Gpu with shader model 3.0 series#

Yes, this is actually a different CAP, but with a high overall instruction limit you allow the compiler more possibility for unrolling/optimizing loops.Д¬onsider GPGPU - There are some pretty cool things being done in this field that don't necessarily require game performance.

The shader requries over 2000 slots in total, even though at most 1/2 of those will actually be exectuted.Д¬onsider looping - if I dynamically iterate over something N times, but each loop has 100 instructions, you'll end up with a *lot* more instructions executed. Say I have a simple if/else block where each branch has 1000 instructions. Long shaders might show artifacts on FP24 hardware, especially when manipulating texture coordinates.Д¬lick to expand.That limit is probably not practical, but it is also not as simple as you might think.Д¬onsider dynamic branching - If I create some super huge shader that handles everything, but is dynamically branched, the shader still has to contain all the instructions even if it doesn't execute them. PS3.0 guarantees at least FP32 precision for the basic arithmetic operations. The longer a shader, the more likely do errors become because of insufficient precision. Radeon X1xxx do, but they are less efficient when it comes to branching in the vertex shader. GeForce6/7 in many cases do not meet this requirement in the case of pixel shaders. These not only require GPUs that merely support branching (PS3.0, VS2_a), but they have to support it efficiently for the granularity needed. When you have loops with a low average but high peak number of iterations, or a common short branch and a rare long one, unrolling those might be prohibitive. Shaders that need dynamic flow control to run efficiently. Those are almost impossible to emulate efficiently, and since NV3x is often too slow, these shaders really need a PS3.0 card. Shaders that take advantage of the gradient instructions that are available with PS2_a (GeForce FX) and PS3.0. Thanks in advance.Д¬lick to expand.There are different kinds of complex shaders that can take advantage of SM3.0:

Gpu with shader model 3.0 series#

If the Shader Model - 3.0 isn't really necessary then why Nvidia introduced it on GeForce6 series boards almost year back?ДЄnd if it was necessary then why ATI didn't made SM3.0 compliant hardware?

Also, if all shader 3.0 effects can be displayed by shader 2.0b hardware then is the X800 series still a good buy?, considering Nvidia with GeForce6 series already supports Shader Model 3.GS appears faster than X800XL with low price and supports Shader Model 3.0, and games are now taking complete advantage of Shader Model-3.0 Is the Shader Model 3.0 significant better than Shader Model 2.0b as far as Image quality is concerned and gaming is concerned?, from what I gather that, long shader programs effects (10-20 Instructions long or even more) can be displayed by Shader Model 2.0/2.0b hardware, but performance may decrease. I have a question, and I cannot able to find precise information I want over web, the question is as below: