Read more.NVIDIA's Forceware 177.39 driver will introduce PhysX support to an array of GeForce cards and promises a "terrific performance boost".
Read more.NVIDIA's Forceware 177.39 driver will introduce PhysX support to an array of GeForce cards and promises a "terrific performance boost".
Well yes, using the GPU to accelerate a CPU only benchmark when no other demands are being made on it will of course give a very nice synthentic benchmark boost
What happens when you have to divert GPU calculations in order to do physics though?
And just how well does the card switch from one API to another, on the fly?
PhysX is enabled via CUDA and most games via D3D. The architecture will need to arbitrate between the two on a near-instant basis.
UT3 will be interesting.
I'm in the process of finding out.
This could sway people to not buy the new 4850/4870's if NVIDIA get this right.
Mobo: DFI LP X48 LT T2R Bios 2008/12/24
CPU: Intel Q6600 G0 @ 3.6ghz (400 x 9)
Ram: OCZ PC2-9200 4GB Flex II @ 1066mhz 5-5-5-15 2T
GC: HIS 5870 @ 900/1300 - Eyefinity
Does this mean that GeForce 8 series cards and the 9800 GX2 are not going to support PhysX?
Here's the answer...
At this moment, PhysX support has only been enabled for those four GPU's:
GeForce GTX 280
GeForce GTX 260
GeForce 9800 GTX+
GeForce 9800 GTX
NVIDIA will progressively enable PhysX support in the other GeForce 9 and GeForce 8 graphics over the next few weeks.
w00t! Now we can pwn all the people who said 9800gtx is just an overclocked 8800gts
Here is an idea of what happens in Vantage chaps.
E2140 & 9800GTX at stock.
175.16
P4078
GPU - 4907
CPU - 2707
177.39 & PhysX driver
P6477
GPU - 5293
CPU - 19666
As you can see a very notable increase in CPU test 2 which is the physics test.
lol, the 8800 series and 9600 are supported in this update with a little ini editing. I just installed the driver on my 8800GTS now and will be trying out the performance difference after watching rambo .
But there is a bit of a difference, the 9800GTX uses slightly better memory iirc hence the better clocking by like 50mhz, still not worth the cash except for you when you got off play .
edit: couldnt use vantage or anything but after installing the graphic quality in crysis seemed better, i tend to run at high 1680x1050 no AA but i used to be able to see lines around the objects (annoying,needed aa i think) now they are gone and images look better in general, think i got about 5fps+ increased to, stays between 45-60 fps ingame.
If you can find anyone to play against...
http://www.strategyinformer.com/edit...-six-months-on
500 players a night, max, worldwide. If that don't scream failure, nothing will.
UT2004 still has many more players than UT3. Its just a better game. Its *fun* UT3 just feels dead. Nothing feels right about it, which is a shame.
Look at Team Fortress 2 - As spammy as hell. A few balance issues, but playing it is fun. tons of players online most nights
ut2004 isn't that much further ahead..
it's still not a failure to them as they must be raking it in with developers buying their sdk to use their engine. UT99 has the most players and is still fun but the graphics take it down. I've been on the UT3 servers and there's always a game to play. I never seem to be playing on a server alone. Might be a failure in terms of buyers/players online but not for those who play
There are currently 1 users browsing this thread. (0 members and 1 guests)