by gigatexal on 5/10/17, 5:18 PM
These tensor cores sound exotic:
"Each Tensor Core performs 64 floating point FMA mixed-precision operations per clock (FP16 multiply and FP32 accumulate) and 8 Tensor Cores in an SM perform a total of 1024 floating point operations per clock. This is a dramatic 8X increase in throughput for deep learning applications per SM compared to Pascal GP100 using standard FP32 operations, resulting in a total 12X increase in throughput for the Volta V100 GPU compared to the Pascal P100 GPU. Tensor Cores operate on FP16 input data with FP32 accumulation. The FP16 multiply results in a full precision result that is accumulated in FP32 operations with the other products in a given dot product for a 4x4x4 matrix multiply,"
Curious to see how the ML groups and others take to this. Certainly ML and other GPGPU usage has helped Nvidia climb in value. I wonder if Nvidia saw the writing on the wall so to speak with Google releasing their specialty hardware called the Tensor hardware that Nvidia decided to use it in their branding as well.
by arca_vorago on 5/10/17, 6:32 PM
More great hardware being stuck behind proprietary CUDA when OpenCL is the thing they should be helping with. Once again proprietary lock in that will result in inflexibility and digital blow-back in the long run. Yes I understand OpenCL has some issues and CUDA tends to be a bit easier and less buggy, but that doesn't detract from the principles of my statement.
by hesdeadjim on 5/10/17, 5:40 PM
I find it so cool that technology created to make games like Quake look pretty has ended up becoming a core foundation of high performance computing and AI.
by mattnewton on 5/10/17, 5:57 PM
Wow, this is just Nvidia running laps around themselves at this point. Xenon Phi still not competitive, AMD focused on the consumer space, looks like the future of training hardware (and maybe even inferencing) belongs to Nvidia. (Disclosure: I am and have been long Nvidia since I found out cudnn existed and how far ahead it was)
by bmiranda on 5/10/17, 5:17 PM
815 mm^2 die size!
That's at the reticle limit of TSMC, a truly absurd chip.
by arnon on 5/10/17, 8:39 PM
This is odd for NVIDIA.
They usually push out revised versions in the second year, not change the entire architecture to the new one.
Feels like they're feeling AMD breathing down their necks with their VEGA architecture, which should be very interesting.
AMD have also stepped up their game with ROCm which might take a chunk out of CUDA.
by Symmetry on 5/10/17, 6:35 PM
by randyrand on 5/10/17, 6:00 PM
What are the silver boxes that line both sides of the card? Huge Capacitors?
by tobyhinloopen on 5/10/17, 5:41 PM
Time to play some games on it
by grondilu on 5/11/17, 4:00 AM
I was wondering if this will be used in supercomputers. Apparently yes:
> Summit is a supercomputer being developed by IBM for use at Oak Ridge National Laboratory.[1][2][3] The system will be powered by IBM's POWER9 CPUs and Nvidia Volta GPUs.
https://en.wikipedia.org/wiki/Summit_(supercomputer)
Summit is supposed to be finished in 2017, though. I'm quite surprised this is possible since the Volta architecture has only just now been announced.
by lowglow on 5/10/17, 7:10 PM
I'm really happy our startup didn't go all in on Tesla (Pascal architecture) yet. These look amazing.
by braindead_in on 5/10/17, 8:35 PM
So when are the new AWS instances are coming?
by 1024core on 5/10/17, 6:39 PM
FTA: "GV100 supports up to 6 NVLink links at 25 GB/s for a total of 300 GB/s."
The math doesn't add up.
by Etheryte on 5/11/17, 11:53 AM
Interesting to note that Nvidia's stock rose about 18% (!, 102.94USD on May 9, 121.29USD on May 10) in a single day after this announcement. I expected the market to react, but this seems disproportionate.
by boulos on 5/11/17, 12:10 PM
My favorite outcome of Volta is that it's the first GPU they've produced that actually can claim this SIMT thing due to its separate program counters (we had a spirited debate about whether or not just doing masking but presenting the programming model meant the chip was SIMT or just that CUDA was but GPUs weren't).
by Athas on 5/10/17, 9:01 PM
Does this architecture improve on 64-bit integer performance? Have any of the GPU manufacturers said anything about that? At some point it becomes a necessity for address calculations on large arrays.
by caenorst on 5/10/17, 6:30 PM
Did they communicate any release date and price during the show ?
by gwbas1c on 5/10/17, 7:36 PM
How long until Tesla sues for trademark infringement? "from detecting lanes on the road to teaching autonomous cars to drive" makes it sound like there is an awful lot of overlap in product function.