facebook rss twitter

GeForce 8 series cards to get PhysX upgrade

by Parm Mann on 18 February 2008, 12:25

Tags: NVIDIA (NASDAQ:NVDA)

Quick Link: HEXUS.net/qalrv

Add to My Vault: x

Software update to bring PhysX to GeForce 8

NVIDIA's aquisition of Ageia raised its fair share of questions but thankfully, one of the key questions coming from GeForce 8 series owners was partly answered during NVIDIA's fourth-quarter financial results conference call.

The question; will GeForce 8 series GPUs benefit from NVIDIA's ownership of PhysX technology?, was answered in part by NVIDIA CEO Jen-Hsun Huang. Though no official announcement has yet been made, Huang stated:

"We're working toward the physics-engine-to-CUDA port as we speak. And we intend to throw a lot of resources at it. You know, I wouldn't be surprised if it helps our GPU sales even in advance of [the port's completion]. The reason is, [it's] just gonna be a software download. Every single GPU that is CUDA-enabled will be able to run the physics engine when it comes."

As all of NVIDIA's GeForce 8 GPUs support CUDA, we can expect PhysX technology to be added across the range via software upgrades.

All we need to know now, is when will it happen? My guess would have to be after the launch of NVIDIA's imminent 9 series.

HEXUS Forums :: 7 Comments

Login with Forum Account

Don't have an account? Register today!
Posted by Lucio - Mon 18 Feb 2008 14:28
Could be handy, but won't it impact on the frame rate if part of the card is busy doing PhysX calculations and lead to worse, rather than better performance on nVidia cards?

Or is this just a ploy to tap into the unused potential of SLI to boost the sales of their motherboards and cards…
Posted by kalniel - Mon 18 Feb 2008 14:47
Yes, to both of those I think Lucio.

We'll have to see which a game uses most - CPU cores or GPU power. If GPU (like most games at the moment) then doing physics on unused cores seems a better way of doing things (and Intel will be pushing this with it's acquisition of Havok).
Posted by Kumagoro - Mon 18 Feb 2008 18:10
I have never liked the idea of giving up Graphics power for physics. I was under the
impression that extra physics puts your GPU under more pressure from all the extra
bits and pieces it needs to render etc.

How much of a trade off is there between running a dedicated PPU chip compared to
it being emulated by the GPU. If there is a significant difference then I would prefer
the PPU integrated onto the GPU die. That way the power of the PPU could be
matched with that of the GPU.

In any case physics will only be properly viable when its a part of directX
Posted by this_is_gav - Mon 18 Feb 2008 19:24
Kumagoro
How much of a trade off is there between running a dedicated PPU chip compared to it being emulated by the GPU. If there is a significant difference then I would prefer the PPU integrated onto the GPU die. That way the power of the PPU could be matched with that of the GPU.
When this was announced I was expecting that to be the case too. I certainly wasn't expecting it to just be a software thing. If that's the case what little impression of PPUs I had has been lowered further - if a GPU can do it alongside it's normal graphics processing then it can't be taking that much of a hit, so did the PPU do much at all? Was it just a marketing gimmick for what was primarily a software layer?
Posted by Zak33 - Tue 19 Feb 2008 09:30
doesn't this make a mockery of PhysX?

I mean….they told us that it ADDED to the gaming experience by doing more…now it turns out the card does it anyway.

I'm baffled.

SEE NEWER »