RISC-V China Summit | OpenMPL attracted attention, RISC-V Summit China 2023 concluded successfully

RISC-V China Summit concluded successfully

     On August 25, 2023, the three-day RISC-V China Summit (RISC-V Summit China 2023) concluded successfully. With the theme of "RISC-V Ecological Co-construction", this summit combines the current new global situation, seizes the new global opportunities, and presents new global perspectives and new trends in RISC-V. It attracted more than 100 companies, research institutions, and open source technology communities to participate in exchanges, and nearly 100 media reported on it.

picture

     Pengfeng Technology is committed to becoming a pioneer in the implementation of RISC-V technology in China, and is also a member of the RISC-V International Foundation. In 2017, Pengfeng Technology launched the Perf-V series, the first domestic RISC-V development board, and this year grandly launched the RISC-V Hengshan series computing servers, which enable advanced computing hardware, scientific research innovation and industry applications, and are suitable for Scientific computing, engineering computing, AI computing, fusion computing and other application scenarios help achieve efficient and green computing power costs.

OpenMPL

        On August 25th from 16:20 to 16:30 in the afternoon, Dr. Zhang Xianyi, founder of Pengfeng Technology, shared the relevant developments of the RISC-V high-performance open source computing library OpenMPL at the meeting.

picture

picture

OpenMPL (Open Math Performance Library) is a set of extremely optimized open source mathematical computing software for various computer instruction set processors and applications in various fields . OpenMPL comes with powerful package management capabilities, so it can not only be used as the basic software of a universal mathematical function library for instruction set CPU platforms such as x86, ARM, RISC-V, etc., but can also be used as a necessary support for various DSA acceleration hardware to meet various application scenarios. The basic software will eventually become an open source mathematical computing software for applications in various fields (such as AI, CAE, EDA, signal processing, etc.).

0 1Multi-platform support:

It is suitable for a variety of instruction set processor platforms, including x86, ARM, RISC-V, etc., and provides universal mathematical function library basic software for different computer architectures.

0 2 Extreme Optimization:
Through in-depth optimization and hardware-specific adjustments, it provides efficient mathematical calculation capabilities to maximize the potential of the hardware.

0 3Powerful package management capabilities:

It has built-in powerful package management functions, allowing users to easily manage and use various functional modules and extensions.

04Wide application areas:

It can be widely used in applications in various fields, including artificial intelligence (AI), computer-aided engineering (CAE), electronic design automation (EDA), signal processing, etc., providing high-performance mathematical computing support for these applications.

Dr. Zhang Xianyi pointed out: In the field of RISC-V computing, although single-core/many-core chips supporting RVV0.7.1 have been released, and the software ecosystem, such as the compilation tool chain, is gradually realizing support for Vector, it is a pity that it has not been accelerated . Optimization of computational libraries . In response to this current situation, Pengfeng Technology will benchmark against excellent international cases and continue to optimize the performance of processors of various architectures such as RISC-V, X86, and ARM, enable advanced computing hardware, and empower scientific research innovation and industry applications.

Thanks to the organizers and sponsors for their strong support, and to all audiences online and offline for their participation and interaction. Let us look forward to next year's RISC-V China Summit and witness the development and growth of RISC-V in China!

Guess you like

Origin blog.csdn.net/weixin_45571628/article/details/132543240