The GTX 280 and 260 GPUs

Like the G80 launch, NVIDIA is providing two GeForce GTX 200 SKUs for today’s launch: the flagship GeForce GTX 280, and its feature-reduced sibling, the GeForce GTX 260. Both SKUs are built off the same GPU, only NVIDIA disables two thread processing clusters on the GeForce GTX 260, leaving just 192 active stream processors. As we mentioned on the previous page, the GTX 260 also has a narrower memory interface. NVIDIA also disables one ROP.

The following chart sums up the differences between the GTX 280 and the GTX 260:

GeForce GTX 280 and 260 Specs
GPU	GTX280	GTX260
Manufacturing Process	65-nm	65-nm
Core Graphics Clock Speed	602MHz	576MHz
Stream Processor Clock	1,296MHz	1,242MHz
# of Stream Processors	240	192
Memory Clock Speed	1,107MHz (2,214MHz effective)	999MHz (1,998MHz effective)
Memory Interface	512-bit	448-bit
Memory Bandwidth	141.7GB/sec	111.9GB/sec
Total Memory	1GB	896MB
ROPs	32	28
Texture Filtering Units	80	64
Texture Filtering Rate	48.2Gigatexels/sec	36.9Gigatexels/sec
HDCP Support	Yes	Yes
HDMI Support	Yes (via HDMI adapter)	Yes (via HDMI adapter)
Connectors	2xDual-link DVI 1xHDTV-out	2xDual-link DVI 1xHDTV-out
RAMDACs	400MHz	400MHz
PCI Express 2.0	Yes	Yes
Form Factor	Dual-slot	Dual-slot
Power Connectors	1x8-pin/1x6-pin	2x6-pin
Max Board Power	236 Watts	182 Watts
GPU Thermal Threshold	105 Degrees Celsius	105 Degrees Celsius
MSRP	$649	$399

As you can see, the clock speeds are slightly slower for the GeForce GTX 260 as well. While the GTX 280 features a 602MHz core with its shaders running at nearly 1.3GHz and 1.1GHz memory, the graphics core on the GTX 260 runs at 576MHz, with 1.24GHz shaders and 1.0GHz memory. And of course, you’ll no doubt notice the difference in price: $400 versus $650.

The first GeForce GTX 280 boards will be hitting retail starting tomorrow, while we’ve been told that GeForce GTX 260 cards won’t be available until June 26th.

Our GTX 280/260 ref boards

Power efficiency

With 1.4 billion transistors inside, power management is crucial for the GTX 200 to remain viable inside today’s PCs. To achieve this, NVIDIA employs a number of power-saving techniques, including clock-gating which shuts down parts of the GPU that aren’t being used, and dynamic clock speed/voltage adjustment.

The end result is a GPU that consumes just 25W at idle according to NVIDIA, while video playback mode (including Blu-ray) requires just 35W. Under full 3D load, with all the functional units inside the GPU cranking to the absolute max, the GPU consumes up to 236W.

The GeForce GTX 280 and 260 also support NVIDIA’s HybridPower technology, so if you have a HybridPower-compatible motherboard, the graphics card can shut itself completely off, consuming 0W, while the motherboard’s IGP handles graphics duties.

GPU-based physics

With NVDIA’s acquisition of PhysX, the two companies have been hard at work porting the PhysX API using NVIDIA’s CUDA SDK. As a result of this effort, any CUDA-capable GPU (this includes every GeForce 8/9/GTX 200 graphics card) is compatible with AGEIA PhysX. We’ve been told that NVIDIA will have a beta PhysX driver for press evaluation sometime later this week, so hopefully we’ll be able to check out GPU-based physics for real very shortly.

Beyond Gaming

Starting with GeForce 8 (and AMD’s own Radeon 2K series) the GPU has evolved from being used solely as a graphics processor to a more diverse general purpose processor that can be used for a variety of tasks that were traditionally handled by the CPU. What kind of applications are on the horizon for the GPU? How does audio and video encoding sound?

The GPU has caches, just like a CPU, and features multiple processing cores with support for multi-threading, not to mention massive amounts of memory bandwidth. In fact, with 240 processing cores running at nearly 1.3GHz, NVIDIA argues that the GeForce GTX 280 makes a compelling alternative to a quad-core Core 2 CPU running at 3.0GHz or even 4.0GHz.

To demonstrate what will be possible with GPU acceleration in the near future, Adobe offered members of the press a sneak peek at their next-generation Photoshop software launching in October. Codenamed Stonehenge, we were shown a 442-megapixel (!) image of Adobe’s HQ in California. This camera was positioned in the lot across from the Adobe campus and the file was over 2GB in size.

Using the GPU to accelerate Photoshop, the Adobe representative was able to zoom in so close we were able to read street signs, license plate numbers of cars in the parking lot, and check out what was happening inside individual offices. With such a high resolution image, you’d assume it would take minutes to accomplish this, but thanks to GPU acceleration these tasks were accomplished in the blink of an eye! Adobe then used the GPU to effortlessly edit a 3D panorama in real-time. It was all quite impressive.

All of this is made possible thanks to CUDA.

But what about video encoding? We were also given a beta copy of Elemental Technologies BadaBoom media converter software to test with. We then made a 300MB MPEG-2 video. Using BadaBoom, we converted this video file to a 640x365 H.264 movie suitable for playback on an iPhone (When it’s released, Elemental’s BadaBoom software can be used to convert various videos for use in portable devices like iPhone/iPod and Zune). The conversion took just 31 seconds. We then performed the same task using QuickTime Pro and a Core 2 QX9770. The process took three minutes and two seconds to perform the same task.

We’ve been told that BadaBoom will be launching sometime in August. A final price hasn’t been determined yet. Elemental also plans to create a plug-in for Adobe Premiere in the near future.