S3 launches DirectX 10.1 Chrome 400 GPUs

Friday 15th February 2008, 11:30:00 AM, written by Arun

Despite losing more than half of their GPU/IGP market share in the last year according to Jon Peddie Research, VIA/S3 is hoping to change that situation with a new family of 65nm/DirectX 10.1/PCI Express 2.0 GPUs. The Chrome 400 Series will become available in late February 2008.

No performance information was available, although it uses a 64-bit memory bus and S3 claims to have achieved the "the highest performance-per-watt ratio ever", clearly hinting at their determination to recapture some share in the notebook market where they used to be relatively strong. The chip integrates DisplayPort, HDMI 1.2 and HDCP capabilities and fully accelerates H.264 video decoding. Similarly to NVIDIA's VP2 engine but unlike AMD's UVD, it does not seem to handle all of the VC-1 decoding process.

This family of GPUs will also be manufactured on Fujitsu's 65nm process, like VIA's recently launched Isaiah processor. No further details on pricing, SKU configurations or design wins were made available at press time.


Discuss on the forums

Tagging

graphics ± via, s3, fujitsu


Latest Thread Comments (65 total)
Posted by itaru on Saturday, 21-Jun-08 22:02:35 UTC
440GTX
geometry shader performance is very good !!
stream out performance ?

Image: http://xs228.xs.to/xs228/08256/cubemapgs440gtx774.jpg
Image: http://xs128.xs.to/xs128/08256/cubemapgs3870793.jpg

Image: http://xs228.xs.to/xs228/08256/metaballs440gtx858.jpg
Image: http://xs228.xs.to/xs228/08256/metaballs3870933.jpg

Image: http://xs128.xs.to/xs128/08256/particlesgs440gtx263.jpg
Image: http://xs228.xs.to/xs228/08256/particlesgs3870619.jpg

Posted by RejZoR on Monday, 23-Jun-08 08:14:25 UTC
Wow, not bad at all. Any possible comparison in something real (like games)?
I'd like to see how it performs there compared to this Radeon card (HD3850 maybe?)

Posted by Morgoth the Dark Enemy on Monday, 23-Jun-08 09:24:42 UTC
Quoting RejZoR
Wow, not bad at all. Any possible comparison in something real (like games)?
I'd like to see how it performs there compared to this Radeon card (HD3850 maybe?)
It's not aimed at the 3850, so it probably wouldn't fare all that well.

Posted by itaru on Monday, 23-Jun-08 13:21:59 UTC
Because TMU and ROP are important by a real game, 440GTX will be slow.
What you should pay attention to is geometry shader performance here

As you know, as for GPU of ATI R600, geometry shader is fast.
But 440GTX is faster.

Image: http://xs128.xs.to/xs128/08261/440cube277.jpg
Image: http://xs228.xs.to/xs228/08261/3870cube731.jpg

Image: http://xs128.xs.to/xs128/08261/440car390.jpg
Image: http://xs228.xs.to/xs228/08261/3870car895.jpg

Posted by phixcoco on Thursday, 26-Jun-08 03:13:49 UTC
[quote=itaru;1180465]
Because TMU and ROP are important by a real game, 440GTX will be slow.
What you should pay attention to is geometry shader performance here

As you know, as for GPU of ATI R600, geometry shader is fast.
But 440GTX is faster.
[quote]

Then, what's your point?

With 440GTX with faster geometry shader, what's the advantage we can take use of?
And can we programmer take the advantage, as S3G doesn't open anything on Chrome 400 serial, eg. programmer's guide!

Posted by Scali on Thursday, 26-Jun-08 08:55:52 UTC
Quoting phixcoco
Then, what's your point?

With 440GTX with faster geometry shader, what's the advantage we can take use of?
And can we programmer take the advantage, as S3G doesn't open anything on Chrome 400 serial, eg. programmer's guide!
Geometry shader is a standard DX10 feature. See the DirectX SDK for information.

Posted by itaru on Thursday, 10-Jul-08 22:01:30 UTC
Chrome440GTX GPUBench

WOW

Streaming: Basic Throughput
Image: http://xs329.xs.to/xs329/08284/streamthroughput699.png
Image: http://xs129.xs.to/xs129/08284/streamthroughputbytes115.png
Streaming: Triangle vs. Quad
Image: http://xs329.xs.to/xs329/08284/streamtriquad649.png
4-Component Floating Point Input Bandwidth
Image: http://xs329.xs.to/xs329/08284/inputfloatbandwidth453.png
Scalar vs Vector Instruction Issue
Image: http://xs129.xs.to/xs129/08284/instrissue_scalar841.png

GeForce8800
http://graphics.stanford.edu/projects/gpubench/results/8800GTX-0003/

Posted by Mintmaster on Thursday, 10-Jul-08 22:29:10 UTC
You realize that your graphs are suggesting 8 teraflops, 2 terabytes/s BW, and 140 GPix/s fillrate, right? :lol:

Posted by itaru on Saturday, 12-Jul-08 17:09:24 UTC
AMD DX10 Shader tests

Image: http://xs329.xs.to/xs329/08286/shadertest315.jpg

Perhaps score of GPUBench will be an error.
Chrome400 tends to have resembled G80 by this test.

2900,8800 used score of this place.
http://techreport.com/articles.x/12458/3

4870,440GTX is actual survey score.

Posted by AnarchX on Tuesday, 26-Aug-08 17:18:32 UTC
Seems 440 GTX got a factory speed-up to 800/1100/800MHz, but there is still headroom for *1030MHz GPU OC*:http://translate.google.ch/translate?u=http%3A%2F%2Fwww.pcgameshardware.de%2Faid%2C657627%2FNews%2FS3_Chrome_440_GTX_erreicht_1030_MHz_GPU-Takt%2F&sl=de&tl=en&hl=de&ie=UTF-8


Add your comment in the forums

Related graphics News

OpenGL 3.0 is here (finally)
[Analysis] TSMC 40G to deliver up to 3.76x the perf/mm^2 of 65G & Power Implications
Old News: AMD CTO resigns, NVIDIA CFO retires, DDR3 for MCP7A, S3, etc.
SwiftShader 2.0: A DX9 Software Rasterizer that runs Crysis
GPGPU and 3D luminaries join 3D graphics heavyweights
The Technology of a 3D Engine - Part One
Samsung joins Qimonda and Hynix in the GDDR5 race
Stanford previews new F@H GUI and GPU client at SuperComputing 2007
Crytek postpone Crysis beta until October 26th
New Ageia PhysX processor on PCIe pictured