The following notes are from the audio portion of NVIDIA's conference call with Tom Peterson and Bryan Del Rizzo covering the GeForce GTX 690.
- Benchmarks will begin to appear at your favorite web site very, very soon.
- Performance is expected to be "very close" to that of the GeForce GTX 680 in 2-way SLI.
- Absolutely the best GPU that NVIDIA has delivered.
- Records were shattered for the time, effort and money that were invested in the GeForce GTX 690.
- No plastic parts.
- Never before used, albeit expensive, materials are part of the GeForce GTX 690 - glossy look, cast aluminum and magnesium.
- Magnesium injected GPU and relevant connections allow LED reporting.
- Clear, scratch-resistant polycarbonate window.
- Nickel-plated vapor chamber.
- Low profile device.
- Ridges cut in baseplate for greater control of airflow and turbulence.
- First time ever SLI scales 100% - GPUs were always thermally challenged, but not the GeForce GTX 690.
- Clock speeds take advantage of the GPU Boost feature.
- "A beast compared to the competition" with a 6 to 9 month head start.
- Most power efficient GPU per watt (PPW).
- Dampening materials for acoustics make the GeForce GTX 690 "quieter" than the GeForce GTX 680 in 2-way SLI.
- A 915MHz default GPU clock speed increases to over 1GHz with GPU Boost.
- Designed for overclocking as GPU clock speed approaches 1300MHz.
- Each GPU operates independently. One shuts off completely and the other enters a low-power state as needed (for example, 2D Windows GUI operations).
- Absolutely the best performance and acoustics.
- Hard launch!
- The GeForce GTX 680 is in short supply due to high demand. NVIDIA is making sure that as many GPUs as possible are being manufactured.
- Powers up to three 2560x1600 monitors running at 120Hz.
- More exciting products and bleeding-edge services are forthcoming...
Update: Tom Peterson touched on micro-stutter during the conference call. Looking over my notes, I see two comments which go something like "the declining quality of those frame rates" and "more and more time not connected to the frame rate."
So yes, NVIDIA invested the time to substantially reduce or eliminate micro stuttering.