videocoreiv-qpu

Shader units --- e.g. the VideoCore's QPUs, which this project is targeting --- are very very weird. Compiling C into them basically isn't possible. They don't really do memory accesses, for example. The VideoCore's own processor, commonly referred to as the VPU or just the VC4, is a perfectly normal architecture with its own vector unit and DSP addons. C can compile into that perfectly easily, and the real-time OS that runs on it alongside Linux is written in it, but there aren't any adequate open source C compilers yet. More info on the QPU here: https://github.com/hermanhermitage/videocoreiv-qpu

Don't forget, also, that the ARM11 in the BCM2835 is actually just a secondary processor; the main CPU is a blisteringly fast dual-core Videocore 4, which lots of interesting DSPish and vector processing instructions. (You can call out to it from the ARM.) See: http://maazl.de/project/vc4asm/doc/index.html And if you're into shader programming, the GPU has a 24 GFLOP shader unit. https://github.com/hermanhermitage/videocoreiv-qpu Only suitable for very specialised purposes, but astonishingly fast.

For those interested, as far as I understand it this release doesn't include VPU or QPU assembler tools. We released a QPU assembler at https://github.com/hermanhermitage/videocoreiv-qpu/blob/mast... a couple of weeks ago. And a reference javascript VPU assembler is due very soon (mirroring the path taken with the QPU assembler). Because our own work was independent and pre-dates this information by a couple of years (in the case of the VPU), we have some alignment work to do in terms of mnemonics. This release adds a lot of knowledge to the table - extending our understanding beyond the core instruction sets (which were already well reverse engineered) into a deeper understanding of the various hardware registers.

Apologies if its bad form to post a reference to my own github. The VPU documentation is now at stage where enough instructions are described to do useful work. I am about to begin writing up a set of tutorials for beginners who want to leverage the VPU. We are looking to push forward with a robust port of GNU Binutils for VideoCore IV, and seeking contributors. Similarly we are looking for contributors to help with ports of GCC and LLVM. Some Background: The SoC in the Raspberry Pi has 3 processors / instruction sets - (1) An ARM v6 @700MHz for userland, (2) a VideoCore IV VPU @250MHz with dual core, dual issue and 16-way SIMD integer vector processor for running the blob, codecs, 2d acceleration etc, and (3) a 24 GFLOPS shader processor (QPUs) for OpenGL ES and OpenVG. Understanding the QPUs is also under way at https://github.com/herma