Imagination Technologies Blog

We've posted a follow-up article where we analyze further optimizations and compare memory traffic and rendering speed for ray traced shadows vs. cascaded shadow maps: http://blog.imgtec.com/multimedia/ray-traced-shadows-vs-cascaded-shado…

Glad to see Imagination is finally putting OmniShield to good use. http://blog.imgtec.com/powervr/omnishield-multi-domain-security-connec…

It's described in the link: http://blog.imgtec.com/powervr/gnomes-per-second-in-vulkan-and-opengl-… the code doesn't use any batching but each gnome is a unique draw call, which is a worst-case scenario for OpenGLES2.

The article linked in the first paragraph provides some good context: http://blog.imgtec.com/powervr/a-look-at-the-powervr-graphics-architec… Graphics developers need to know this stuff because certain operations can be very, very expensive on tile-based deferred renderers. Things like reading from the framebuffer, for example. On an immediate mode renderer, the GPU just has to queue up a DMA. On a tile-based deferred renderer, the previous commands have to be flushed first. And these differences are part of the motivation for new APIs like Metal and Vulkan.

That's a good point, but I wonder if coming up from below with a cleaner architecture might be able to accomplish something similar. Is MIPS64 cleaner than X86, and could this let them create lower power higher performance chips or manufacture equivalent performance chips at a lower cost? http://blog.imgtec.com/mips-processors/loongson-mips64-processors-perf…

Interesting benchmark results: http://blog.imgtec.com/wp-content/uploads/2015/09/64-bit-CPU-performan… • Both MIPS and ARM have surpassed AMD in the low-power SoC area (the only where AMD is even remotely competitive right now), and are comparable to their desktop/server products • A (2? 3?) years old Intel architecture still beats all three by a wide margin. And the recent Skylakes are again some 10% more efficient.

The GPU in question here, the PowerVR G6430, does not support OpenGL 4.5. It tops out at OpenGL 3.2 since it's a mobile part so everyone is using GLES. The blog post only somewhat touches on it but I believe a key part of why this is so much faster is that it sounds like Vulkan is actually exposing tiled-based rendering. Note that PowerVR is a tile-based deferred renderer under the hood[1]. So if the app is doing the work of chunking draws into tiles instead of the driver trying to extract that from the OpenGL stream that's going to be significantly more efficient. 1: http://blog.imgtec.com/powervr/a-look-at-the-powervr-graphics-architec…

Their new GPU comes with something called an "RTU", which accelerates the ray intersections. This article is showing off one way to use their fancy new toy: http://blog.imgtec.com/powervr-developers/powervr-gr6500-ray-tracing

Fun fact: the new Horizons's (redundant) processors are 12MHz MIPS R3000s. http://blog.imgtec.com/mips-processors/mips-goes-to-pluto I am really curious what it uses for storage. I am guessing something much more modern, with extensive redundancy.

Maybe there isn't a spec yet, but it seems there are demos running through vulkan: http://blog.imgtec.com/powervr/trying-out-the-new-vulkan-graphics-api-…

Blog post here: http://blog.imgtec.com/mips-processors/meet-mips-i6400-warrior-cpu-for…

"life-like reflections"[1]? allow me to disagree. the reflection is too green. probably not the hardware's fault, or is it? [1] http://blog.imgtec.com/wp-content/uploads/2014/03/5_-PowerVR-Ray-Traci…

Good question about the precision; we only hear little snippets about hitting virtual windows in space within 100km or so. New Horizons is running a MIPS-1 R3000 processor[0], that has double-precision floating point available but presumably fixed-point arithmetic could be more suitable here. But I don't know that there would be so much accumulated error, there were a few course corrections between Earth and Pluto. Maybe the precision is not so much of an issue? As for the fly-by, well. Speed is a problem. At the point of closest approach to Pluto it was going at 13.78 km/s, and it's about the mass of a grand piano[1] so that's about 45TJ of kinetic energy. Or some large number anyway. Let's just carry on with Newtonian stuff and assume the numbers get worse if you do it properly. At current speed its orbital radius[2] would be about ... 4600 metres. Three miles! That's somewhere deep within the icy core, I'm guessing. In order to orbit at the distance of the photos we're getting now, it would need to slow right down to 264m/s, shedding basically all of that kinetic energy. Back of the envelope numbers suggest that would require about the equivalent of a 90 megaton explosion! Of course, if we strapped the Tsar Bomba[3] onto New Horizons, there would be more mass, so we'd need more stopping power ... and so on. Other reasons for doing a flyby, I suspect, is that we get to see more stuff this way. There are a few moons, and an uncharted Kuiper Belt. Seems a…