. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.5 Graphics Hardware Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.6 網路連接硬體 . . . . . Braille Displays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.8 Hardware Speech Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.9 周邊設備與其他硬體 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.3.3.1 Testing hardware compatibility with a Live-System . . . . . . . . . . . . . 14 3.3.4 網路設置 . . . . . . . . . . .

build a unified and open OS that supports multiple processor architectures, and to advance the hardware and software application ecosystem. On December 31, 2019, the openEuler open source OS community machine, check that the physical machine meets the compatibility and minimum hardware requirements. For details about hardware compatibility, see the openEuler compatibility list at https://openeuler.o Raspberry PI. 21. RISC-V: The RISC-V platform supports kernel- based virtual machines (KVMs). 22. Hardware: Hi1822 iNICs are supported. 12 13 openEuler WHITE PAPER openEuler 21.03 Technical White Paper

switch to another software after a period of use. In that case, they need to redeploy such as hardware and data, resulting in a high transfer cost, so when users select the use of closed source software Pitchbook, Yunqi Partners 附录 1 开源硬件 v.s. 开源软件 Appendix 1 Open Source Hardware v.s. Open Source Software 开源硬件，即 Open Source Hardware，是可以通过公开渠道获得的硬件设计，如电路图、 材料清单和开发板布局数据。并且硬件设计的源代码的特定的格式可以为其他人获得，以方便 Open Source Hardware is hardware designs such as circuit diagrams, BOMs, and development board layout data available through public channels. Furthermore, the source code of a hardware design is available

infrastructure scenarios. It has a rich history of helping enterprises develop their software, hardware, and applications. The openEuler open source community was officially established on December 31 install openEuler on a physical machine, check that the physical machine meets the compatibility and hardware requirements. For a full list, visit https://openeuler.org/en/compatibility/. VMs openEuler the compatibility and hardware requirements. Embedded Devices To install openEuler on an embedded device, check that the embedded device meets the compatibility and minimum hardware requirements. openEuler

storage 选项卡中，自定义存储选项。 f. 在 Select clone options 中，选择 Customize this virtual machine’s hardware。 g. 在 Customize hardware 选项卡中，点击 VM Options → Advanced。 可选：覆盖 vSphere 中的默认 DHCP 网络。启用静态 IP 网络： i. 设置静态 IP guestinfo.ignition.config.data.encoding：指定 base64。 disk.EnableUUID：指定 TRUE。 h. 在 Customize hardware 选项卡的 Virtual Hardware 面板中，根据需要修改指定的值。确保 RAM、CPU 和磁盘存储的数量满足机器类型的最低要求。 i. 完成配置并打开虚拟机电源。 $ export IPCFG="i storage 选项卡中，自定义存储选项。 f. 在 Select clone options 中，选择 Customize this virtual machine’s hardware。 g. 在 Customize hardware 选项卡中，点击 VM Options → Advanced。 从 Latency Sensitivity 列表中选择 High。 点击 Edit Configuration，然后在

并行计算语言（函数式编程）：Scala Ø 加速方法： Ø 算法的优化 Ø 算法的并行化 Ø CPU: 多核 CPU Ø GPU： 多核处理器 Ø 硬件仿真：算法计算在FPGA里实现，输入和输出在 PC 端实现。 Ø Hardware in the loop simulation 加速计算 (Accelerate Computing) 多核处理器系统 8 Ø HDL：硬件描述语言 (Verilog, VHDL)，通过与或非 Zynq or ZynqMP PetaLinux supported platform 13 • Programmable logic circuits are presented as hardware libraries called overlays. These overlays are analogous to software libraries. A software engineer configurable and re-used as often as possible in many different applications. 解释： Ø PYNQ：软件框架 Ø Hardware library: overlay，硬件IP 库 Ø Software library: 软件库 Ø 通过API 调用 Ø Overlay 怎么产生：需要FPGA 工程师设计 IP Ø

并行计算语言（函数式编程）：Scala Ø 加速方法： Ø 算法的优化 Ø 算法的并行化 Ø CPU: 多核 CPU Ø GPU： 多核处理器 Ø 硬件仿真：算法计算在FPGA里实现，输入和输出在 PC 端实现。 Ø Hardware in the loop simulation 加速计算 (Accelerate Computing) 多核处理器系统 8 Ø HDL：硬件描述语言 (Verilog, VHDL)，通过与或非 Zynq or ZynqMP PetaLinux supported platform 13 • Programmable logic circuits are presented as hardware libraries called overlays. These overlays are analogous to software libraries. A software engineer configurable and re-used as often as possible in many different applications. 解释： Ø PYNQ：软件框架 Ø Hardware library: overlay，硬件IP 库 Ø Software library: 软件库 Ø 通过API 调用 Ø Overlay 怎么产生：需要FPGA 工程师设计 IP Ø

并行计算语言（函数式编程）：Scala ➢ 加速方法： ➢ 算法的优化 ➢ 算法的并行化 ➢ CPU: 多核 CPU ➢ GPU： 多核处理器 ➢ 硬件仿真：算法计算在FPGA里实现，输入和输出在 PC 端实现。 ➢ Hardware in the loop simulation 加速计算 (Accelerate Computing) 多核处理器系统 8 ➢ HDL：硬件描述语言 (Verilog, VHDL)，通过与或非 Zynq or ZynqMP PetaLinux supported platform 13 • Programmable logic circuits are presented as hardware libraries called overlays. These overlays are analogous to software libraries. A software engineer configurable and re-used as often as possible in many different applications. 解释： ➢ PYNQ：软件框架 ➢ Hardware library: overlay，硬件IP 库 ➢ Software library: 软件库 ➢ 通过API 调用 ➢ Overlay 怎么产生：需要FPGA 工程师设计 IP ➢

理 MSA (Microservice Architecture)基础架构演进史 Hardware Virtualizati on OS Containers Middlewar e Runtime Data Application s FaaS Function Hardware Virtualizati on OS Containers Middlewar Runtime Data Application s SaaS Function Hardware Virtualizati on OS Containers Middlewar e Runtime Data Application s PaaS Function Hardware Virtualizati on OS Containers Middlewar Runtime Data Application s CaaS Function Hardware Virtualizati on OS Containers Middlewar e Runtime Data Application s IaaS Function Hardware Virtualizati on OS Containers Middlewar

硬件加速全览 Decoder Encoder Other support Internal （AVHWaccel） Standalone Hardware output Standalone Hardware input Filtering Hardware context Usable from ffmpeg CLI CUDA / CUVID / NVENC N Y Y Y Y Y Y 已经在各种Linux 分发版本中支持 • 支持的硬件更加广泛 • 更多的Codec的支持 • VAAPI 作为一个Linux上的Video 硬 件加速接口，同时也可以支持AMD / Nvidia hardware with Mesa. • 与其他标准APIs集成较好 (EGL/OpenGL, OpenCL). • 某些Cases下图像质量更好 • 可能支持的转码路数更多一些 (particularly ffmpeg.org/wiki/HWAccelIntro • FFmpeg VAAPI的使用 • https://trac.ffmpeg.org/wiki/Hardware/VAAPI • https://wiki.libav.org/Hardware/vaapi • Buffer Sharing • https://www.khronos.org/registry/OpenCL/extensio