Thursday, June 3rd 2010

Galaxy Designs GeForce GTX 480 with Dedicated PhysX GPU

While every NVIDIA GPU since GeForce 8 series, with over 256 MB of memory supports the CUDA GPU compute technology, and with it PhysX GPU acceleration, the prospect of dedicated GPUs for computing PhysX interests many, especially enthusiasts, after NVIDIA stopped production of the PhysX processor from erstwhile Ageia. EVGA first devised a graphics card that has a high-end GPU processing graphics, and a mainstream GPU dedicated to PhysX processing, with its GeForce GTX 275 Co-op. Building on the same principle, Galaxy designed an enthusiast-grade graphics card that uses an NVIDIA GeForce GTX 480 for DirectX 11 compliant graphics processing, while its sidekick on board is a 40 nm GeForce GT 240 GPU.

The GeForce GTX 480 sticks to specifications, complete with 1536 MB of GDDR5 memory across a 384-bit wide, while the GT 240 has its own 512 MB of GDDR5 memory across a 128-bit wide interface. Both GPUs share the system bus over an NVIDIA nForce 200 bridge chip, which gives both GPUs a PCI-Express 2.0 x16 link. The GTX 480 GPU packs 480 CUDA cores, and features the latest GPU technologies, including 3D Vision surround. Since it is independent from the GT 240, the GTX 480 can pair with three more of its kind for 4-way SLI, without affecting the functionality of the GT 240 in any way.

The GT 240 packs 96 CUDA cores, and is DirectCompute 4.1 compliant. It has enough compute power for offloading PhysX processing from the main GPU. Since the GT 240 in its standalone card form is very energy efficient by not requiring auxiliary power, its presence didn't affect the electricals of this card in any big way. The GTX 480 component seems to be powered by a 6+2 phase digital PWM circuit, while the GT 240 uses a simpler 2+1 phase PWM circuit. Power inputs remain 8-pin and 6-pin. More details as they emerge.
Add your own comment

44 Comments on Galaxy Designs GeForce GTX 480 with Dedicated PhysX GPU

#2
Fourstaff
Another marketing ploy to get people to buy GT2xx! Good job Galaxy :rockout:
Posted on Reply
#3
raptori
nice now they need to put powerful cooling not the reference.
Posted on Reply
#4
Assimilator
Is it just my imagination or is that board the same size as the standard GTX 480 PCB? Would be very interesting if Galaxy has managed to squeeze an extra GPU in there!
Posted on Reply
#5
xkche
What's the perfomance of GT240 on PhysX?
is enough to move high FPS to the GTX480??
Posted on Reply
#6
OnBoard
by: Assimilator
Is it just my imagination or is that board the same size as the standard GTX 480 PCB? Would be very interesting if Galaxy has managed to squeeze an extra GPU in there!
Pretty much. They used the air hole space for the extra GPU. Same way they made GTX 470 shorter, get rid of empty space and put it to use :)
Posted on Reply
#7
ktr
by: xkche
What's the perfomance of GT240 on PhysX?
is enough to move high FPS to the GTX480??
PhysX is mathematical calculations and is independent of graphic rendering.
Posted on Reply
#8
xkche
by: ktr
PhysX is mathematical calculations and is independent of graphic rendering.
Yes, but.. how can fast can calculate GT240 PhysX, to stay on par of the renders of the GTX480 ?
Posted on Reply
#9
ShRoOmAlIsTiC
the 240 does just fine with my 5850. Batman AA gets an average of 84fps in the bench with physx on high.
Posted on Reply
#10
phanbuey
just dont install the current WHQL's for it... as they seem not to work with 400 and older GPU's for PhysX.
Posted on Reply
#11
newtekie1
Semi-Retired Folder
Totally pointless. When I tested a 9600GT with my GTX470, there was no performance difference between having the 9600GT do PhysX and just letting the GTX470 do it. The GTX400 series cards, especially the GTX480, is more than powerful enough to handle PhysX by themselves.

Plus the same problem I had with the special HD5770 with the integrated Killer NIC applies here, when I want to replace the card, I loose the PhysX card.
Posted on Reply
#12
ktr
by: xkche
Yes, but.. how can fast can calculate GT240 PhysX, to stay on par of the renders of the GTX480 ?
Again, it is independent, therefore it does not have to "stay on par" with the GPU. PhysX aids the CPU, not the GPU.
Posted on Reply
#13
newtekie1
Semi-Retired Folder
by: ktr
Again, it is independent, therefore it does not have to "stay on par" with the GPU. PhysX aids the CPU, not the GPU.
I think what he is asking is if the GT240 is fast enough to calculate PhysX and not hold back the GTX480. The same way a weak CPU would hold back a stong GPU.

The issue is that people think PhysX takes a lot of computing power, when it really doesn't, so a very basic card can handle it easily without limitting the more powerful graphics card.
Posted on Reply
#14
$immond$
I wonder 2 of these in SLI would be like.
Posted on Reply
#15
xkche
by: ktr
Again, it is independent, therefore it does not have to "stay on par" with the GPU. PhysX aids the CPU, not the GPU.
by: newtekie1
I think what he is asking is if the GT240 is fast enough to calculate PhysX and not hold back the GTX480. The same way a weak CPU would hold back a stong GPU.

The issue is that people think PhysX takes a lot of computing power, when it really doesn't, so a very basic card can handle it easily without limitting the more powerful graphics card.
tnks! that's the point!
Posted on Reply
#16
v12dock
Anyway they can get rid of there massive G92 stock
Posted on Reply
#17
OnBoard
by: v12dock
Anyway they can get rid of there massive G92 stock
It's not G92 anymore. GT215 40nm and DX10.1.
Posted on Reply
#18
wahdangun
shame they don't put hydra chip in there, its maybe more useful than that nforce chip, so if you don't want to use physix you can still get performance boost
Posted on Reply
#19
trt740
Galaxy is really coming around this will be a great card.
Posted on Reply
#20
ShRoOmAlIsTiC
nvidia should start integrating a gt240 physx processor into motherboards and make it so they dont disable them when ati is present. Then lisence the chips to amd and intel.
Posted on Reply
#21
LittleLizard
did anyone noticed that it STILL USE 6+8, even with the gt240 on board?!?!?!
Posted on Reply
#22
Steevo
They should shut off the 480 core and memory when in 2D clocks and save a shitload of power. Then it would be worth it.
Posted on Reply
#23
_JP_
by: LittleLizard
did anyone noticed that it STILL USE 6+8, even with the gt240 on board?!?!?!
Your question can be easily answered by reading the whole news. Especially here:

by: btarunr

The GT 240 packs 96 CUDA cores, and is DirectCompute 4.1 compliant. It has enough compute power for offloading PhysX processing from the main GPU. Since the GT 240 in its standalone card form is very energy efficient by not requiring auxiliary power, its presence didn't affect the electricals of this card in any big way. The GTX 480 component seems to be powered by a 6+2 phase digital PWM circuit, while the GT 240 uses a simpler 2+1 phase PWM circuit. Power inputs remain 8-pin and 6-pin. More details as they emerge.
:D

I'm looking forward to see reviews of this card! Galaxy is making some very bold moves with these new exiting cards!
Too bad they don't sell in my country, though. :(
Posted on Reply
#24

I don't need PhysX, what am i going to do with it?

Where dah games at to run it?
#25
SteelSix
As Steevo said, imagine this thing shutting down the 480 core for surfing/idle. Hmm..
Posted on Reply
Add your own comment