BTW, talking to the NVidia rep at NAB, by far the #1 request they received was for a Thunderbolt CUDA engine. The response was that they aren't sure that they can get the performance out of Thunderbolt to make it worthwhile.
Wikipedia: "Thunderbolt is interoperable with DisplayPort 1.2 compatible devices. When connected to a DisplayPort compatible device the Thunderbolt port can provide a native DisplayPort signal with 4 lanes of output data at no more than 5.4 Gbit/s per lane. When connected to a Thunderbolt device the per-lane data rate becomes 10 Gbit/s and the 4 lanes are configured as 2 channels with each bidirectional 10 Gbit/s channel comprising one lane of input and one lane of output."
I don't know how much performance they need.
Tim Sassoon SFD Santa Monica, CA
On May 3, 2012, at 1:40 PM, Chris Bobotis wrote: Wouldn't it be great if NVIDIA or someone who is very GPU savvy at Adobe could chime in. Just saying. Cheers, Chris Mettle.comSkype: Mettlecom On 2012-05-03, at 4:36 PM, Dave Bittner wrote: I'm fairly sure that Rob picked that up from me. :)
So, to be fair, it's not exactly independent confirmation.
On May 3, 2012, at 3:25PM, Chris Meyer wrote:
Diving through this thread seems to confirm the info Rob Birnholz picked up at NAB that the GTX cards are optimized for writing to screen and hobbled at writing data back to the CPU, compared to the Quadro cards which are more generalized graphics processors than "just" display cards.
- Chris
+---End of message---+
To unsubscribe send any message to <ae-list-off@media-motion.tv>
+---End of message---+ To unsubscribe send any message to <ae-list-off@media-motion.tv>
|