelled). I'm also working on greatly expanding the "Generating C" paper to a comprehensive journal paper.
We have started work on a new backend for strymonas: WASM.
|
Report
(2 results)
Research Products
(17 results)