The owner’s name, followed by the cluster name used in the most official announcement.
If the name or the owner’s name is unknown, then other identifying information may be used to distinguish the supercomputer (such as location, year, chip type, an index, etc).
Supercomputers that have several phases are notated by “Phase 1”, “Phase 2”, etc

Oak Ridge NL Frontier

Existing: The supercomputer (as described in this entry) is believed to be operational (>80% of the GPUs available for training)
Planned: The supercomputer is not yet operational, and this entry describes plans for when it will be operational.
Decomissioned: The supercomputer used to exist, but has since been decomissioned.

Existing

How likely does (or will) this cluster exist as roughly as specified?
Confirmed: Official announcement that seems highly trustworthy or expert confirmation. Corresponds to >90% certainty.
Likely: Cluster most likely exists/ will exist roughly in the specified form. 50-90% certainty.
Unlikely: Cluster probably doesn’t/won’t exist roughly in the specified form. 10-50% certainty
Confirmed false: Cluster does not/ will not exist roughly in the specified form. Use e.g. if an expert or media report revealed earlier announcements were wrong. (90% or more likely the cluster does not exist in roughly the specified form)

Confirmed

Yes: The source clearly implies a single supercomputers deployed on a continuous campus without significant physical separation, connected by high-bandwidth networking fabric, forming a single system. There is no evidence otherwise
Unclear: No compelling evidence either way.
No: The cluster the source refers to is, in fact, several distinct supercomputers.

Yes

The primary AI chip used in the supercomputer. This links to a specific chip in the Epoch ML Hardware Database.

AMD Radeon Instinct MI250X

Total number of the primary AI chips in the supercomputer.
A GB200 is counted as two AI chips for this dataset.

37,632

Note to describe the all the AI chips in the supercomputer.

AMD MI250X

The base 10 logrithm of the maximum theorhetical operations per second the superocmputer can achieve (out of 32, 16, or 8 bit performance)

19.158756074028645

Base 10 logrithm of the maximum theorhetical performance of the supercomputer in 16-bit FLOP/s

19.158756074028645

Divides the supercomputer’s Max OP/s field (in 32, 16, or 8 bit) by an H100’s FP8 FLOP/s. Note, H100 equivalents isn’t a well defined measurement, and this number should be treated more as a “sense” instead of a precise measurement. It works fairly well for comparisons among supercomputers since ~2021, but makes less sense for older supercomputers.

7,283

When the supercomputer was first fully operational (ie, you could run a workload on at least 80% of the cluster). This will often be an approximation to the nearest month or so, since the exact data a supercomputer became operational is rarely announced. By default, we are conservative with this date, and set it to the first date that we have confirmation that it was operational, which will often be a few months after it was actually first operational.
This data field is of date type, and will automatically be calculated based on the text in the “First operational note” field

2022-05-30

Country in which the supercomputer is physically located.

United States of America

The entity that owns the AI supercomputer itself (the hardware within the datacenter). This can include several entity if it is a joint venture. This can sometimes differ from the entity that owns the datacenter itself, or the entity that has a long term arrangement to rent the supercomputer.

US Department of Energy

Private: Owned by a commercial entity.
Public: Owned by a government or university
Public/Private: A colaboration between private company and government or university if actors from the public and private sectors have both paid for at least 25% of the supercomputer.

Public

The peak power capacity of the system, in megawatts. Will be the reported power capacity if available, and the calculated power capacity if not.

40

The cost to acquire the hardware for the system in 2025 US dollars. Includes the cost for the GPUs, CPUs, networking, etc, but not the datacenter itself. Will be the reported hardware cost if it exists, and the calculated hardware cost otherwise.

$620,445,966.19

Log of the 16-bit FLOP/s per watt. The numerator is the theorhetical max 16-bit FLOP/s of the system, and the denomenator is the peak power capacity of the system.

360326400000

This supercomputer’s rank in the world by performance (FLOP/s, maximum of 32, 16, or 8 bit precision) the day it first became operational

2

Specific location where the supercomputer is physicaly located.

Oak Ridge National Laboratory 5200, 1 Bethel Valley Rd, Oak Ridge, TN 37830

Any significant entities known to use the supercomputer. “Cloud” refers to the supercomputer being offered publicly via the cloud. Will often just be the owner.

US Government, Academia, Industry

A quote from the source specifying the number of and type of chips and/or the reported or manual FLOP/s number if applicable.

The system has 74 Olympus rack HPE cabinets, each with 128 AMD compute nodes, and a total of 9,408 AMD compute nodes… Each Frontier compute node consists of [1x] 64-core AMD “Optimized 3rd Gen EPYC” CPU (with 2 hardware threads per physical core) with access to 512 GB of DDR4 memory. Each node also contains [4x] AMD MI250X

Notes that give additional context or helpful information about the supercomputer.

(9408 nodes)x(4 GPUs/node) = 37,632 GPUs

When the supercomputer was first fully operational (ie, you could run a workload on at least 80% of the cluster). This will often be an approximation to the nearest month or so, since the exact data a supercomputer became operational is rarely announced. By default, we are conservative with this date, and set it to the first date that we have confirmation that it was operational, which will often be a few months after it was actually first operational.
Will be “Planned” or “Planned for [date]” if supercomputer in construction or planned

2022-05-30

A briefly explanation of why the given certainty level was selected

Details released on official government website

If a supercomputer was built in multiple phases, this links to the supercomputer corresponding to the preceding phase.

If a supercomputer has a later phase (eg, was upgraded later in time), and that phase is already operational, this links to the supercomputer entry corresponding to that phase

True if we think there is a >20% chance that this supercomputer is a duplicate of a supercomputer that already exists in this database. In circumstances where we think two entries might correspond to the same supercomputer, we will mark this field False for the entry with which we have more information, and True for the entry with which we have less information (this allows the user to filter out all entries with “Possible Duplicate”=True for analysis)

If “Possible duplicate” is marked True, this will link to the entry that we think this supercomputer is a duplicate of

The secondary AI chip used in the supercomputer, if there was any. This links to a specific chip in the Epoch ML Hardware Database.

Total number of the secondary AI chips in the supercomputer, if there are any.
A GB200 is counted as two AI chips for this dataset.

NaN

Corresponds to the number of AI chips in the supercomputer. Maxium of {Primary AI chip quantity + Secondary AI chip quantity; Manual total chips number; number of chips calculated from FLOP/s}

37,632

The GPU supplier for the Primary AI Chip type

AMD

The GPU supplier for the Secondary AI Chip type (if applicable)

True if the entry should be excluded from analyses. This could be becase it falls below the inclusion threshold, or there is another reason for excluding it. False otherwise.

True if this entry qualified to be used in the analyses we used for the original paper. That is, (Status=“Existing” or “Decomissioned”) AND (Exclude=False) AND (Certainty = “Confirmed” or “Likely”) AND (Possible Duplicate=False) AND (Single Cluster?=“Yes”)
False otherwise

True

The maximum theorhetical operations per second the superocmputer can achieve (out of 32, 16, or 8 bit performance)

1.4e+19

The maximum theorhetical operations per second the superocmputer can achieve in 8-bit numerical precision

1.4e+19

The maximum theorhetical operations per second the superocmputer can achieve in 16-bit numerical precision

1.4e+19

The maximum theorhetical operations per second the superocmputer can achieve in 32-bit numerical precision

3.6e+18

The calculated power capacity of the supercomputer in megawatts based on the number and type of chips. See paper for full methodology

40.066790399999995

Reported power capacity of the supercomputer in megawatts

40

Calculates the cost (in 2025 US dollars) based on number and type of chips. See paper for full methodology

$967,488,702.64

Reported total cost of the supercomputer in US dollars at the time.

$600,000,000

Reported total cost of the supercomputer, adjusted to 2025 US dollars

$620,445,966.19

Quote from the source saying how much the supercomputer cost. Might refer to a currency besides dollars, or not yet be inflation adjusted

$600M

The name of the largest known supercomputer existing on the first operaitonal date of this supercomputer

Microsoft GPT-4 cluster

The maximum theorhetical FLOP/s (32, 16, or 8 bit) for the this supercomputer, divided by the maximum theorhetical FLOP/s for the largest known supercomputer existing at the time.

0.9239138461538378

URL for source on the AI supercomputer. Generally saved as a Wayback Machine link to preserve the link

https://web.archive.org/web/20240720224959/https://docs.olcf.ornl.gov/systems/frontier_user_guide.html