Professor of Networked Systems at UCL and Networks at OpenAI
Joined May 2012
- Tweets 1,155
- Following 115
- Followers 11,040
- Likes 865
245 Photos and videos
Mark Handley retweeted
There is a lot of news about compute being the bottleneck for AI. There is less visibility into the engineering it takes to make large-scale compute actually work reliably.
In my view, this is one of the most interesting computer science problems in the industry right now. It is not just about getting more GPUs. It is about making every layer of the system work: networking, scheduling, hardware health, storage, orchestration, reliability, observability, security, and the developer experience for researchers.
This blog gives a rare preview into the depth of engineering happening across the stack at OpenAI, starting with MRC and supercomputer networking. We're excited to start sharing more about designing, building, and operating compute at planet scale.
openai.com/index/mrc-superco…
Join us: openai.com/careers/software-…
AI supercomputers need a new kind of network to stay in sync at massive scale.
OpenAI’s @markjhandley and @poyntingatgreg join @AndrewMayne to discuss what it takes to move data across record numbers of chips reliably and efficiently, the new Multipath Reliable Connection (MRC) networking protocol, and why it's available for the whole industry to use.
13
26
468
879,902
AI supercomputers need a new kind of network to stay in sync at massive scale.
OpenAI’s @markjhandley and @poyntingatgreg join @AndrewMayne to discuss what it takes to move data across record numbers of chips reliably and efficiently, the new Multipath Reliable Connection (MRC) networking protocol, and why it's available for the whole industry to use.
We’ve partnered with @AMD, @Broadcom, @Intel, @Microsoft, and @NVIDIA, to release Multipath Reliable Connection (MRC), a new open networking protocol that helps large AI training clusters run faster and more reliably, with less wasted GPU time.
openai.com/index/mrc-superco…
142
152
1,640
1,104,971
We’ve partnered with @AMD, @Broadcom, @Intel, @Microsoft, and @NVIDIA, to release Multipath Reliable Connection (MRC), a new open networking protocol that helps large AI training clusters run faster and more reliably, with less wasted GPU time.
openai.com/index/mrc-superco…
214
698
6,024
1,101,850
Excited to be able to share what I've been working on for the last few years!
May 6
Today we shared MRC (openai.com/index/mrc-superco…), a networking protocol developed with @Microsoft, @nvidia, @AMD, @Broadcom, and @intel to improve how large AI training systems move data and recover from failures. This innovation has come full circle for me personally, it was initiated by @OpenAI with my team at @intel then when I was leading the networking business there and it's great to see it come to life at scale!
As training clusters scale, networking becomes a critical part of overall compute efficiency. It is not enough to add more capacity. You also need systems that keep jobs running reliably, use bandwidth well, and reduce wasted GPU time.
MRC is one example of the kind of infrastructure work required to make frontier model training more efficient and more resilient. It reflects a broader view we have at OpenAI: progress in AI depends not just on better models, but on better compute systems across the stack.
1
1
8
604
Mark Handley retweeted
3 Oct 2024
When I became a wheelchair user, I thought coastline trails had become inaccessible to me. Fortunately, I was wrong!
On a three day adventure in North Wales, I explored some of the accessible sections of @WalesCoastPath for @countrylivinguk:
countryliving.com/uk/wildlif…
1
10
40
2,448
Mark Handley retweeted
28 Jun 2024
this is the alcohol / politics / statistics crossover we’ve all been waiting for
13
18
90
19,657
10 May 2024
These are taken on a Oneplus Nord 2T in manual mode: 30 second exposure, 100 iso. I'm really impressed what this 2 year old mid-range phone can do!
2
2
1,189
Mark Handley retweeted
29 Mar 2024
@rossjanderson Professor Ross Anderson, FRS, FREng Dear friend and treasured long term campaigner for privacy and security, Professor of Security Engineering at Cambridge University and Edinburgh University, Lovelace Medal winner, has died suddenly at home in Cambridge.
73
291
803
487,742
Mark Handley retweeted
17 Mar 2024
My uncle @MarkJHandley turned up this weekend with an all-terrain wheelchair he's designed & made!
Can't express what it means to be back on my favourite trails.
Getting my freedom back one muddy path at a time 💛
ALT Beth, a young woman with brown hair, is sat in an all terrain wheelchair whizzing along a muddy path
ALT Beth, a woman in her 20s with brown hair, is covered in mud and giving a thumbs up
5
6
78
3,758
28 Dec 2023
Suspiciously long journey our rental car did on the 25th. Either the sleigh wasn't working this year, or Kia's software has a bug.
1
6
2,819
6 Oct 2023
Just realised it's more than three years since I fixed a failed Mac Mini by baking it in the oven to reflow the solder. Surprised and happy to report it's still going strong (mostly used as an Octoprint server for my 3d printer). #cantbelieveitstillworks
2 Apr 2020
When needs must: repaired failed Mac by heating logic board in the oven at 200C for 8 minutes to reflow solder. Typing this on it now. #cantbelieveitworked
1
1
11
4,882
3 Oct 2023
Got this Bard answer via @larrypress. I've done space nets research, but:
- This paper doesn't exist
- Space Communications last issue was 2013
- I've never been at Bristol Uni
- I don't work on optics
But, #HeyGoogle, could you hallucinate citations for it in Google Scholar...
4
1,602
Mark Handley retweeted
28 Apr 2023
In the past 10 years, Japan has moved at a tremendous pace to install elevators & escalators in subway and train stations for the disabled and the elderly. These wheelchair accessible escalators are unique to Japan
[read more: buff.ly/3G1WQXW]
73
859
5,145
533,186
26 Mar 2023
The ILCAs seem oblivious to the cuckoo in their nest @UKLA_ILCA_UK
1
1,488