Recent content by arjan de lumens

Each of the four quad lanes in an execution engine has an FMA plus a separate ADD, so in terms of flop numbers, you should get that: G71 at 1000 MHz should be able to do 32*12*(2+1)*1000M = 1152 GFlops/s. While this does fall a bit short of XB1/PS4, it should generally be a bit easier to get...

This kind of efficient FSAA functionality (usually in the form of multisampling more than supersampling) is something that's still widely present in modern 3d hardware, however, modern rendering techniques tend to do a lot of screenspace post-processing, in ways that interact with AA in such a...

The IEEE-754-2008 standard defines a few decimal floating-point formats (32-bit, 64-bit and 128-bit) that generally use an encoding that packs three decimal digits into 10 bits. This has been implemented in some processors, such as IBM's POWER processors from POWER6 and up; its main use case is...

Result for Mediatek's MT6755, using a Mali-T860, apparently an MP2: 2.5 fps for the Car Chase, 8.1 fps for the tessellation test (both offscreen). Like all previous Malis, Mali-T860 does not have any tessellation HW; it's all compute shaders.

This seems like a fairly standard floating-point issue, quite similar to classical floating-point annoyances like "why does 6 times 0.1 not equal 0.6?". Values on the form N/65535 are generally not exactly representable in floating-point, which means that every time you sample such a value, you...

To expand on mczak's point with a few numbers: The Galaxy S6 has a screen resolution of 2560x1440 (~3.7 million pixels), so its sustained performance on that T-rex test (which, as the anandtech article pointed out just below that graph, is an on-screen test), measured in pixels per second, is...

You can kill a shader like that in the middle just fine, as long as you don't actually kill it inbetween side-effects.

Oh. OK. That clearly solves the issue I had in mind, at least for GLES3.1 and Vulkan. Thanks.

Ummm, the impossibility of avoiding this kind of virtual-texture overfetch is a consequence of how the feature is defined at the API level, not specifics of their hardware design. For immediate-mode renderers like those of Nvidia and AMD, the only harm done is that you fetch somewhat bigger...

Well, yeah; the point here is that the APIs for virtual texturing (AMD's mechanism, DX11.2, GL4.x etc) are being defined in such a way that it is not possible to avoid doing this for overdrawn geometry.

I don't know what rpg.314 is referring to either, but I can think of at least one thing that can break both the overdraw removal and the submission order independence of TBDRs quite badly: Virtual Texturing. Especially if the form introduced by AMD...

Well, yes; they admittedly get rather visible when flipping back and forth at 8x magnification. Cutting bitrate in half was probably a bit overoptimistic of me for this kind of comparison; pushing bitrate that hard forces the compressor to make some rather hard choices between...

It seems to me that it's possible to push the ASTC bitrate really quite far down before you cross over from "they just have different artifacts" to "one is clearly worse than the other". Here are your three images, except that I cut the bitrate for the ASTC image in half: I have replaced the...

Let's see. Petri Nordlund has become a development director at Unity Technologies. The Tuomi brothers, on the other hand, have formed a startup "Siru Innovations Oy", which according to their website http://siru.fi/ is a "leading Finnish graphics IP vendor", with very little information beyond...

OK, but is that really a battle you would want to avoid? In particular, it would seem to me that no matter how large input images you use, this particular battle will just start up again and re-diverge the encoders as soon as you start constructing a mipmap pyramid? This seems like an argument...