Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

El Capitan: New supercomputer is the fastest

spectrum.ieee.org

Post reply on HN
Print view View original

21–30 of 106 posts

Re: El Capitan: New supercomputer is the fastest

#21

Post by amelius »

JumpCrisscross wrote: post #16
olao99 wrote: post #7

I fail to understand how these nuclear bomb simulations require so much compute power. Are they trying to model every single atom? Is this a case where the physicists in charge get away with programming the most inefficient models possible and then the administration simply replies "oh I guess we'll need a bigger supercomputer"

> Are they trying to model every single atom? Given all nuclear physics happens inside atoms, I'd hope they're being more precise. Note that a frontier of fusion physics is characterising plasma flows. So even at the atom-by-atom level, we're nowhere close to a solved problem.

Or maybe it suffices to model the whole thing as a gas. It all depends on what they're trying to compute.

Re: El Capitan: New supercomputer is the fastest

#22

Post by JumpCrisscross »

amelius wrote: post #21
JumpCrisscross wrote: post #16

Earlier quoted context omitted.

> Are they trying to model every single atom? Given all nuclear physics happens inside atoms, I'd hope they're being more precise. Note that a frontier of fusion physics is characterising plasma flows. So even at the atom-by-atom level, we're nowhere close to a solved problem.

Or maybe it suffices to model the whole thing as a gas. It all depends on what they're trying to compute.

> maybe it suffices to model the whole thing as a gas

What are you basing this on? Plasmas don't flow like gases even absent a magnetic field. They're self interacting, even in supersonic modes. This is like saying you can just model gases like liquids when trying to describe a plane--they're different states of matter.

Re: El Capitan: New supercomputer is the fastest

#23

Post by pama »

Noting here that 2700 quadrillion operations per second is less than the estimated sustained throughput of productive bfloat16 compute during the training of the large llama3 models, which IIRC was about 45% of 16,000 quadrillion operations per second, ie 16k H100 in parallel at about 0.45 MFU. The compute power of national labs has fallen far behind industry in recent years.

Re: El Capitan: New supercomputer is the fastest

#24

Post by alephnerd »

MobiusHorizons wrote: post #3

> El Capitan, housed at Lawrence Livermore National Laboratory in Livermore, Calif., can perform over 2700 quadrillion operations per second at its peak. The previous record holder, Frontier, could do just over 2000 quadrillion peak operations per second. > El Capitan uses AMD’s MI300a chip, dubbed an accelerated processing unit, which combines a CPU and GPU in one package. In total, the system boasts 44,544 MI300As,…

> Seems like a nice win for AMD

Yep! They've been part of the Exascale project for a long time, and it's good to see their commitment on HPC actually succeeded unlike Intel's during the same time period.

Re: El Capitan: New supercomputer is the fastest

#25

Post by handfuloflight »

pama wrote: post #23

Noting here that 2700 quadrillion operations per second is less than the estimated sustained throughput of productive bfloat16 compute during the training of the large llama3 models, which IIRC was about 45% of 16,000 quadrillion operations per second, ie 16k H100 in parallel at about 0.45 MFU. The compute power of national labs has fallen far behind industry in recent years.

Any idea how that stacks up with GPT-4?

Re: El Capitan: New supercomputer is the fastest

#26

Post by alephnerd »

pama wrote: post #23

Noting here that 2700 quadrillion operations per second is less than the estimated sustained throughput of productive bfloat16 compute during the training of the large llama3 models, which IIRC was about 45% of 16,000 quadrillion operations per second, ie 16k H100 in parallel at about 0.45 MFU. The compute power of national labs has fallen far behind industry in recent years.

Training an LLM (basically Transformers) is different workflow from Nuclear Simulations (basically Monte Carlo simulations)

There are a lot of intricates, but at a high level they require different compute approaches.

Re: El Capitan: New supercomputer is the fastest

#27

Post by pkaye »

p_l wrote: post #15
olao99 wrote: post #7

I fail to understand how these nuclear bomb simulations require so much compute power. Are they trying to model every single atom? Is this a case where the physicists in charge get away with programming the most inefficient models possible and then the administration simply replies "oh I guess we'll need a bigger supercomputer"

It literally requires simulating each subatomic particle, individually. The increases of compute power have been used for twin goals of reducing simulation time (letting you run more simulations) and to increase the size and resolution. The alternative is to literally build and detonate a bomb to get empirical data on given design, which might have problems with replicability (important when applying the results to r…

Are they always designing new nuclear bombs? Why the ongoing work to simulate?

Re: El Capitan: New supercomputer is the fastest

#28

Post by danhon »

pkaye wrote: post #27
p_l wrote: post #15

Earlier quoted context omitted.

It literally requires simulating each subatomic particle, individually. The increases of compute power have been used for twin goals of reducing simulation time (letting you run more simulations) and to increase the size and resolution. The alternative is to literally build and detonate a bomb to get empirical data on given design, which might have problems with replicability (important when applying the results to r…

Are they always designing new nuclear bombs? Why the ongoing work to simulate?

It's also to check that the ones they have will still work, now that there are test bans.

Re: El Capitan: New supercomputer is the fastest

#29

Post by handfuloflight »

alephnerd wrote: post #26
pama wrote: post #23

Noting here that 2700 quadrillion operations per second is less than the estimated sustained throughput of productive bfloat16 compute during the training of the large llama3 models, which IIRC was about 45% of 16,000 quadrillion operations per second, ie 16k H100 in parallel at about 0.45 MFU. The compute power of national labs has fallen far behind industry in recent years.

Training an LLM (basically Transformers) is different workflow from Nuclear Simulations (basically Monte Carlo simulations) There are a lot of intricates, but at a high level they require different compute approaches.

Can you expand on why the operations per second is not an apt comparison?

Re: El Capitan: New supercomputer is the fastest

#30

Post by bryanlarsen »

pama wrote: post #23

Noting here that 2700 quadrillion operations per second is less than the estimated sustained throughput of productive bfloat16 compute during the training of the large llama3 models, which IIRC was about 45% of 16,000 quadrillion operations per second, ie 16k H100 in parallel at about 0.45 MFU. The compute power of national labs has fallen far behind industry in recent years.

A 64 bit float operation is >4X as expensive as a 16 bit float operation.
Post reply on HN