GTX 285 Shows its 55nm Face

The new GeForce GTX 285 GPU uses the same great architecture (the GT200) that is behind the GTX 280 though shrunk down to a 55nm process technology. What the hell does that mean for you? Higher clocks, lower power and lower prices!

A week ago today we brought you the scoop on the new GeForce GTX 295 graphics card from NVIDIA that utilized a pair of 55nm GT200-based GPUs on a single card to produce the single fastest graphics card on the planet.  At the end of that write up, I mentioned that NVIDIA was already telling the world about the upcoming GeForce GTX 285 – a single GPU version of the same (or mostly similar GT200) 55nm GPU that would essentially replace the 65nm GeForce GTX 280 and offer slightly better performance with slightly lower power consumption.

Right on time, and with before launch availability, NVIDIA has delivered on its promise and brought us the GeForce GTX 285. 

Oh, and I’m not going to lie – this article is going to be short and to the point – we just got the cards in YESTERDAY so we didn’t have a lot of time with them yet so benchmarks are limited.  We did get to overclock them (but popular request in the forums) and take shiny pictures too.  Oh, and it’s about 4:15am here as I write this opening sentence…

What’s changed with this GTX 285 thing?

I get the impression that NVIDIA itself is actually tired of releasing new products.  After the onslaught in the past couple of months with new drivers, 3D Vision and the GTX 295 card, the GTX 285 is really getting a raw deal from the PR committee – but maybe that’s because it doesn’t really need any. 

What is essentially a die-shrink update to the existing (and EOL’d) GTX 280, the GTX 285 takes the same GT200 architecture and moves it from the 65nm process to the smaller, more power efficient 55nm process we have seen on other NVIDIA GPUs since mid-to-late last year.  This added power efficiency does lead to one strong positive for the new design: it no longer requires a 6-pin AND an 8-pin ATX power connector and can now run on a pair of 6-pin connections instead. 

55nm GT200: NVIDIA GeForce GTX 285 1GB Review - Graphics Cards 50
Anonymous GTX 285

If we look at the specs of the GTX 285 you should see identical specs to those found in our original GTX 280 article:
  • 240 stream processors
  • 80 texture units
  • 1GB GDDR3 frame buffer
  • 512-bit memory bus
Reference speeds did get a nice boost though:
  • Core: 648 MHz (GTX 280: 615 MHz)
  • Shader: 1476 MHz (GTX 280: 1350 MHz)
  • Memory: 1242 MHz (GTX 280: 1107 MHz)
Nothing too dramatic but a healthy bump up in speed; with the move to a more power efficient design NVIDIA is able to push the clocks a bit higher and maintain a similar (actually ends up being a bit lower) power envelope.  Also of note is that the two GeForce GTX 285 cards we got in for review were both running at overclocked speeds so it would seem NVIDIA let the flood gates open early on that front.

55nm GT200: NVIDIA GeForce GTX 285 1GB Review - Graphics Cards 51

Because the GT200 architecture hasn’t changed much, if at all, but the process has shrunk from 65nm to 55nm, we can assume that NVIDIA is saving around 20% die space on the new cores.  That saves them money per die and also potentially increases the yield rates giving the company a bigger boost in the bottom line.  Unfortunately, we can’t measure this ourselves on the card because of the above heat spreaders NVIDIA is utilizing.  The new 55nm chip (on the left) is actually using the same size heat spreader as the original GT200 (on the right). 

« PreviousNext »