Published The new version of the project VortEx, developing open gpgpu based on architecture of the RISC-V commands, designed for Performing parallel calculations using the OpenCl API and the SIMT (Single Instruction, Multiple Threads). The project can also be used in conducting studies in the field of 3D graphics and in the development of new GPU architectures. Schemes, descriptions of hardware blocks in Verilog, Simulator, Drivers and accompanying design documentation are distributed under the license Apache 2.0.
GPGPU basis is the typical ISA RISC-V, expanded by some additional instructions necessary to support the GPU and flows management functions. At the same time, changes in the architecture of the RISC-V command set are minimized and, if possible, existing vector instructions are used. A similar approach is used in the RV64X project, which also develops an open GPU based on RISC-V technology.
The main features of Vortex:
- Support for 32- and 64-bit architectures of a set of commands RISC-V RV32IMF and RV64IMAFD.
- Number of nuclei, tasks (Warps) and flows.
- Custom, fpu, lsu and sfu for each core.
- Customizer width (Pipeline Issue Width).
- Optional shared memory and cache levels L1, L2 and L3.
- Support for the OpenCl 1.2.
- The possibility of implementation on the basis of FPGA Altera Arria 10, Altera Stratix 10, Xilinx Alveo U50, U250, U280 and Xilinx