Search results

Is it dot product? I thought FTRV was the big one, which is 4x4 matrix multiplication. According to the docs I've read it has a 4 cycle cost. Is there another instruction that I've missed? From my readings the equivalent NEON code sequence would be 8 cycles, which means it would be covered by a...

Unfortunately it's not *quite* this simple (is anything)? The X86 has a mix of instructions, some of which will execute in a single cycle (simple ops between two registers, for example), and some of which are more complex (for example, incrementing the value in a memory location). However...

The question I have (that I haven't yet seen an answer to) is - how delayed is it all going to be? So the original estimate was end of November, they would start sending out Pandoras. With delays, obviously this is no longer really plausible. Indeed, we will probably end up with orders being...

It can be pipelined for (some) operations if you switch on various flags. Although it doesn't have forwarding in these cases.

I found it somewhere on beyond3d.com - they have some interesting forums. Not much about TBDR, but there is a little, and there are a couple of PowerVR people chatting on that board. They do some pretty cool stuff, I reckon it would be an amazing company to work for.

Apparently PowerVR have a patent on something related to compressing the tiles (basically performing a partial Z elimination) before writing them out. Assuming they have fairly efficient local buffers, I assume that they compress the buffers before they have to spill them, so that odds are the...

I reckon it might just be counted as a 4x4 matrix transform - and if you're using triangle strips, do you only need to translate one point? so the first triangle in a polygon would be 3 vertices, but any subsequent triangles would be just one point. My memory of all this stuff is very sketchy...

Surely with such a limited speed it would have some issues doing vertex geometry? For example, just translating a single point to the screen would cost 16 multiply-adds (assuming a 4x4 matrix and a 4 element vector). Or does it have custom hardware to perform transformation? I mean, assuming...

I dunno, I'm sure I read somewhere that the Cortex A-8 has a brand new 'jazelle' - which instead of executing bytecode directly instead is designed to support jitters better by adding support for null pointer checking etc. Indeed, 'standard' jazelle seems to be particularly crippled...

Cool, what did you use for the recompilation? Is it a custom thing for Albion, or would it work on other games?

This has the potential (that will probably be screwed up :() to be fantastic Consider, it is the perfect Myth style box. USB2 hard drive, USB2 wifi network, and you have yourself a mini opensource PVR that makes *NO* noise. No fan noise, no noise period. It can even do encoding if need be. Not...

On 13.4.2.4 (phew) of the MP2520F manual thing on the wiki, it seems to indicate that it supports 5 layers (or blocks of RGB data that can be overlayed, so not just a YUV plane). It looks like you can then layer this over a YUV plane if you so choose as well. It's not exactly obvious however...

Hrm, they don't make it easy do they! From the documentation, I thought that it handled it in 8bit paletted and 15bit mode. But I only skimmed it briefly. Do you need more than 15 bit mode for the SNES?

I skimmed through the MagicEyes documentation, and it looks like it can do multiple surfaces together, including alpha-blending. Is this any use at all, or is it too general? I was hoping it might actually do almost all of it, until I think you mentioned that people could change the pallette...