backtop


Print 10 comment(s) - last by NuronV.. on Jun 10 at 4:24 AM

Manufacturers tell DailyTech where AGEIA is headed

I recently had the opportunity to sit down with AGEIA board partners to see where the company was going. 

Arguably, the largest complaint AGEIA board partners have is there is no single entity guiding reviewers and marketing.  BFG, ASUS and others must rely on their own marketing teams to push the review guides and benchmarks.  Furthermore, the largest problem with AGEIA's marketing right now is there is no competition.  There is no singular way to benchmark PhysX against a competitor like NVIDIA-Havok or ATI's physics processing, which should be announced later today.

The golden chalice for physics processing would ultimately be a $100 add-on card, even if the card provided only half of the performance supplied by a full blown AGEIA card.  Currently, at $280, a PhysX card costs more than a serious video card upgrade, such as a GeForce 7900GT.  Almost anyone will tell you, if you're upgrading a system from a 6600GT the benefit is in grabbing a new video card rather than a PhysX processor.  However, at $100, the gain for someone using a mid-range card becomes much more justifiable if PhysX supports a title you own already.

Unfortunately, manufacturers tell us ASIC that AGEIA runs on costs more than $100.  As a result, the company is currently exploring plans to go the other direction with PhysX.  Corporate planning roadmaps reveal that the company is "looking into" the possibilities of multiple-PhysX cards and multiple PhysX chips on a single PCB.  Instead of appealing to the masses, the direction appears to appeal to the super-enthusiast.  Certainly, if AGEIA can substantially ramp production of its ASICs, we may see such a card.  Perhaps UMA-esque memory support to reduce the cost of onboard memory as well.

Ultimately, AGEIA's success will still hinge on middleware support.  It's not to say that AGEIA's middleware isn't good. On the contrary, from the developers we've talked to, AGEIA's middleware is excellent even when compared to the likes of Havok, but there is no easy "fix" to implement without incorporating hooks in the game engine in very early stages.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Future? What about now?
By PandaBear on 6/6/2006 1:02:06 PM , Rating: 2
Sounds like they have the right tool but the wrong market, or the wrong time to enter the market.

If you are asking developers to do extra work just for 1% of their target audience to use it, they are in big trouble. Unless they can work with MS to get it in DirectX or pay id to put it in the next gen engine, they will be in serious trouble.

ATI is already tweaking their design to do physics, I don't think nVIdia will help much either. We already learned from 3DFX that their route of targetting only high end is not going to work.




RE: Future? What about now?
By epsilonparadox on 6/6/2006 1:10:29 PM , Rating: 2
I don't think Ageia anticipated such a strong response or viable solutions from NVidia or ATI this quickly.


RE: Future? What about now?
By jkostans on 6/6/2006 4:51:39 PM , Rating: 2
Ageia will be gone by next year... well at least those stupid physx processors will. ATI seems to have a much better solution, a 1600XT is cheaper and faster than a physx card. By the time you will actually be debating buying a physics processor, there will be much better solutions from nvidia and ati for the same or better price.


Stupid eh?
By Cincybeck on 6/7/2006 12:11:58 PM , Rating: 1
I personally think it's clear that ATi and Nvidia are trying to cash in on AGEIA's endeavor. The Havok SDK has ran on the CPU since year 2000. I think one should ask themselves why all of a sudden they want to allow the SDK to utilize their GPU's now that a more realistic more advanced solution is present. With the current solution I doubt Havok will ever be able to do what PhysX is promising. Yes it is some what expensive, but how do you get stupid from that? Your ignorance must be bliss. I'm sure there was people just like you saying the same thing about 3Dfx and their stupid graphics processors 10 years ago.


Broken by design
By Griswold on 6/8/2006 11:50:38 AM , Rating: 2
AGEIA created the demand by offering a product. That is why the two behemoths didnt come up with their own cheapo solutions before AGEIA kicked the ball off - they apparently didnt believe in demand. Now, everybody is talking about it and at least ATI offers the possibility to buy or use an old, weaker card to do the same. It may not be as powerful as the AGEIA PPU but it might deliver the same results until AGEIA fixes whatever is holding it back (probably bandwith due to PCI bus limitations). But errors like this often kill a bird before it learns to fly.

I dont believe in physx in its current shape and form. Too expensive and not fast enough due to weak design (PCI instead of PCIe).


RE: Stupid eh?
By jkostans on 6/9/2006 4:53:21 PM , Rating: 2
Not really, I actually wanted a voodoo back in the day. When physics accelerators show a major visual and performance improvement and don't cost a fortune, I'll definately buy one. Just putting 1000's of physics objects in a game doesn't make it better, just gimmicky.


RE: Future? What about now?
By TheDoc9 on 6/7/2006 6:44:35 PM , Rating: 2
I think no one doubts AGEIA has a good potential product. but that's all it is - potential. These cards are WAY to expensive, At least with 3dfx you saw the difference in the game, with AGIA you lose performance?! It reminds me of turning on advanced, accelerated audio processing from creative or Aureal years ago. When you turned on this faster, better multi-channel audio with it's own dedicated processor, your game slowed down 15-20%!!

AGEIA needs to both - broker deals to get thier processor incorporated into graphics cards, and drop the current add on card price to $99.


By Master Kenobi (blog) on 6/8/2006 3:42:25 PM , Rating: 2
Agreed. Right now my 1900xtx works just fine, physX would be a neat addon but I don't see the need to drop 280$ for it, when it doesn't give me anything tangible just yet.


RE: Future? What about now?
By NuronV on 6/10/2006 4:24:33 AM , Rating: 2
A GPU might be faster but its not a physics processor, by that logic a cpu core would give many times the performance.

Unless they can get more support from developers and sort out the performance issues they might be better off selling their intellectual property to ati or nVidia to incorperate it onto their cards


By lemonadesoda on 6/8/2006 5:01:45 PM , Rating: 3
This is a prize case study...

... A great product concept... and probably water-tight NDA (non-disclosure agreements) that kept the lid on the AGEIA product until some weeks ago...

BUT...

1./ Marketing launch of product BEFORE the product was ready
2./ At a price point that was way too high (note that many many products are launched at a loss only to recover profit once manufacturing and sourcing gets cheaper due to volumes)
3./ First application demonstrates flaws in the system. They needed much better demonstation of concept applications
4./ Underestimation of ability of mega-corporations (like ATI and nVidia) to stomp all over their marketing efforts
5./ "Going it alone". A little company like AGEIA should have got the backing of a much bigger, branded, company to help legitimise the product and ensure sufficient sales
6./ Limited scope. If only these guys had seen that market credibility could have been earned in the non-gamer and corporate world. A co-processor with appropriate CODEC pumping out DivX movies at 5 minutes per film, or EXCEL add-ins producing giga-calculations per seconds, or photoshop plug-in filters crunching "live", or Super-pi running in 5 seconds (http://www.xtremesystems.com/pi/). These would have earned a lot of street cred.




“We do believe we have a moral responsibility to keep porn off the iPhone.” -- Steve Jobs














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki