HackerTrans
TopNewTrendsCommentsPastAskShowJobs

alechammond

33 karmajoined 3 lata temu

Submissions

Who Needs DRAM? We Have Fiber

arxiv.org
12 points·by alechammond·przedwczoraj·2 comments

Scholar Labs: An AI Powered Scholar Search

scholar.googleblog.com
2 points·by alechammond·8 miesięcy temu·0 comments

Magic Leap partners with Google to advance XR

magicleap.com
4 points·by alechammond·2 lata temu·0 comments

Nvidia Blackwell Arch Technical Brief

nvdam.widen.net
5 points·by alechammond·2 lata temu·0 comments

A gymnasium for machine learning assisted computer architecture design

blog.research.google
2 points·by alechammond·3 lata temu·0 comments

Silicon Photonics Key to Unlocking AI’s Full Potential

eetimes.com
2 points·by alechammond·3 lata temu·0 comments

Build data apps in Julia. Fast

geniecloud.io
1 points·by alechammond·3 lata temu·0 comments

The Bill Is Coming Due for China’s ‘Capitalist’ Experiment

nationalreview.com
13 points·by alechammond·3 lata temu·0 comments

Sophia: Stochastic Second-Order Optimizer for Language Model Pre-Training

arxiv.org
7 points·by alechammond·3 lata temu·0 comments

Meta unveils AI inference accelerator

ai.facebook.com
3 points·by alechammond·3 lata temu·1 comments

comments

alechammond
·2 lata temu·discuss
Thanks for the reply!

> So running 500 sequential kMC simulations would take approximately 125 years - not ideal.

Ah I see. Ya it’s hard to get around this…

> is fundamentally stochastic, and not differentiable

This is actually a common “inverse design pattern” for a variety of applications, and luckily there are tricks to efficiently compute gradients here. In my domain (nanophotonics) we’re often simulating incoherent sources, which are similarly stochastic. But there are ways to reformulate the problem to drastically reduce the number of forward and adjoint solves you need during the design process [1].

That being said, given the cost of the forward problem (3 months per iteration!) this doesn’t help you much…

[1] https://doi.org/10.1007/s00158-022-03389-5
alechammond
·2 lata temu·discuss
Interesting paper, thanks for sharing! Whenever I see physics-based inverse design using ML surrogates, I always ask, “why not optimize the problem directly” (and eg compute the gradient using an adjoint-variable method)? The paper implies that the forward simulation process isn’t differentiable, but is this true? Thanks!
alechammond
·2 lata temu·discuss
> They require significantly different fabrication processes, and we don't know how to fab them into the same chip as electrical ones.

There are actually a few commercial fabs that will monolithically integrate the photonics, analog electronics, and digital electronics, all in the same CMOS process. See for example GF’s process:

https://www.cmc.ca/globalfoundries-fotonix-45spclo/

Integrating good optical sources in silicon remains a challenge, but companies like Intel have mastered hybrid bonding and other packaging techniques. TSMC too has a strong silicon photonics effort.
alechammond
·3 lata temu·discuss
Note that Meta recently added a Julia ruleset to its buck2 prelude set. An important step toward working with Julia on prod environments.

https://github.com/facebook/buck2/tree/main/prelude/julia
alechammond
·3 lata temu·discuss
I wonder why ORNL used AMDGPU.jl directly rather than something like KernelAbstractions.jl, which doesn’t require you to overspecialize to a particular architecture. I realize Frontier is all AMD. But the DOE labs have flipped back and forth between HPC platforms a few times. (Which is partly why Sandia invested so heavily in developing Kokkos).
alechammond
·3 lata temu·discuss
> Greer believes that this is one of the first demonstrations of 3D printing of metal structures at the nanoscale.

I may be missing something, but MIT has been doing something very similar for awhile now. The student (Dan Oran) even spun out a startup.

https://www.media.mit.edu/projects/implosion-fabrication-1/o...