Lightning Talk: Creating a GPU With C++ and an FPGA
summary
Concrete claim: a working VGA-capable GPU implemented on an rta7 FPGA development board from about less than 100 lines of C++ using Xilinx HLS to Verilog. The process shows pipelined inner loops, bit-banged-style outputs for HS/VS/R/G/B, texture lookup via large RGB arrays, and barycentric triangle rasterization; the result is a tiled sky texture and a red triangle displayed over VGA at the target clock frequency.
AI-written summary · about
chapters (3)
related talks
01h 02mTurning software into computer chips – Hastlayer
18mgVulkan: Scalable GPU Pooling for Pixel-Grained Rendering in Ray Tracing
45mAutomated, Simulation-Based Flow for Low-Cost FPGA-Accelerated Devices with Zeph...
01h 00mGPU Accelerated Computing & Optimizations on Cross-Vendor Graphics Cards with Vulkan & Kompute
23mHello GPU
50mUnder the HAT: Empowering GPU Acceleration for Java
09mLightning Talk: RISC-V and AI: How an Open-Standards ISA Underpins Portability f...
16mBuild your own Open Hardware CPU in 25 minutes or less