Slashdot Mirror
NVIDIA Unveils Next-Gen Turing Quadro RTX Professional Graphics Cards (hothardware.com)
MojoKid shares a report from Hot Hardware: We been hearing a lot about NVIDA's next-generation GPU architecture since late last year, and today NVIDIA is announcing the first products based on Turing. NVIDIA is targeting the professional graphics market first with its new Quadro RTX 8000, RTX 6000 and RTX 5000 GPUs. Turing GPU architecture replaces Pascal, which has served both the consumer and professional markets since 2016. But as its 8th generation GPU architecture, NVIDIA is ushering in a number of advances with Turing and it's billed as the "world's first ray-tracing GPU." When it comes to content creators, NVIDIA claims that with the power of Turing, "applications can simulate the physical world at 6x the speed of the previous Pascal generation."
Getting down to brass tacks, the entry-level Quadro RTX 5000 has 3,072 CUDA cores, 384 Tensor cores, and will come with 16GB of 14Gbps GDDR6 memory. Its ray-tracing performance is dialed in at 6 GigaRays/sec, according to NVIDIA. Both the Quadro RTX 6000 and RTX 8000 have 4,608 CUDA cores and 576 Tensor cores; the only difference between the two is that the former has 24GB of GDDR6, while the latter doubles that to 48GB. Ray-tracing performance for both of these GPUs tops out at 10 GigaRays/sec. NVIDIA is also claiming up to 16TFLOPs compute performance for the Quadro RTX 8000. NVIDIA's new Quadro GPUs will also be among the first to support both USB-C and VirtualLink for next-generation virtual reality and mixed reality headsets. Other VirtualLink backers include AMD, Oculus, Microsoft and Valve. The Quadro RTX 5000, RTX 6000 and RTX 8000 will all be available during the fourth quarter of 2018 priced at $2,300, $6,300 and $10,000 respectively.
Getting down to brass tacks, the entry-level Quadro RTX 5000 has 3,072 CUDA cores, 384 Tensor cores, and will come with 16GB of 14Gbps GDDR6 memory. Its ray-tracing performance is dialed in at 6 GigaRays/sec, according to NVIDIA. Both the Quadro RTX 6000 and RTX 8000 have 4,608 CUDA cores and 576 Tensor cores; the only difference between the two is that the former has 24GB of GDDR6, while the latter doubles that to 48GB. Ray-tracing performance for both of these GPUs tops out at 10 GigaRays/sec. NVIDIA is also claiming up to 16TFLOPs compute performance for the Quadro RTX 8000. NVIDIA's new Quadro GPUs will also be among the first to support both USB-C and VirtualLink for next-generation virtual reality and mixed reality headsets. Other VirtualLink backers include AMD, Oculus, Microsoft and Valve. The Quadro RTX 5000, RTX 6000 and RTX 8000 will all be available during the fourth quarter of 2018 priced at $2,300, $6,300 and $10,000 respectively.
24 * 2 = ?
1% APY, No fees, Online Bank https://captl1.co/2uIErYq Don't let your $$$ sit in a no-interest acct.
"The only difference between the two is that the former has 24GB of GDDR6, while the latter doubles that to 96GB. "
So does someone need a typing or a maths class?
If only I had an application that would be tractable on this sort of hardware. The ray tracing side of things is boring to me, i'd be interested to know how this works out for large matrix operations, or signal processing type applications. Still, more is more when it comes to compute!
The professional cards are used for a different workflow where stability and visual fidelity matter the most and it involves certification of hardware and drivers for compatibility with various CAD and healthcare applications. This is principally the reason for high prices. Usually they cannot match the FPS scores of non-professional cards from the same generation as the latter ones are used primarily for fragging and in this case they focus on squeezing out every bit of performance while being more relaxed on stability concerns.
The cards themselves aren't that much different or better. Typically, they have more memory (possibly ECC as well) than your run of the mill gaming card and NVidia might include tensor cores or other specialized hardware in some of these high-end lines, but otherwise it's the same design as their consumer cards. What you're really paying for is the professional drivers. Basically you pay extra to get a lot of optimizations for your CAD program as well as certification that it will produce reliable results in those applications.
Crap, I just bought a P4000.
Note that if you do RTFA, it's still badly explained, wrongly written but you get 48GB on the top end Quadro (not in the article : it's up from the older 24GB on GDDR5, because GDDR6 is available with twice the density)
If you have two Quadro RTX 8000 linked with a bridge, you get a real and genuine 96GB memory space in the same way it happens on modern dual CPU systems. Coherent, seamless but bandwidth is much slower if you go over the border of course. This feature was already available on Quadro GP100 (Pascal with fast double precision floating point ; 16GB HBM2 on each of two cards, combined 32GB)
It is a NUMA system, much like Opteron and dual Xeon E5 before.
Accessorily the article says the bridge feature (NVLink) is available on all three new boards, so two RTX 5000 is 16GB + 16GB, two RTX 6000 is 24GB + 24GB.
On the picture with the boss showing two artistic (stylized, unrealistic) representations of Pascal GP102 vs Turing, it's written : 24GB for the GP102, 48GB + 48GB for the large Turing.
(GP102 is using on GTX 1080 Ti with 11GB GDDR5X RAM, on Titan Xp with 12GB RAM, on Quadro P6000 with 24GB RAM)
I don't believe that Google gets a licensing fee on common math. Yet.
https://en.wikipedia.org/wiki/...
Yes. You run a week long simulation only to find an error caused by unstable hardware.
Only the State obtains its revenue by coercion. - Murray Rothbard
Ah, I see your mistake here. I bet you've been buying the latest, younger cards instead of the older, more mature ones. My GTX 650 has always been quiet, on time and never gave me any trouble.
#DeleteFacebook
The pro cards (Quadro, and also some NVS cards) nowadays come with four DisplayPort outputs. The consumer cards tend to have one DisplayPort and a mishmash of HDMI and DVI. That is annoying, since there are converters from DisplayPort to both of these but not the other way round. The GeForce cards also tend to be dual-slot, power-hungry monsters while the lower-end Quadro cards fit in a single slot and don't need auxiliary power.
So I think there are reasons to prefer the Quadro line, apart from the question of "certification" (which does look like like a bit of a racket).
In principle you can license and build a GeForce card with four DisplayPort connectors, but they are very hard to find. The best I was able to buy was a GeForce 1060 with dual DisplayPorts.
-- Ed Avis ed@membled.com
Couldn't make it TWO WORDS without a grammatical mistake? Literally the minimum necessary number of words for such a mistake. That's damn impressive.