Glenn K. Lockwood
banner
glennklockwood.mast.hpc.social.ap.brid.gy
Glenn K. Lockwood
@glennklockwood.mast.hpc.social.ap.brid.gy
I am a supercomputing enthusiast, but I usually don't know what I'm talking about. I post about large-scale infrastructure for #HPC and #AI.

🌉 bridged from https://mast.hpc.social/@glennklockwood on the fediverse by https://fed.brid.gy/
Andreas Dilger is now working for The Lustre Collective (https://thelustrecollective.com) after leaving DDN. I am glad to see his leadership continue to drive Lustre into the future. Say what you will about it, Lustre is the standard to which every other #hpc file system is compared.

#sc25
November 18, 2025 at 11:51 PM
When I spent a Thanksgiving week after SC writing the non-MPI layer for Darshan years ago, I thought to myself “surely this work will make me famous!”

I guess my ship finally came in at the PDSW keynote by Rob Ross.

#sc25
November 17, 2025 at 3:27 PM
Chatting with a pal reminded me of a fun pre-SC activity: looking back at old conference takes that aged like milk. Remember this one?

#hpc #zettascale #hedoesntworkthereanymore
November 14, 2025 at 10:40 PM
🎶 It's the most wonderful time of the year 🎶

#sc25
November 14, 2025 at 8:22 PM
I’m in North Dakota this week celebrating Scandinavian heritage.

That liquid on top is butter.
September 24, 2025 at 10:32 PM
Visited some of my old pals at NERSC this afternoon. It feels like a lot has changed, but at the same time, it’s the same old place. I do miss working here, but am glad I now get a chance to visit under the auspices of my new role at VAST.
September 19, 2025 at 11:51 PM
Getting on a plane for VAST for the first time on Monday. But I’m not so sure about the destination 🥵
August 9, 2025 at 5:04 PM
GPT-5 is cool and all, but it's still not to be trusted. Didn't take me long to get "PhD-level" made-up facts.

What really distinguishes "PhD-level" from whatever GPT-4 was is the ability to critically think, be skeptical, and stand behind the information […]

[Original post on mast.hpc.social]
August 9, 2025 at 12:52 AM
One thing I really enjoy about working at VAST (or perhaps that I enjoy about not working at Microsoft) is that I can go out and talk to people again as part of my job. Here’s a view from where I got to spend my afternoon today.

#nothpc #butthatsok
August 5, 2025 at 5:21 AM
I am a sucker for photos of cool #hpc infrastructure, and here is a dense GB200 NVL72 cluster going up somewhere in Canada (I think). Impressive to see this many racks in row; the DC must have facility water which is still uncommon in hyperscale. Source […]

[Original post on mast.hpc.social]
July 16, 2025 at 4:46 PM
In the last 18 hours, I’ve learned way more about adjuvanted vaccines and adverse reactions to them in cats than I ever cared to. And the real kick is that the choice to re-up the cat’s vaccines was an afterthought, because we had to take other cat in for […]

[Original post on mast.hpc.social]
July 8, 2025 at 5:34 PM
Photos of a new, big, naked Cerebras cluster in Oklahoma appearing on the socials today. Pretty neat. Wonder if this is another G42 install.
June 28, 2025 at 5:10 PM
The #isc25 conference has a strong keynote speaker lineup from around the world this week. Looking forward to hearing all three.
June 10, 2025 at 7:16 AM
Here's a photo of an Azure datacenter coming online later this year that "will run a single, large cluster of hundreds of thousands of interconnected NVIDIA GB200 GPUs," "exabytes of storage," and "millions of CPU compute cores."

Source […]

[Original post on mast.hpc.social]
May 27, 2025 at 8:50 PM
Cleaning out old photos and I found this blast from the past. ISC19 specifically. My how times have changed.

Not shown: the slide that followed talking up Optane DC PMM.
May 8, 2025 at 4:36 AM
And here is the silicon photonics switch that was announced at #gtc25. The laser modules slot in along the top and look like familiar transceivers, just without the fibers. Ring resonators modulate the laser light out of the MPOs(?) that connect directly to the switch below the lasers. Wild.
March 21, 2025 at 6:17 AM
Here is the business end of a proof of concept for NVIDIA’s NVL576 rack, dubbed Kyber, on display at #gtc25. Compute blades in the front, NVLink Switch blades in the back, and a midplane. Reminiscent of Cray EX, but much more dense.
March 21, 2025 at 6:08 AM
A hotel breakfast worthy of #gtc25: Keurig coffee and INNOVATION!
March 20, 2025 at 1:48 PM
Right behind the entrance to #gtc25 is what appears to be an anti-antitrust wall. Look at all the companies who are manufacturing these identical, bespoke components for an NVL72 rack!
March 19, 2025 at 4:42 AM
If you pose the question of “how will we get to training on a million GPUs,” you skip the part where you question whether that’s even a reasonable or well-formed question.
March 19, 2025 at 4:41 AM
GPU node fault distributions from Crusoe. Pretty consistent with my experience, but actually presented publicly. Nice.

#gtc25
March 18, 2025 at 11:40 PM
GPU node fault distributions from Crusoe. Pretty consistent with my experience, but actually presented publicly. Nice.

#gtc25
March 18, 2025 at 10:28 PM
20 minutes into a talk by Ian Buck and his focus has been almost exclusively scientific computing. Little/no mention of AI, in sharp contrast to Jensen’s keynote. Surprising.

#gtc25
March 18, 2025 at 9:23 PM
I've been playing with OpenAI's Deep Research, and I asked it for help in figuring out what I want to do with my life. It wound up telling me to learn from this guy named Glenn who was in a similar position.

#ffffffff
March 1, 2025 at 1:07 AM
I am not an expert, but I find this news very confusing. The Nature paper accompanying this announcement contains a bunch of statements like this which appear to contradict the claims in this blog post. The paper only shows "a necessary ingredient of […]

[Original post on mast.hpc.social]
February 19, 2025 at 6:22 PM