berkut

51 ms·

511.
by berkut 10y ago
The problem with subscriptions is that it allows the app maker to potentially not maintain and improve their app while still earning money. I've been employed by companies (who I've now left) where the management explicitly want
512.
by berkut 10y ago
But that doesn't really scale - unless the per-thread stuff is small (and then you'd have to page on a sub-mesh level - i.e. split up the mesh spatially and page it as you traverse through it for internal scattering or something),
513.
by berkut 10y ago
Why would you need per-thread geometry caches? That doesn't really make sense (memory duplication)... What you'd do is just map the geometry + bvh into a serialisable format (like vraymesh) such that they can be fread() in one go
514.
by berkut 10y ago
24 hours (wall clock, not CPU time) is a little long, but it's not unheard of. With MC path tracing, the noise you get is pretty difficult to resolve, and a lot of companies are using denoisers afterwards to clean up the renders, as it
515.
by berkut 10y ago
By "geomapping" I assume you mean paging geo? Easiest way is to have a two-level BVH, where the lower level just has overall AABBs of every object in the scene which contains a "geometryInstance", each of which then cont
516.
by berkut 10y ago
That's somewhat true for VFX in general, but not really for Pixar: They mostly use procedural textures (they use them a lot more than most other VFX studios can) so texture IO is a lot less, and is one of the reasons they can get awa
517.
by berkut 11y ago
Maya, Mari, Modo, Nuke use C++ Qt, Katana uses PyQt, other apps like Houdini use their own solution. And before someone says "that means they're not native" - well obviously, and anyway, they have their own theme which is qui
518.
by berkut 11y ago
Pixar and Disney mostly use Linux (as do most of the VFX / CG studios). Artists can use Windows / OS X in some cases, but given the main apps they use (Maya, Mari, Modo, Katana, Nuke) all run on Linux, there's no point real
519.
by berkut 11y ago
"easy to compile anywhere" Doesn't this imply there needs to be the Rust compiler / Cargo (if it needs that) installed on the system? And what versions - stable / unstable ?
520.
by berkut 11y ago
Phantom's can get quite high: http://www.dnaindia.com/scitech/report-watch-a-stupid-drone-...
521.
by berkut 11y ago
As someone trying to learn/experiment with Rust, the syntax is pretty complex. Now granted it comes with some benefits, and I realise there are a limited number of characters available to use (at least that everyone in the world has
522.
by berkut 11y ago
Most independent people who need 3D performance for CAD might be using Windows, but the VFX industry need 3D performance on the desktop and they heavily use Linux and NVIDIA (at least at the high level). AMD is almost nowhere to be found -
523.
by berkut 11y ago
But in practice it's quite often the case that you know optimisations are worth it based on past experience. For example, with floating point calculations, divides are slower (on x86 anyway) than multiplies, so quite a common one in gr
524.
by berkut 11y ago
Those results look slightly suspicious to me, in that I've seen TBB concurrent_hash_map scale much better than that... I guess it depends on the workload ratio... Would have been nice to see a binned/sharded hashmap in the results
525.
by berkut 11y ago
std::map is a RB tree, not a hash map.
526.
by berkut 11y ago
The tiling concept is similar, but the overall algorithm is a bit different, in that with REYES micropolygons are shaded first before being visibility tested: that's why REYES is so amazing at doing displacement. It does means however
527.
by berkut 11y ago
It does. I benchmarked compilers for some of my stuff just over a year ago (GCC, LLVM, ICC): http://imagine-rt.blogspot.co.nz/2014/12/c-compiler-benchmar... and surprisingly I found ICC was no longer always the fa
528.
by berkut 11y ago
32 years is quite a long time in computing terms! Both the computing tech and raytracing/graphics theory and algorithms have changed significantly over that time, with better algorithms and much more powerful computing resources allowi
529.
by berkut 11y ago
I think register access is technically faster than a L1 hit, but due to store forwarding from the write back cache, this extra latency is usually not visible, so in practice they're the same.
530.
by berkut 11y ago
Maybe, depends on whether you're doing CG animation (Pixar, Dreamworks style) or more realistic stuff - VFX, photoreal stuff. If the former, then maybe a bit. But the beauty of physically-based is that in many cases, you can take a mod
531.
by berkut 11y ago
It won't - at least for general ray intersection itself, the deferred sorting doesn't buy you that much - the main benefit it gives you is much more coherent texture accesses, which Disney needs because they use PTex which needs
532.
by berkut 11y ago
Because in path tracing, when you have many lights, you don't want to be sampling each light each bounce - that just doesn't scale due to the explosion in number of rays you have to send to test occlusion. The whole point of pat
533.
by berkut 11y ago
sigh Why does pretty much all the marketing and examples of "groundbreaking" new/faster tech for raytracing/pathtracing always just show outside IBL-lit scenes? These are easy to resolve (generally within 32 progress
534.
by berkut 11y ago
It's only really SCons which gives Python as a build system a bad name in terms of speed in my experience.
535.
by berkut 11y ago
Tagged pointers - hiding a few bytes in a pointer address.
536.
by berkut 11y ago
You don't need to these days (since 2011) - there have been numerous benchmarks over the years showing ICC generate binaries that perform better than MSVC++ for AMD processors.
537.
by berkut 11y ago
This can run native x86 code though (although to take advantage of the architecture code might have to be re-written to some extent). So this means you can run the same code on a desktop Xeon as this one.
538.
by berkut 11y ago
Cool - I'm working on Manuka at Weta atm so the logistics might be a bit difficult :) Yeah, there's a bit of c++11 going on here - but until all the DCC apps move next year, as you say, it's going to be a while.
539.
by berkut 11y ago
Really? So you can completely control the layout of where items on the heap are allocated, get hold of pointer addresses, and store things within the least significant bits (for tagged pointers)?
540.
by berkut 11y ago
And VFX software! :) Not even fully on C++11 yet! Vil: How's it going?

More ›