On Wednesday 14 August 2002 9:00 pm, Tim Roberts wrote: | On Wed, 14 Aug 2002 18:46:38 +0400, Vadim Plessky wrote: | >On Tuesday 13 August 2002 8:33 pm, Keith Packard wrote: | >| Around 18 o'clock on Aug 13, Vadim Plessky wrote: | >| > * what video drivers offer hardware support for RENDER ext.? // | >| > checked http://www.xfree86.org/4.2.0/RELNOTES2.html and found no | >| > asnwer for this | >| | >| There is some support for Render acceleration in the Matrox driver and | >| I believe the closed-source Nvidia driver has a bit. Mark is currently | >| designing a new XAA architecture which will include significant support | >| for Render. | > | >So, we are (still) limited to Matrox cards and (aprtially) NVIDIA? | >What about ST Microelectronics Kyro II and Intel 815/845 (built-in into | >chipsets)? | | Let us be very clear that we all understand the question you asked. You | did NOT ask which drivers support the RENDER extension. Rather, you
yes, that's correct. | specifically asked which drivers offer HARDWARE support for the RENDER | extension. Those are very different questions. | Right. Let me quote part of e-mail from our off-list discussion ------------------------------------------------------------- > Yep, libart is worth a look for the RENDER people. > Vadim, do you know if RENDER will be/is hardware > accelerated? If not, I dont see much point in RENDER over libart > currently. Just calling some "draw me curves here and here and..." functions via an X-protocol should be much faster than blitting a pixmap via the protocol. Libart would be faster if it used DRI, which it doesnt, or am i wrong? >From a non-performance pov: i think having a standard rendering infrastructure never hurts. It might creep into Qt someday. ------------------------------------------------------------- Point was that if RENDER is not hardware-accelerated, there is no reason to use it, instead of libart. Note that we are speaking here about vector-drawings (SVG, EPS, PostScript), not about font technologies for *general text rendering* (Word Processor, browser) where current AA is used. | Almost all of the current crop of drivers support the RENDER extension. | That means RENDER-based applications will work correctly. However, in all | but a few cases, that support is implemented in high-performance software | simulations. The primitives required by RENDER are rather specific, and | there has not been a lot of time to figure out if the existing hardware | (most of which was designed long before RENDER was conceived) can be | wrapped around those requirements. To my best knowledge, ATI chips are rather advanced, and ST's Kyro is quite advanced, too. // NOTE that I haven't asked about Trident, S3, Realtek, etc. I am not graphics-chip/hardware expert, that's why I ask this question on <render> list. | | The i815 is not a terribly sophisticated engine. I'm not sure it will do | much better than the simulations. And what about 845 with built-in graphics? I have seen reviews that Intel finally got graphics *right* in that chip (after 3 or 4 years of attempts to build something useful :-) P.S. yes, I recognize thta my laptop is already 2 years old. But Linux still works fast on it (Pentium/III-600Mhz, 128MB RAM). I can afford to buy a new notebook, but I should see a reason for this. Matrox chips are *not present* in notebooks, and NVIDIA chips are only in high-end models, which IMO are out of reasonable price range. Besides, 85dpi for standard LCD panel (15" 1024x768) is not very good dpi value. If resolution/dpi can be at 150dpi-170dpi (say, 2048x1536, or at least 1600x1200, for 15" panel), than it's a reason to upgrade. Current panels with low dpi are not very good, say I am not impressed... -- Vadim Plessky http://kde2.newmail.ru (English) 33 Window Decorations and 6 Widget Styles for KDE http://kde2.newmail.ru/kde_themes.html KDE mini-Themes http://kde2.newmail.ru/themes/ _______________________________________________ Render mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/render
