All results must be verified on at least one of the following RISC-V hardware environments, except the results of challenge S2312, which should be verified on SG2002:
If required, please contact our staff by email: rvspoc@cyberlimes.cn
Summary:Natural voice interaction is an important form of human-machine interfacing. However, as high-precision ASR models are computationally intensive, these computation projects are usually run on the cloud, worsening the experience. This challenge aims to port high-precision streaming ASR to Sophgo’s SG2002 processor. You may choose between Chinese and English models. The goal is to achieve minimum word error rate whilst running on limited RAM (256MBytes) and a rate of RTF<1. You may reference Kaldi, Wenet, and other open source speech recognition projects for your port.
Summary:Port Xv6 to Milk-V Duo, with the ability to interact via serial connection, support for basic peripherals, and the ability to compile sample system and peripheral applications.
Summary:Port Zephyr to Milk-V Duo, with the ability to interact via serial connection, support for basic peripherals, and the ability to compile sample system and peripheral applications.
Summary:Port OpenGauss to SG2042/Pioneer Box (openEuler recommended, but any Linux distribution is allowed), optimise benchmark results as much as possible.
Summary:Using an SG2042/Pioneer Box, run x86 or ARM64 applications using Qemu-like emulators or translators. You may need to implement support for runing Windows applications via Wine. The goal of this challenge is to optimise binary translation performance as much as possible.
Summary:Port OpenHarmony to SG2042/Pioneer Box with its graphical environment, along with the ability to browse the Web and perform productivity tasks.
Summary:It is a feat to get Baby LLaMA 2 running on small development boards like the Milk-V Duo. This challenge aims to improve Baby LLaMA 2 performance on Milk-V Duo, as assessed by tokens processed per second. The participant will make use of lightweight techniques and compiler optimization principles to develop a story-telling robot with support for microphone and command-line input. The demo should output robotic voice via speakers.
Summary:Run Linux distribution with GNOME/KDE desktops, OpenGL 3D benchmarks, video playback, and 2D/3D games on a combination of SG2042/Pioneer Box + an arbitrary (host sanctioned) Chinese domestic GPU.
Summary:ncnn is a high-performance neural network compute framework with good RISC-V support. This challenge assesses the participant’s ability to maximise ncnn performance on a Duo, as benchmarked by its internal performance testsuites.
Summary:Implement bugfixes and optimization on upstream Node.js for RISC-V, as well as platform-specific optimizations for SG2042/Pioneer Box, optimising benchmark results as much as possible.
Summary:Improve OpenCV performance on Milk-V Duo. The participant must port OpenCV to the Milk-V Duo platform, optimising its performance. The assessment rubrics will focus on both accuracy and performance: accuracy will be benchmarked using OpenCV’s regression testsuites (build/bin/opencv_test_); performance will be assessed using OpenCV’s performance testsuites (build/bin/opencv_perf_). The host will judge based on both the pass rates of regression testsuites and performance metrics and assign scores to each submission based on a predetermined weighted scale.
Summary:ROCm is an open source compute framework from AMD, which is an NVIDIA’s CUDA competitor. This challenge asks the participant to port ROCm to an arbitrary SG2042/Pioneer Box + AMD GPU platform, making it possible to perform high-performance heterogenous computing on RISC-V hardware with a combination of open source hardware and software.