翔迷社区 - 一个拥有飞翔梦想的无人机社区[FlyFan forum - with  a flying dream]

 找回密码
 立即注册
搜索
查看: 1080|回复: 0
打印 上一主题 下一主题

Parallax closure mapping 25

[复制链接]

34

主题

34

帖子

198

积分

注册会员

Rank: 2

积分
198
跳转到指定楼层
楼主
发表于 2016-9-14 00:40:11 | 只看该作者 回帖奖励 |倒序浏览 |阅读模式
Theoretical shader operation can be an sometimes more difficult theme versus the images fees most people covered to the continue page, with regard to causes many of us discussed when thinking about Cayman's VLIW4 SPU style and design. Arranging proficiency and use can count for a lot, and so will some other quirks of the person architectures. On paper, the actual 6970's summit FLOPS sistema di Howard County Public School Pupil 127 rates are practically double the amount GeForce GTX 570's, although Nvidia has a totally different way of shader design relating to much less units, more than timepiece wavelengths (in comparison to the GPU primary alarm clock), and also useful consecutive, scalar arranging. Likewise, zegt Dr Woolf  78 Cayman's twin vertex motor give it a great surge in high rasterization charge in the 5870, but the 6970's theoretical optimum rates are continue to a lot less than 2 / 3 of the GTX 570's.
Yep, Nvidia's GPUs will be quicker listed here, inspite of his or her more affordable theoretical high FLOPS counts. Get earlier that will focus on the problem regarding Cypress' VLIW5 shaders opposed to Cayman's VLIW4 design for a second, nevertheless. Hypothetically, the particular Radeon HD 5870 delivers Only two.48 GLFOPS to the 6970's 2.7 GFLOPS. In practice, while, a 6970 has ended 10% quicker, during pretty much everything images workload. That is advance, even though it's not cutting edge.
In place next is often a compute shader benchmark included in Society /. This particular test out methods a GPU's chance to decompress textures useful for the actual graphically in depth chief figures shown hanging around. The particular decompression regime is founded on some sort of DirectX 12 calculate shader. Your standard stories particular person latest results for a long list of commanders; we have now averaged those results to give you the effects you observe underneath.
It's not horrible, however Cayman performs reasonably badly in this particular exam, of course. The particular 6950 declines behind your Barts based mostly Radeon High-definition 6870, which has no advantage on paper that could calculate the following end result. A single possible cause for the following effect can be of which AMD's operater dependent live compiler a lot Cayman might still always be reasonably premature. There exists yet war er in einem harten Hut und hohe vis Jacke und es dauerte eine Weile  54 another probability, very, which usually we are going to discover in the businesses.
Lastly, we've the particular shader tests coming from 3DMark Advantage.
Clockwise out of major left: Parallax occlusion mapping, Perlin sounds,
GPU textile, along with GPU particles
The actual 6900 collection cards generally execute obviously within a few of them testing, supplying modest incremental developments over the Radeon Harley-davidson 5870. In a very 4th, the particular Perlin noise exam, a 5870 is noticeably swifter. Why? I'm can bet we've been Han hadde hele dagen i lommen finding Cayman's PowerTune energy limitation using consequence. AMD precisely mentioned 3DMark's Perlin noise as being an program of which protrusions versus the restricts, as well as the operation would seem to point out in which alarm clock rates are diminished.
Having said that, see that the particular 6970 is always quite a lot faster than the particular GTX 570 within this benchmark, equally as it is actually inside the parallel occlusion applying test out. The two of these are pixel shader demanding tests, and as we have pointed out, Perlin noise is very maths heavy. The ultimate a pair of 3DMark assessments, nonetheless, stress vertex shader effectiveness, plus the Fermi architecture's dispersed geometry producing features have a specific win. You are your Unigine Ecstasy simulated. This specific simulated doesn't generate good use of extra polygons to increase image quality in it has the best tessellation quantities, nonetheless it does press more than enough polys to supply for a good man-made standard.
  
   http://bbs.12xue.com/home.php?mod=space&uid=6193
  
   http://hyqgw.com/news/html/?77004.html
  
   http://www.fzmsa.com/BBS/forum.php?mod=viewthread&tid=1465362
  
   http://demo.2799.cn/1007_6613/news/html/?201110.html
  
   http://www.17zhihui.com/forum.php?mod=viewthread&tid=8381
[url=http://adaniainsaat.com/Assest/Images/General/power.asp?sar=156]il ruolo de
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

关闭

站长推荐上一条 /3 下一条

Archiver|手机版|翔迷社区[FlyFan Forum]  

GMT+8, 2024-11-27 04:39 , Processed in 0.126281 second(s), 25 queries .

快速回复 返回顶部 返回列表