Subject: Editorial | May 10, 2017 - 09:45 PM | Josh Walrath
Tagged: nvidia, earnings, revenues, Q1 2018, Q1, v100, data center, automotive, gpu, gtx 1080 ti
NVIDIA had a monster Q1. The quarter before the company had their highest revenue numbers in the history of the company. Q1 can be a slightly more difficult time and typically the second weakest quarter of the year. The Holiday rush is over and the market slows down. For NVIDIA, this was not exactly the case. While NVIDIA made $2.173 billion in Q4 2017, they came remarkably close to that with revenues of $1.937 billion. While $250 million is a significant drop, it is not an unexpected one. In fact, it shows NVIDIA being slightly stronger than expectations.
The past year has shown tremendous growth for NVIDIA. Their GPUs remain strong and they have the highest performing parts at the upper midrange and high end markets. AMD simply has not been able to compete with NVIDIA, much less overcome the company with higher performing parts at the top end. GPUs still make up the largest portion of income that NVIDIA receives. NVIDIA continues to invest in new areas and those investments are starting to pay off.
Automotive is still in the growth stages for the company, but they have successfully taken the Tegra CPU division and moved away from the cellphone and tablet markets. NVIDIA continues to support their Shield products, but the main focus looks to be the automotive industry with these high performing, low power parts that sport advanced graphical options. Professional graphics continues to be a stronghold for NVIDIA. While it did drop quite a bit from the previous quarter, it is a high margin area that helps bolster revenues.
The biggest mover over this past year seems to be the Data Center. Last year NVIDIA focused on delivering entire solutions to the market as well as their individual GPUs. The past two years have seen them have essentially no income in this area to having a $400 million quarter. This is simply tremendous growth in an area that is still relatively untapped when it comes to GPU compute.
NVIDIA continues to be very aggressive in their product design and introductions. They have simply owned the $300+ range of graphics cards with the GTX 1070, GTX 1080, and the recently introduced GTX 1080 Ti. This is somewhat ignoring the even higher end TitanXp that is priced well above most enthusiasts’ budgets. Today they announced the V100 chip that is the first glimpse we have of a high end part running on TSMC’s new 12nm FinFET process. It also features 16 GB of HBM2 memory and a whopping 21 billion transistors in total.
Next quarter looks to be even better than this one, which is a shock because Q2 has traditionally been the slowest quarter of the year. NVIDIA expects around $1.95 billion in revenues (actually increasing from Q1). NVIDIA also is rewarding shareholders with not only a quarterly dividend, but also has been actively buying back shares (which tends to keep share prices healthy). Early last year NVIDIA had a share price of around $30 while today they are trending well above $100.
If NVIDIA keeps this up while continuing to expand in automotive and data center, it is a fairly safe bet that they will easily overtop $8 billion in revenues for the year. Q3 and Q4 will be stronger if they continue to advance in those areas while retaining marketshare in the GPU market. With rumors hinting that AMD will not have a product that will top the GTX 1080Ti, it is a safe bet that NVIDIA can easily adjust their prices across the board to stay competitive with whatever AMD throws at them.
It is interesting to look back when AMD was shopping around for a graphics firm and wonder what could have happened. Hector Ruiz was in charge of AMD and tried to leverage a deal with NVIDIA. Rumors have it that Huang would not agree to it unless he was CEO. Hector laughed and talked to ATI who was more than happy to sell (and cover up some real weaknesses in the company). We all know what happened to Hector and how his policies and actions started the spiral that AMD is only now recovering from. What would that have been like if Jensen had actually become CEO of that merged company?
Subject: Graphics Cards | May 10, 2017 - 01:32 PM | Ryan Shrout
Tagged: v100, tesla, nvidia, gv100, gtc 2017
During the opening keynote to NVIDIA’s GPU Technology Conference, CEO Jen-Hsun Huang formally unveiled the latest GPU architecture and the first product based on it. The Tesla V100 accelerator is based on the Volta GPU architecture and features some amazingly impressive specifications. Let’s take a look.
|Tesla V100||GTX 1080 Ti||Titan X (Pascal)||GTX 1080||GTX 980 Ti||TITAN X||GTX 980||R9 Fury X||R9 Fury|
|GPU||GV100||GP102||GP102||GP104||GM200||GM200||GM204||Fiji XT||Fiji Pro|
|Base Clock||-||1480 MHz||1417 MHz||1607 MHz||1000 MHz||1000 MHz||1126 MHz||1050 MHz||1000 MHz|
|Boost Clock||1455 MHz||1582 MHz||1480 MHz||1733 MHz||1076 MHz||1089 MHz||1216 MHz||-||-|
|ROP Units||128 (?)||88||96||64||96||96||64||64||64|
|Memory Clock||878 MHz (?)||11000 MHz||10000 MHz||10000 MHz||7000 MHz||7000 MHz||7000 MHz||500 MHz||500 MHz|
|Memory Interface||4096-bit (HBM2)||352-bit||384-bit G5X||256-bit G5X||384-bit||384-bit||256-bit||4096-bit (HBM)||4096-bit (HBM)|
|Memory Bandwidth||900 GB/s||484 GB/s||480 GB/s||320 GB/s||336 GB/s||336 GB/s||224 GB/s||512 GB/s||512 GB/s|
|TDP||300 watts||250 watts||250 watts||180 watts||250 watts||250 watts||165 watts||275 watts||275 watts|
|Peak Compute||15 TFLOPS||10.6 TFLOPS||10.1 TFLOPS||8.2 TFLOPS||5.63 TFLOPS||6.14 TFLOPS||4.61 TFLOPS||8.60 TFLOPS||7.20 TFLOPS|
While we are low on details today, it appears that the fundamental compute units of Volta are similar to that of Pascal. The GV100 has 80 SMs with 40 TPCs and 5120 total CUDA cores, a 42% increase over the GP100 GPU used on the Tesla P100 and 42% more than the GP102 GPU used on the GeForce GTX 1080 Ti. The structure of the GPU remains the same GP100 with the CUDA cores organized as 64 single precision (FP32) per SM and 32 double precision (FP64) per SM.
Click to Enlarge
Interestingly, NVIDIA has already told us the clock speed of this new product as well, coming in at 1455 MHz Boost, more than 100 MHz lower than the GeForce GTX 1080 Ti and 25 MHz lower than the Tesla P100.
Click to Enlarge
Volta adds in support for a brand new compute unit though, known as Tensor Cores. With 640 of these on the GPU die, NVIDIA directly targets the neural network and deep learning fields. If this is your first time hearing about Tensor, you should read up on its influence on the hardware markets, bringing forth an open-source software library for machine learning. Google has invested in a Tensor-specific processor already, and now NVIDIA throws its hat in the ring.
Adding Tensor Cores to Volta allows the GPU to do mass processing for deep learning, on the order of a 12x improvement over Pascal’s capabilities using CUDA cores only.
For users interested in standard usage models, including gaming, the GV100 GPU offers 1.5x improvement in FP32 computing, up to 15 TFLOPS of theoretical performance and 7.5 TFLOPS of FP64. Other relevant specifications include 320 texture units, a 4096-bit HBM2 memory interface and 16GB of memory on-module. NVIDIA claims a memory bandwidth of 900 GB/s which works out to 878 MHz per stack.
Maybe more impressive is the transistor count: 21.1 BILLION! NVIDIA claims that this is the largest chip you can make physically with today’s technology. Considering it is being built on TSMC's 12nm FinFET technology and has an 815 mm2 die size, I see no reason to doubt them.
Shipping is scheduled for Q3 for Tesla V100 – at least that is when NVIDIA is promising the DXG-1 system using the chip is promised to developers.
I know many of you are interested in the gaming implications and timelines – sorry, I don’t have an answer for you yet. I will say that the bump from 10.6 TFLOPS to 15 TFLOPS is an impressive boost! But if the server variant of Volta isn’t due until Q3 of this year, I find it hard to think NVIDIA would bring the consumer version out faster than that. And whether or not NVIDIA offers gamers the chip with non-HBM2 memory is still a question mark for me and could directly impact performance and timing.