NVIDIA creating a line of TITAN cards

Maximum PC is reporting that NVIDIA is going to produce a whole line of TITAN cards.

There will be a higher end Titan Ultra:

This SKU will reportedly utilize all of the GK110’s 15 SMX clusters, giving it an edge over the regular Titan which has one cluster disabled with a total of 2,688 CUDA cores and 240 TMUs. As as result, the Titan Ultra will boast the full quota of 2880 shader cores and 256 texture units. Further, as opposed to the original’s 837 MHz clock speed and 206-Watt power appetite, the Ultra will have a 950 MHz clock speed and a 225W TDP

And a TITAN LE:

it will reportedly have 2496 CUDA cores, 208 TMUs, 40 ROPs, 320-bit DDR memory interface, 5 GB of GDDR5 and a 180-190W TDP

I would totally love an ULTRA TItan, as that is just a sick amount of CUDA.

Bare Feats has posted a part 2 to it’s EVGA NVIDIA GEFORCE GTX 680 Review with Professional Apps

Bare Feats has posted it’s second part of it’s Geforce GTX 680 Review and the results are mostly good. It does get beat in Cinema 4D and Motion 5 by the GTX 680MX in the top of the line iMac, but that is likely because of the older hardware and buss speed of the now ancient MacPro over the newer hardware in the iMac. Still it does win for DaVinci Resolve, After Effects, and Octane. Hopefully it would really whoop ass in a new MacPro if that ever happens.

Awesome to have the choice now and not have to get something off of ebay that is many generations old!

Barefeats speed test the EVGA NVIDIA Geforce GTX 680 Mac Edition and it is good!

Barefeats has some speed tests and things are looking good for gaming, though I would like to see some Premiere Pro and After Effects Speed tests (especially with the Quadro 5000K in the tests).

Now if only Apple would allow more than 2GB of RAM (yes I know the QUADRO 5000K has 4GB of RAM, but you still have to hack your mac to get a PC NVIDIA card to run that has more than 2GB of RAM).

NVIDIA Releases the GTX Titan

titan

I read about this amazing new video card in Maximum PC magazine.

This is the CUDA leap editors have been looking for without having to invest in the super expensive QUADRO line. Based on the Kepler GK110 instead of the GK104, this single processor card has twice the transistors and almost double the CUDA of the GTX 680! The GK110 are the chips used in the TESLA K20X GPU.

2688 CUDA cores in 14 SMX unites and 6GB of GDDR5 memory at 6GHz and a clock speed of 836MHz with a boost to 876Mhz and a 384 bit memory interface! Damn! And for $999. It also has GPU boost which will let you overclock it and have it change performance based on heat!

You can check it out over at NVIDIA’s home page as well. And Maximum PC has released Benchmarks as well, which show it not as fast as the dual processor GTX 690, though for a single CPU it is faster (though they are the same price), but it will likely be better for CUDA for NVIDIA which at least on a Mac can’t deal with the 2 processors of the GTX 690, though of course we can only hope to get this running on a Mac anyway!

EVGA Announces GeForce GTX 680 Mac Edition!

evga_gtx_680_mac

Macnn has the news on this exciting new video card for macs! And EVGA has it’s page up on this monster. Finally a great Mac video card for a reasonable price. It has 1536 CUDA cores, 2GB of GDDR5 RAM and an HDMI, Display Port, and 2 DVI connections, and Newegg is claiming it comes out APril 8th for $600. Must have solved the power issue, as usually the 680’s require 8 pin power and draw more than a MacPro can handle, which is why I have an EVGA GTX 670 in my mac.

Obviously they have not gotten around the Apple enforced 2 GB limit on Video cards, which is why I had to hack my machine to get my 4GB card running, but hopefully it does mean better video drivers for the 6 series Kepler cards! Woohoo! Should give me a performance boost for sure!

You can download the PDF spec sheet right here and start drooling!

Finally Premiere Pro has an official video card that will really kick ass and not break the bank.

Arstechnica on new NVIDIA 700M GPUs

Arstechnica has an article on the NVIDIA 700M GPU’s, which NVIDIA announced yesterday.

And basically it sounds like the the new NVIDIA GPU’s are the same as the 600M series, but with some slightly higher clock speeds, and a new GPU Boost 2.0, though these are all the lower end versions, and they haven’t talked about the higher end versions as of yet.

More speed is always good, and faster CUDA even better, but I would rather see a new series be a new architecture than just a speed bump.