bit-tech Shopping: BFG Technologies ageia PhysX Physics Card

Rating: 8

Time with product: 15 Days

Strengths: Noticeable increase system performance and special effects form the instant that I put it in.

Weaknesses: None so far other than limited amount of titles that use this technology.

Summary: After my rebate I ended up paying under $100 from CompUsa.com. I had to upgrade my power supply because the Dell PC I had FALSELY stated that the power supply I have had a connector to power the accelerator. After hardware upgrade, supported games look terrific. I notice things that were "not there before." I would prefer that the PCI express slot could have been utilized instaed of my last PCI slot

Rating: 8

Time with product: 7 Days

Strengths: Increased system performance & special effects.

Weaknesses: No PCI-E version & lack supported games to date.

Summary: After my rebates I ended up paying under $100.00 which was a plus for me. I would have rather had a PCI-E version as I only have 2 PCI slots on my computer. Ageia had intended to make one in this flavor but has subsequently dropped the ball on this project. Their original design that was sent out for reviews and testing actually incorporated both slot configurations on the same card. Maybe over time with PCI becoming obsolete and if more people wine, Ageia will once again consider the PCI-E. I'm also hoping that over time more titles will become available. Currently alot of game titles are using HAVOK to employ physics, animation, etc. into games. Although this engine works very well, its still using the CPU & GPU to process all the additional information. To me this is a nobrainer, an actual physics processing unit "PPU" is the way to go.

Rating: 6

Time with product: 4

Strengths: Really makes a noticable impact for certian applications, even on a very high-end system.

Weaknesses: There are times it seems to not help at all; PCI interface; included game is lousy.

Summary: I bought this card when BFG was doing a holdiay rebate promotion and I had a coupon for the retailer that I didn't have much use for, and combined that made my purchase price for the card around $60. At that price, I figured I'd but it (if nothing else than for UT3).At that price, I'm happy with it -- but I don't think I'd have been happy at any higher price. The technology still isn't used by many apps or games, and sometimes it's even hard to tell if ti's making a difference. (In UT3, for instance, it really only seems to make a difference on the levels specficially designed for the Physx card and not realy much on any of the other levels.)Granted, I am running this card in a very high-end (and lovingly overclocked) system, so it might be of greater incremental benefit for someone who doesn't have a processor/memory combo that can easilt handle the physics in question. But to bw honest, using the Aegia demos, it's easy to see where the card makes a difference (with a ridiculously high polygon count) even on the high-end system.This discrepancy in seeming benefits between demo and "real world" apps may be caused by bottlenecking through the PCI bridge, or it simply may be that there are other bottlenecks in the system that keep the card form doing all it can in a more interactive setting.Regardless, if it can be gotten for a good price, it may be worth your while... Especially in a system that might have a good video card but non-stellar CPU... But it's very difficult to strongly endorse a product that comes bundled with software (the game CellFactor) that doesn't even NEED the Physx hardware to run smoothly (and which is just not worth the effort to even install as that game won't hold your interest for more than 3 seconds). Hopefully ythe next gen of Physx cards -- PCI-Expressed based ones that will hopefully make a more noticable difference in more settings -- will also come out when there are enough titles that support the technology to make it cost-effective. But for now, UT3 fans might want to get this PCI version of the card just to run the two CTF boards smoothly.

Rating: 8

Time with product: 3

Strengths: Latest price drop ($160), powerful physics, new game play options

Weaknesses: Limited support, power consumption, still waiting for promised titles.

Summary: It's a little bit like the chicken vs. egg discussion. With currently only one game supporting PhysX to generate a little bit better graphics (by rendering more particles etc. behave physically correct), Ghost Recon: Advanced Warfighter as the only one (other than a few other patches) was also not a good reason to put down $300. The end of 2006 appears to not change this by much and we're waiting for 2007 to make a difference. CellFactor (RC36) has been debated for its claim to require PhysX and some users actually found that i can run without it. What gets lost in those debates is the fact that the game scales down and some demanding effects are either disabled or excruciatingly slow (i.e tearing cloth). Nevertheless, PhysX is software which is supported by dedicated hardware. Hence it should not surprise anybody that it can run without a card -- how fast is a different question. While CellFactor's gameplay received some shoulder shrugging, the big next thing might be UT2007 (and other games based on Unreal 3 engine), since PhysX is implemented for a variety of effects. Scheduled for Q1, we will see what that really means, but it should reach much further than GR:AW. The hardware looks a little conventional and the chip in 130nm technology signs responsible for the relatively high power consumption (40W under load / 20W idle). (Be sure to have a AT power plug available as the card requires external power -- much like a high-performance graphics card. The installed fan is noticeable in otherwise quiet systems, but my x1900XT sure does tune it out. The biggest problem for PhysX besides the lack of a wide support by important games (as of right now), is the announced HavokFX which uses the GPU as the math cruncher. There is also the notion that this might save the expense for the PhysX hardware. Yet, in the process it gets often overlooked that HavokFX is likely to require a dedicated GPU to run similar effects as PhysX does. So in other words, the SLI/Crossfire is suddenly downgraded to a single graphics processor. The majority is likely to have only one card installed which not only reduced allowable physics but also takes away from the graphics power. On the other hand, HavokFX doesn't have to deal with the PCI bottleneck. Overall, the new pricing (at some vendors) and the expected release of at least two killer games (CellFactor and UT2007) make it worth the consideration as a present for the 2006 Christmas -- either to yourself or a loved one. If just Ageia would release more demos or other little gadgets that shorten the time. (The one included is no good showcase for any of the physics since it's way to easy to challenge an entry-level CPU.)

Rating: 8

Time with product: 15 Months

Strengths: Adds extra effects in games like GRAW2, Cellfactor, Warmonger and UT3, versatile software development kit, great future potential, price has come down to about $120 online

Weaknesses: More games that use the tech are needed, will be better utilized in coming months, difficult (for developers) to "add-on" to an existing game engine, no PCI Express version

Summary: (Jan 6th, 2008) The Ageia PhysX solution is a add-on PCI card that is solely dedicated to calculating physics within computer games. It is your standard PCI board with a slightly loud 40mm fan (note to Ageia, add a silent heatsink and PCI Express interface on the next go around) and is about the size of a Creative Audigy soundcard. It costs about $75 -150 (US) and is now available at major Etailers (never buy Best Buy, when you can save $50 online). Currently, there are now more than ten games which can utilize the PhysX API, which must be built into the game engine from the start. Some of the major games are Ghost Recon Advanced Warfighter (both I and II), Cellfactor (more of a tech demo), Warmonger (free download) and of course UT3. All of these games now utilize PhysX for more integral gameplay as compared to earlier titles that just used PhysX for graphical special effects gimmicks. (and you can run many of these games in software mode without PhysX but with much lower frame rates). The problem still hounding PhysX at this early point is that the gaming community is going to need more exciting titles along the lines of UT3 and GRAW2 for PhysX to develop a following. The use of PhysX in games can really be broken down into two stages; the earlier games like GRAW I where the only thing you could see was a bit more debris and particulate spray. And the latest games like GRAW II and UT3 where you still see the special effects but they are now more integrated into the gameplay and level objectives. In the earlier games, the particle effects were interesting the first few times you saw them, but in no way helped or hindered the success of your respective in-game objectives (blowing up a car did not send debris which shatters a nearby window that you must crawl thru to continue your mission). It was similar to optional AF and AA filtering for your graphics card. On the other hand, the Island level of GRAW2 and the add-on tornado levels for UT3 allow you to use PhysX effects to accomplish in-game objectives (blowing up a barrel CAN shatter a nearby window that you must crawl thru to continue to your objective) with much greater effect than say the gravity gun in Half-Life 2. PhysX would still find much greater acceptance if for example, there were optional add-on mods for Oblivion (Elder Scrolls IV) that required, hitting a large boulder with a certain spell so that it has to roll down a hill along a specific physics determined trajectory and close an oblivion gate (this would be difficult because the Gamebryo engine of Oblivion has not been optimized for the PhysX foundation). Then, you would see folks more eager to rush out to Best Buy, Newegg, etc to stock up. Currently, PhysX is beginning to move out of the novelty stage and is starting to enter a more mature phase where its API is being put to better use in more titles. Although, most owners are still early adopters and GRAW junkies, I do believe the technology has great promise. Not surprisingly, the SDK can also be used outside of the game realm to model real world phenomena such as elastic collisions, neural networks, strong gravitational fields etc. Currently, with several months on the card, I have reprogrammed it to work with neural network models where its speed is faster than a single core in context specific applications. However, this is the nature of technology. Just like the early 3D acclerators (graphics cards) in the 90s, this card is currently underutilized. It make take two or three years before we see physics calculations in games that are essential to the gameplay instead of primarily special effects gimmicks. I applaud Ageia for trying to get this to market and push the boundaries for game developers, and I look forward to future titles that put the tech to good use.