The NVIDIA Titan V Deep Learning Deep Dive: It's All About The Tensor Cores
by Nate Oh on July 3, 2018 10:15 AM ESTBenchmarking Testbed Setup
Our hardware has been modified for deep learning workloads with a larger SSD and more RAM.
CPU: | Intel Core i7-7820X @ 4.3GHz |
Motherboard: | Gigabyte X299 AORUS Gaming 7 |
Power Supply: | Corsair AX860i |
Hard Disk: | Intel 1.1TB |
Memory: | G.Skill TridentZ RGB DDR4-3200 4 x 16GB (15-15-15-35) |
Case: | NZXT Phantom 630 Windowed Edition |
Monitor: | LG 27UD68P-B |
Video Cards: | NVIDIA Titan V NVIDIA Titan Xp NVIDIA GeForce GTX Titan X (Maxwell) AMD Radeon RX Vega 64 |
Video Drivers: | NVIDIA: Release 390.30 for Linux x64 AMD: |
OS: | Ubuntu 16.04.4 LTS |
With deep learning benchmarking requiring some extra hardware, we must give thanks to the following that made this all happen.
Many Thanks To...
Much thanks to our patient colleagues over at Tom's Hardware, for both splitting custody of the Titan V and lending us their Titan Xp and Quadro P6000. None of this would have been possible without their support.
And thank you to G.Skill for providing us with a 64GB set of DDR4 memory suitable for deep learning workloads, not a small feat in these DDR4 price-inflated times. G.Skill has been a long-time supporter of AnandTech over the years, for testing beyond our CPU and motherboard memory reviews. We've reported on their high capacity and high-frequency kits, and every year at Computex G.Skill holds a world overclocking tournament with liquid nitrogen right on the show floor.
Further Reading: AnandTech's Memory Scaling on Haswell Review, with G.Skill DDR3-3000
65 Comments
View All Comments
mode_13h - Monday, July 9, 2018 - link
Nice. You gonna water-cool it?https://www.anandtech.com/show/12483/ekwb-releases...
wumpus - Thursday, July 12, 2018 - link
Don't forget double precision GFLOPS. Just because fp16 is the next new thing, nVidia didn't forget their existing CUDA customers and left out the doubles. I'm not sure what you would really benchmark, billion-point FFTs or something?mode_13h - Thursday, July 12, 2018 - link
Yeah, good point. Since GPUs don't support denormals, you run into the limitations of fp32 much more quickly than on many CPU implementations.I wonder if Nvidia will continue to combine tensor cores AND high-fp64 performance in the same GPUs, or if they'll bifurcate into deep-learning and HPC-centric variants.
byteLAKE - Friday, July 13, 2018 - link
Yes, indeed. Mixed precision does not come out of the box and requires development. We've done some research and actual projects in the space (described here https://medium.com/@marcrojek/how-artificial-intel... and results give a speedup.ballsystemlord - Monday, September 30, 2019 - link
Both myself and techpowerup get 14.90Tflops SP. Can you check your figures?https://www.techpowerup.com/gpu-specs/titan-v.c305...