Introducing a new card with a familiar name
NVIDIA is updating the GTX 260 to be more competitive with the recently Radeon HD 4800 releases from AMD. Can this upgraded GTX 260 model put NVIDIA back into the clear lead for performance gaming? We take a look at an overclocked BFG model to find out!
Pulling out the StopsThat old adage of “competition is good for consumers” is over played in my eyes; competition is GREAT for consumers. Take today’s review for an example: compared to NVIDIA’s GeForce 8800 GTX, available in early 2007, the new GeForce GTX 260 card gets 1.4x the performance in Bioshock for about 1/4 the price of the 8800 GTX when it was released. Equivalent performance THEN, with the GeForce 8800 GTX, would have cost you about $840 if we pretend we could scale a GPU 1:1 by price/performance to today’s $249 GTX 260 cards. Not too bad.
One of the side effects of such a rapidly changing segment of the PC market is…change. We see more new GPU and graphics SKU released than with any other component your system – both NVIDIA and AMD are valiantly fighting for the title of “best performer” or “best value” in order to capture to dollar of the enthusiast and mainstream gamer. Today we are looking at the third iteration of the GT200 GPU that NVIDIA brought to market in June of 2008: the GeForce GTX 260…wait, what?
The GeForce GTX 260 gets a Refresh
Don’t be confused my readers, there is an explanation for this. Today NVIDIA (or more accurately, NVIDIA’s partners) are releasing a new revision of the GeForce GTX 260 graphics board that will be faster performing, maintain the $299 MSRP of the original and drive the price down of the “original” GTX 260.
First, a quick background; if you remember from our initial GT200 article, the GTX 280 card (as the top-performing flagship GPU) sports 240 shader processors in 10 blocks of 24 as seen in the diagram below:
The GPU also carried with it a full 512-bit memory bus, indicated by the 8 64-bit memory controllers along the bottom of the image. To offer a lower end part, NVIDIA took this basic GPU design and disabled/chopped off a small portion in order to improve yields and lower the price. The result was the GTX 260 seen here:
With two fewer SP blocks the GTX 260 had 192 shader processors and a 448-bit memory bus, and all was right with the world. Until AMD released their Radeon HD 4870 512MB graphics card that beat out the NVIDIA GeForce GTX 260 cards at a much lower cost. NVIDIA responded by lowering the prices on the GTX 280/260 series but that wasn’t enough to get the price/performance crown back into their favor.
NVIDIA’s reaction is what you see today; a “new” GTX 260 card seen here:
NVIDIA’s reaction is what you see today; a “new” GTX 260 card seen here:
This release also leaves us with quite a confusing product stack from NVIDIA:
- GeForce GTX 280
- GeForce GTX 260 (new, 216 shaders)
- GeForce GTX 260 (original, 192 shaders)
- GeForce 9800 GTX+ (55nm)
- GeForce 9800 GTX (65nm)
- ….
NVIDIA is NOT creating any kind of new brand for this product however: there is no GTX 260+ designation, no move to the GTX 270, nothing like that. Instead NVIDIA is leaving the “branding” part up to their partners and AICs so expect each board vendor to come up with their creative name for these updated GeForce GTX 260 cards. For our testing sample, BFG Tech came up with the GeForce GTX 260 MAXCORE models. No doubt this exercise in branding will lead to a high number of confused buyers trying to discover why one GTX 260 card is so much better than another, both from vendor X and both sporting the same friggin’ name. Likely this is a great “accidental” side effect of NVIDIA’s decision as it will help current GTX 260 cards to sell better as the GTX 260 brand is brought to light once again with this launch.