Enter Dojo: Tesla Reveals Design for Modular Supercomputer & D1 Chip

By Oliver Peckham

August 20, 2021

Two months ago, Tesla revealed a massive GPU cluster that it said was “roughly the number five supercomputer in the world,” and which was just a precursor to Tesla’s real supercomputing moonshot: the long-rumored, little-detailed Dojo system. “We’ve been scaling our neural network training compute dramatically over the last few years,” said Milan Kovac, Tesla’s director of autopilot engineering. “Today, we’re barely shy of ten thousand GPUs. … But that’s not enough.”

Enter Dojo, the design for which was revealed during Tesla’s AI Day event, along with the design for its constituent D1 chip.

Dreaming of Dojo

“There’s an insatiable demand for speed, as well as capacity, for neural network training,” said Ganesh Venkataramanan, Tesla’s senior director for autopilot hardware. To that end, a few years ago, Tesla CEO and co-founder Elon Musk asked Venkataramanan’s team to build the company a “super-fast” training computer, aiming to achieve the best AI training performance, enable larger and more complex neural net models and achieve both power efficiency and cost efficiency.

“For Dojo, we envisioned a large compute plane filled with very robust compute elements, packed with a large pool of memory, and interconnected with very high-bandwidth and low-latency fabric,” Venkataramanan said. “We wanted to attack this all the way – top to bottom of the stack – and remove all the bottlenecks at any of these levels.”

Tesla’s D1 chip. Image courtesy of Tesla.

The D1 chip & training tile

The aforementioned precursor cluster relied predominantly on Nvidia’s A100 GPUs for acceleration. Not so for Dojo, which will consist almost entirely of Tesla’s decidedly unique D1 chip. D1, Venkataramanan said, supported FP32, BFP16 (aka bfloat16 or brain floating point) and a new format called CFP8 (“configurable FP8”). Optimized for machine learning workloads, D1 (which consists of 354 “training nodes”) is manufactured using a 7nm process and, at just 645 square millimeters, contains 50 billion transistors. “There is no dark silicon, there is no legacy support,” Venkataramanan said of the chip, designed completely by Tesla’s internal engineers. “This is a pure machine learning machine. … This chip [has] GPU-level compute with CPU-level flexibility.”

The integrated training tile, which contained 25 D1 chips. Image courtesy of Tesla.

Tesla put a strong emphasis on modularity across the hardware. D1 is equipped with 4TBps off-chip bandwidth on each of its lateral edges – all four equipped with connectors – allowing it to connect to and scale with other D1 chips without sacrificing speed.

The next step up is Tesla’s “training tile,” a wedge less than a cubic foot in size that contains 25 of the D1 chips. The training tile operates with similar modularity to the chip itself: power and cooling are conducted through the top of the tile, allowing its four lateral edges to be outfitted with high-output connectors designed for maximum bandwidth (a total of 36TB/s of off-tile bandwidth).

Dojo

“By now, you must have realized our modularity story is pretty strong,” Venkataramanan said. “We just put together some tiles. We just tile together tiles!” And, indeed, it’s modularity all the way down: 354 training nodes make a D1 chip; 25 D1s in a training tile; six training tiles (2×3) in a “training matrix,” which constitutes a tray; two trays in a cabinet; and ten cabinets in what Venkataramanan calls the ExaPOD, a massive machine learning machine with uniform bandwidth. (How many ExaPODs Dojo will contain is unclear.)

Details on the ExaPOD. Image courtesy of Tesla.

With each D1 chip providing 22.6 teraflops of FP32 performance, each training tile will provide 565 teraflops and each cabinet (containing 12 tiles) will provide 6.78 petaflops – meaning that one ExaPOD alone will deliver a maximum theoretical performance of 67.8 FP32 petaflops. (Tesla preferred to offer performance in BFP16 and CFP8, and by those metrics, an ExaPOD will deliver 1.1 exaflops – hence the name.) All of this performance, Venkataramanan said, will be made accessible through a high-performing compiler, which operates automatically without human involvement and requires minimal effort to initiate on the researchers’ part.

“This is what [Dojo] will be,” Venkataramanan said. “It will be the fastest AI training computer.”

Dojo, however, hasn’t arrived just yet – in fact, Venkataramanan said that the first functional training tile had only arrived the previous week. Next up, he said, they were building the cabinets (“pretty soon”).

“And,” he continued, “we’re not done.” Tesla, Venkataramanan said, had a “whole next-generation plan already,” with their eyes set on the next tenfold increase in performance.

Learn more

Ahead of ‘Dojo,’ Tesla Reveals Its Massive Precursor Supercomputer

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Under The Wire: Nearly HPC News (June 13, 2024)

June 13, 2024

As managing editor of the major global HPC news source, the term "news fire hose" is often mentioned. The analogy is quite correct. In any given week, there are many interesting stories, and only a few ever become headli Read more…

Quantum Tech Sector Hiring Stays Soft

June 13, 2024

New job announcements in the quantum tech sector declined again last month, according to an Quantum Economic Development Consortium (QED-C) report issued last week. “Globally, the number of new, public postings for Qu Read more…

Labs Keep Supercomputers Alive for Ten Years as Vendors Pull Support Early

June 12, 2024

Laboratories are running supercomputers for much longer, beyond the typical lifespan, as vendors prematurely deprecate the hardware and stop providing support. A typical supercomputer lifecycle is about five to six years Read more…

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently king of accelerated computing) wins again, sweeping all nine Read more…

Highlights from GlobusWorld 2024: The Conference for Reimagining Research IT

June 11, 2024

The Globus user conference, now in its 22nd year, brought together over 180 researchers, system administrators, developers, and IT leaders from 55 top research computing centers, national labs, federal agencies, and univ Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst firm TechInsights. Nvidia's GPU shipments in 2023 grew by more Read more…

Under The Wire: Nearly HPC News (June 13, 2024)

June 13, 2024

As managing editor of the major global HPC news source, the term "news fire hose" is often mentioned. The analogy is quite correct. In any given week, there are Read more…

Labs Keep Supercomputers Alive for Ten Years as Vendors Pull Support Early

June 12, 2024

Laboratories are running supercomputers for much longer, beyond the typical lifespan, as vendors prematurely deprecate the hardware and stop providing support. Read more…

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently Read more…

Highlights from GlobusWorld 2024: The Conference for Reimagining Research IT

June 11, 2024

The Globus user conference, now in its 22nd year, brought together over 180 researchers, system administrators, developers, and IT leaders from 55 top research Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

ASC24 Expert Perspective: Dongarra, Hoefler, Yong Lin

June 7, 2024

One of the great things about being at an ASC (Asia Supercomputer Community) cluster competition is getting the chance to interview various industry experts and Read more…

HPC and Climate: Coastal Hurricanes Around the World Are Intensifying Faster

June 6, 2024

Hurricanes are among the world's most destructive natural hazards. Their environment shapes their ability to deliver damage; conditions like warm ocean waters, Read more…

ASC24: The Battle, The Apps, and The Competitors

June 5, 2024

The ASC24 (Asia Supercomputer Community) Student Cluster Competition was one for the ages. More than 350 university teams worked for months in the preliminary competition to earn one of the 25 final competition slots. The winning teams... Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Leading Solution Providers

Contributors

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

GenAI Having Major Impact on Data Culture, Survey Says

February 21, 2024

While 2023 was the year of GenAI, the adoption rates for GenAI did not match expectations. Most organizations are continuing to invest in GenAI but are yet to Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

Intel Plans Falcon Shores 2 GPU Supercomputing Chip for 2026  

August 8, 2023

Intel is planning to onboard a new version of the Falcon Shores chip in 2026, which is code-named Falcon Shores 2. The new product was announced by CEO Pat Gel Read more…

How the Chip Industry is Helping a Battery Company

May 8, 2024

Chip companies, once seen as engineering pure plays, are now at the center of geopolitical intrigue. Chip manufacturing firms, especially TSMC and Intel, have b Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire