berkut

57 ms·

601.
by berkut 12y ago
Not necessarily: even if the data's on the GPU so doesn't have to pay the PCI-E transfer penalty, GPUs still have cache hierarchies and these have latencies as well, and they can be worse than for CPUs as the branch-predictors and
602.
by berkut 12y ago
That's only the case for lighting and comp for us. The earlier stages (Assets, texturing) need internet for reference :)
603.
by berkut 12y ago
This has happened at plenty of places! Ever worked on a Marvel film? They're insane about security - only certain parts of the studio are actually allowed access to that show's dailies, assets, etc.
604.
by berkut 12y ago
I guess the app would have to handle not being allowed permission to do these things gracefully. That's possibly a tall order based on my experience with apps these days, although for a difference reason: a lot of them assume there&#x2
605.
by berkut 12y ago
Movie would have been rendered at 2k ( maybe 4k but that's still rare). GPU's still don't cut it for high-end VFX/Animation rendering, as you're always memory / IO bound, e.g. reading in +20GB geometry and &gt
606.
by berkut 12y ago
Yeah, previs (pre-visualisation) is generally always done these days, as it lets the director preview what things "look" like, but that only has very basic geometry, low res textures (if any) and no proper cloth/skinning defo
607.
by berkut 12y ago
It would in theory work - our solution doesn't at the moment.
608.
by berkut 12y ago
You wouldn't need MPI, just have multiple concurrent writes to a tiled EXR. We do it occasionally for stupidly big (>64K) frames. Doesn't work for deep though :)
609.
by berkut 12y ago
Bandwidth allowing of course - that's often the bottleneck when you're rendering production scenes with +300GB of textures :)
610.
by berkut 12y ago
PBRT's an excellent book if you want to play with / build / understand raytracing renderers. Some of the knowledge is not-quite state of the art any more, but everything's still relevant. And there's a third edition
611.
by berkut 12y ago
The radar and weapon control systems are "locked on" to from the aircraft's perspective - I don't really understand what you're talking about... A particular target has been identified, it is being tracked (maybe at
612.
by berkut 12y ago
Or an earlier version of bash (4.1), which I'm assuming (haven't installed any patches within the last month on a centOS6 machine) hasn't got the issue?
613.
by berkut 12y ago
For really simple apps, no. But more complicated ones, with resizable widgets/layouts, selection changes, scroll view positions and widgets which display a lot of data, or layering/parenting of widgets make things a lot more compl
614.
by berkut 12y ago
Isn't using immediate mode going to make implementing more complex widgets like multi-line text-edits with selection and style, and more importantly tree controls with collapsed/expanded persistent state that lazily load children
615.
by berkut 12y ago
Why does what language they write it in affect that? There are apps written in C/C++/Obj-C that can efficiently open/mmap extremely large files. Unless you're talking about more high-level languages possibly having an is
616.
by berkut 12y ago
It's really just a bandwidth issue - VFX studios do this all the time with their renderfarms - textures are the main issue - prod/archviz like Ikea stuff are generally really clean and don't have THAT many textures - whereas
617.
by berkut 12y ago
Yeah, I think the difference is "Physically-based" and "Physically-plausible" :) Although strictly speaking a lot of VFX rendering is still about how much you can fake stuff, although it is a lot more real (as in done in
618.
by berkut 12y ago
Renderers and shaders for one of the biggest VFX companies in the world :)
619.
by berkut 12y ago
+1 - sorry, I do this stuff all day, so am used to the slang :) Fresnel also effects just reflection as well - it's the reason you don't often (depending on the material) see much of a reflection head-on on a shiny surface, but do
620.
by berkut 12y ago
Should be "Next-Gen Game Lighting". This is just Physically-based shading basically, which has been done for the past 4/5 years in VFX. Essentially it's energy-conserving materials with the correct fresnel effect based o
621.
by berkut 12y ago
Trees are static though - they only deform - they don't transform around the scene itself.
622.
by berkut 12y ago
No, this is a gamma correction issue. Software which work in full-float linear (Nuke, for example) don't have this issue, as the RGB values are already in luminance-adjusted values as they're linear. The issue only affects softwar
623.
by berkut 12y ago
This is one of the reasons using things like boost's boost::shared_ptr (which uses atomics for ref counting) can have a huge time overhead in multithreaded code. Often people think atomics are much faster than mutexes for simply checki
624.
by berkut 12y ago
How big are the binaries? ICC inlines a lot more heavily, it's possible (I've seen it before) the code size has bloated, not fitting into the IC... Having said that, most of ICC's general gains over GCC and LLVM are due to th
625.
by berkut 12y ago
On the flipside, using pre-increment introduces a dependency, which means the processor can't do Out-of-order execution on instructions using that value...
626.
by berkut 12y ago
In a traditional render farm that is an issue, as the vast majority of renderfarms for VFX companies are CPU only. renderfarms aren't just using for renderer rendering, they're used for comp rendering, fluid sims (often using up
627.
by berkut 12y ago
Also less than half the price of Arnold and VRay now... It's possible they were hemorrhaging money thanks to the Arnold and VRay competition over the past few years. Looks like ILM are being "coerced" into using PRMan again f
628.
by berkut 12y ago
For doing production VFX shots, unless there's fur/hair or volumetrics involved, you're pretty much IO constrained in terms of pulling in textures and paging them in memory. Often scenes can have > 300GB of textures total.
629.
by berkut 12y ago
As have Pixar for PRMan 19 for some of the BVH build stuff :)
630.
by berkut 12y ago
Good tips, however it's still often beneficial to do reciprocal float division: i.e. float rcpDiv = 1.0f / value; float finalValue = myValue * rcpDiv; In a lot of cases when speed matters. With fpmath=fast, I've seen

More ›