Hi, Did you even look or try the stuff you posted, or as normally spam with unrelated things? ;)
Quoting Juhana Sadeharju <[EMAIL PROTECTED]>: > Hello. CS had a software renderer. Is it still in CS? What is > its quality and does it support latest shaders? Yes it does, but no, the quality isn't comparable to hardware rendering and no it doesn't support programmable shaders (and that is not just because they would be terribly slow;). The software renderer is just there as a debugging aid (valgrind for example does not always play nice with opengl drivers) and for cases where you just cannot use opengl but still want some output. > > I don't have up-to-date graphics card, but would like to > test latest rendering tricks. Sorry, thing you are out of luck there. There _are_ no software renderers out there that supports modern shader programs and at the same time get you anything close to realtime performance. The closest you get is Pixomatic from RadGameTools, which will cost you some $10k or so (and then you get Dx7 level features). > > WildMagic4 has a software renderer with shaders. Perhaps it > could be used as an up-to-date software renderer in CS. > http://www.geometrictools.com > WM4 is released as LGPL software. Previous versions are and the > next version will be "viewable source" licensed due user's > wishes (perhaps they are not willing to publish their modifications > as open source as required by LGPL?!). The WM4 software renderer, while seeming a good thing is incredibly slow, have non-fixed fill convention (you will see triangle gaps) and is not a good choice for _improving_ CS software renderer. An example of the speed you get is the DefaultShader test, which is a flat-shaded model of a pretty low polygon model. The WM4 software renderer will get ~10fps on my computer. Also the Portals test-case gives sub-one fps, while the much more complex "flarge" level in CS gives tenths of fps (on same computer). Notable is also that even though WM4 support "software" shaders these are written in c++, have to be compiled into the project and are much more limited than what you get from using Cg on modern hardware. > > Author of WildMagic4 is one of original developers of > NetImmerse/Gamebryo engine. I'm trying to compile WM4 because > it may be more suitable for rendering the NIF files, but also > because Mesa3D seems to not provide software shaders at all. As said above there is a reason Mesa3D doesn't do this :) > > There is an open source GPU emulator at > http://personals.ac.upc.edu/vmoya/ > Someone expert should take a look at it if it could be used > as a part of software renderer in CS. If anyone compiles it > and gets it working, please let me know. That is an emulator for a proposed GPU architecture, not any existing. Also it is not realtime, you take a opengl recording log and give to it and a few minutes later it spits out lots of statistics about how this proposed GPU architecture would work. Not really something useful for a realtime 3d engine :) > > As mentioned here earlier, Mesa3D was/is too slow for real-time > game graphics. Would WM4 be faster? Would Vmoya be faster or > would it require Mesa3D on top of it (as if Vmoya emulates the > GPU hardware only)? If you had spent at least an hour on investigating your proposals you would have been able to answer that by yourself and not wasted everyones time. (The answer to all is no). - M ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Crystal-main mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/crystal-main Unsubscribe: mailto:[EMAIL PROTECTED]
