LLM n C, 1.6nm, xz vul, turing 2024, let’s encrypt, chatdev, Ethernet vs IB, Slingshot, Tailscale ssh, videos, 42 rules, CNI, Cilium

Origins of deep learning: interesting post. At the beginning all was Matlab and CPU bounded. repo

LLM in C: post and repo.

A16: 1.6nm process for 2026. More frequency, less power.

xz vulnerability repo: Something I need to check in the VP

Turing Award 2024: zero-knowledge-proof.

Cloudflare and Let’s Encrypt’s certificate change: I haven’t heard of this until recently. I use Let’s Encrypt so as far as I can read, makes sense what they are doing. But didnt know 2% Cloudflare customer were using the “cert”

ChatDev: Communicate agents for software development. I am a not a developer but I would use this just as a starting point If I have any idea for a project. I would remove the C-suite agents, at least for low level projects.

IB vs Ethernet: A bit of bias here (the author is from Broadcom -> Ethernet). I have no hands-on experience with IB, but I have read the cables are not cheap… Let’s see when UltraEthernet gets into the market. Another view.

Slingshot and Juniper: A bit of bias again as HP bought Juniper. So how will these interconnects fade inside the company? As far as I know, most supercomputers use some “special” interconnect so not much ethernet there. But the money nowadays is in AI infra… Paper for slingshot (haven’t read it)

Tailscale SSH, wireguard throughput: These are things I should a spend a bit of time one day and consider if I should use them (I dont like it is not opensource though). This netmaker?

Videos:

Jocko Willink: Discipline = Freedom. Remember but not dwell. Good leader, delegate. Be a man -> take action, bonding (pick your activity)

Jimmy Carr: Imposter syndrome each 18 months, so you have to stand-up. People crave the success not the journey. Teaching comedy good for communicating.

Sam Altman – Stanford 2024: First time I see him talking. It has some funny moments. More powerful computers. I missed a question about opensource LLM and closed ones.

Find a girlfriend: I know just a little bit about the person (I want to read one of his books) from other books and videos. I would think he would have already a girlfriend or family. From the three methods, definitely, the face to face approach in the street looks so much better (and that’s what I would like to do)

Jordan Peterson original 42 rules

CNI performance: I have used kubernetes since I studied for CKAD but still I am interested in the networks side. I didn’t know about Kube-router and it did great! I am bit surprised with Calico as I have read more and more about Cilium.

Cilium for network engineers. I have to read this fully (worried that Cisco bought it…)

Blueprint: Build a bulletproof body

I want to get fitter so I try to learn from experts who have proved themselves. And this is a good example. The author has done some amazing things. Quite jealous to be honest.

For me, I just want to get stronger at climbing and get back to (proper) running, all without (more) injuries. I have read before about the different cycles that top athletes so it was interesting to read about it directly and all the science behind that. I know I need to add (more) strength training and endurance. I should create a proper work plan for each week, a bit of less climbing but better prepared? But I am not clear how to use the knowledge from the book to climbing, when, at the end of day, you want to climb hard every week, as I am not going to compete or anything like that. Or saying in a different, how to handle your ego and jealousy.

Season:

  • Recover Mesocycle: low volume, low intensity. This is the chapter I liked the most from the book.

— 2 strength-based rehab routines per week

— 2 endurance-based rehab routines per week at low intensity and no more than 45 minutes.

— 2 days rest

  • Base Mesocycle: increase volume, low intensity

— 3 strength-based rehab routines per week

— 4 low-intensity endurance-based rehab routines per week in zone 2 (aerobic).

— 1 strongman strength session

— 1 day rest

  • Build Mesocycle: reduce a bit volume, increase intensity

— 3 strength and speed sessions per week (force-velocity curve)

— 3 low and long (10km) open water swims operating in zone 2 (aerobic).

— 3 high-intensity interval sessions in zone 4-5 (anaerobic) (separated by 48h)

— 1 day rest

  • Peak Mesocycle: reduce volume, increase intensity:

— 2 strength and speed sessions per week for maintenance of fitness

— 2 low and long (5km) open water swims operating in zone 2 (aerobic). Focus in tecnique.

— 3 high-intensity interval sessions in zone 4-5 (anaerobic)

— 2 day rest

Shoulder pain: Practice hangs from a bar. That’s what our “ancestors” did… Simple

Eudaimonia: fulfilment. It’s different from happiness since it openly accepts that pain and struggling should form part of the process. Happiness without fulfilment is a failure.

The greater the difficulty, the more glory in surmouting it. Skilful pilots gain their reputation from storms and tempests

Epictetus

Askesis: healthy hardship. “The comfort zone is the great enemy to creativity, moving beyond it necessitates intuition, which in turn configures new perspectives and conquers fears” Dan Stevens.

Why we adventure to combat spiritual decay:

A man has achieved his present position by being the most aggressive and enterprising creature on earth….. The comfortable life lowers a man’s resistance… the comfortable life causes spiritual decay

From 1956 book “The Outsider” by Colin Wilson:

Broccoli with Garlic Sauce

I want to eat more broccoli. And this is a new recipe:

Ingredients:
1 head of broccoli
2-3 pieces garlic
small piece ginger
1 tbsp soy sauce
1 tsp sugar (brown if possible)
olive oil
1 tsp corn starch
2 tbsp water
1 tsp sunflower oil
2 tsp paprika
1 tsp sesame seeds

Directions:

  1. Boil 1.5 litres of water in a saucepan
  2. Remove the florets from the broccoli head. Be sure the head is cleaned.
  3. Blench the broccoli in the boiling water for 30-45 sec. Just to see how it gets a bright green color! Remove from water and pass it via cold water. Set aside.
  4. Grate the garlic and ginger.
  5. Prepare the sauce by combining together soy sauce, sugar, oil, corn starch and water. Set the sauce.
  6. Heat up a nonstick pan to medium heat. Add a bit of olive oil.
  7. Cook the broccoli for 2-3min each side. Until you see a bit of char. Put aside
  8. Clean the pan, and put at medium-low heat, add olive oil and paprika (replacement of chili oil). Mix clean so it doesnt burn.
  9. Add the garlic and ginger. Gently cook for 2-3min
  10. Add the broccoli and turn the heat to medium. Sauté for about 1min
  11. Add in the stir fry sauce and if you have a lid, place it immediately so you can steam the broccoli steam for about 30-45 seconds. If not, just toss the broccoli so it gets the sauce spread in all bits.
  12. Plate the broccoli and garnish with white sesame seeds

rsync go, NASA SP287, git options, Undersea cable failures in Africa, Quotes, Log4j, done list, Dan Lynch, Systems-based Productivity, Run Africa

rsync go: Interesting talk about rsync, as it explains how it works and it is something I didnt know. But then, all other things/projects mentioned are cool and related. I need to try to install rsync go in my vm. ccc slides and repo

NASA to the moon: This is an engaging and provocative video regarding the Artemis III (project back to the moon II). He makes some hard questions to the people in charge (I have no clue about physics) and it seems he has a point. Not sure it this will get any effect but again, looks “smart”. When he mention the NASA SP287 (What made Apollo a success) document as the grial for going back to the moon, I wanted to get a copy (here) so I could read it one day.

Git options: Nice post about popular git config options. I am a very basic git user (and still sometimes I screw up) but the options to improve diff looks interesting so I will give it a go at work.

Undersea cable failures in Africa: It is clear that Africa relays heavily in submarine cables (it doesnt look like there are many cable systems intra continent). And the Red Sea is becoming a hot area due to different conflicts…

Quotes: I like the ones regarding simplicity:

A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system. (John Gall)

In programming, simplicity and clarity are a crucial matter that decides between success and failure. (Edsger Dijktra)

Log4j: This is old news but when it came out I tried to run the PoC but I failed 🙁 This is just a reminder. It was annoying because I manged to install all tools but never managed to exploit it.

Done List: I feel totally identified. The to-do list is never done and you feel guilty. Done-list, much healthier.

Dan Lynch: He passed away, and as usual on my ignorance, it seems he is one of the unsung heroes of Internet, migrating ARPANET to TCP/IP.

Systems-Based Productivity: TEMPO refers to five dimensions of productivity: T (Time Management), E (Energy Management), M (Mindset), P (Proficiency) and O (Organization).

Run Africa: very jealous.

Infraops challenge, Devika, Daytona, NTP 2038, Linux Crisis Tools, videos, Chocolonely, LLM, Transformers, Enforce-first

InfraOps challenge: A bit beyond me, but interesting If you could try without applying for the job.

Devika: Agent AI. Another thing I would like to have time to play with it. If you have API keys for some LLMs, looks like it shouldn’t be difficult to run and you dont need a powerful laptop (?)

Daytona: My development environment is a joke, just python envs. But I guess for more serious devs, could be interesting

NTP and year 2038: Agree, when it is not DNS, it is likely NTP (seen this with VPNs and SSL certs in boxes with NTP unsync), or something blocking UDP.

Linux crisis tools: I haven’t got my hands dirty with BPF but I am surprised with so many tools. I would add nc, netstat, lsof, traceroute, ping, vim, openssl etc but because I do pure networks.

Jim Kwik: How to improve your reading speed. One improvement is you use your finger or a ruler. Need to watch again.

Rich Roll: The guy is super chill. I would like to be able to do some ultra at some point in life… Very personal conversation.

Ferran Adria: I didnt know much about the person apart from being one of the best Chefs in history. I like how he starts the interview and take over for 15 minutes. Haven’t watched till the end. But just the beginning is priceless.

Mark Manson: I have read all his books and his emails. Interesting his story.

Chocolonely: I didnt know it was a dutch company and interesting history behind. I want to try one day, but I haven’t found a dark choco version.

LLM in 1000 lines puce C: I was always crap at C. But interesting this project as something educational and intro in LLM.

Visual intro to transformers: The easy joke, unfortunately, this is not about Optimus Prime.

Indonesia Heavy Metal Girls: Unexpected. Respect.

Enforce-first-as: I dint know about this until last week. Cisco defined by default. Juniper disabled by default. And this makes sense with Route Servers.

Feta and Spinach Filo Pie

I haven’t done this for a long time so it was due. I used to use his book quite often. Recipe:

Ingredients:

100g nuts (I used sunflower and pumpkin nuts)
5 large eggs
300g feta cheese – crumbled
50g grated cheese
oregano
1 lemon zest
olive oil
1 knob of butter
500g fresh spinach
1 x 270g pack of filo pastry
cayenne pepper
nutmeg for grating

Process:

  • Preheat the oven to 200°C
  • Toast the nuts in a large ovenproof frying pan over a medium heat until golden.
  • Crack the eggs into a large mixing bowl add feta, grated cheese, oregano and pepper
  • Add the toasted nuts and mix.
  • In the same frying pan, add a bit of olive oil and add half the spinach. Stir until wilted.
  • Add the lemon zest, grated nutmeg and piece of butter to the spinach.
  • Add the rest of spinach. Stir until all is wilted.
  • Then add the spinach to the egg mix. And clean the frying pan.
  • Take a piece of ovenproof paper roughly 1.5 bigger than your pan. Pass it through water so it is wet.
  • Lay it out on a clean work surface, rub lightly with oil and flatten out again.
  • Keep adding sheets of file on top of the ovenproof paper. Add a bit of oil and other spices in each layer. The idea is to cover the frying pan later. I used 6 layers in total.
  • Move the ovenproof paper and filo to the frying pan. Pour the egg/spinach mix. Add a bit more grated cheese.
  • Close the pie folding the edges (as it should be 1.5 bigger than the pan).
  • Fry at medium heat the pie so the bottom is crunchy. 1-2 minutes. Dont burn it!
  • Add some spices and olive oil on top.
  • Then move the pan to the oven for 20 minutes aprox or golden and crisp.

Quite happy with the result:

Narconomics

Very interesting book. It explains the mechanics of a drug cartel from the point of view of economics. I didnt think issues like supply chain, HR/PR, competition/merges, offshoring, R&D, online business, diversification, etc were part of drug cartel, as you only think of those as part of a licit business. There were many things that I didn’t know like the birth of “legal highs” in NZ (and Matt Bowden)

The goal is to fully understand the “business” because the current laws/actions, etc against drugs are clearly not working. So this way you can really offer a different approach to tackle the issue. You are not going to destroy them 100%. Most of the actions are at the source of the drug business: growing the plant (decrease in growing area causes minimum increase in retail price). But the book shows that is not effective and prevention (done in the consumer’s land: like rehab, education in jails, etc) is much more productive (for the same investment). As well as legalization (ie marijuana) as that brings control (“safer drugs”, tax revenue, etc) and put out of the market the dealers/cartels.

This is a difficult pill to swallow (punt intended) for governments and citizens but the writing is in the wall.

GPU Fabrics, Optimizations, Network Acceleration, Learning Cambridge, British Library

Several posts worth reading. There are plenty of things go over my knowledge. I already posted this, it is a good refresher.

GPU Fabrics: The first of the article is the one I am more lost as it about training and the communications between the GPU depending on the take to handle the models. There are several references to improvements as the use of FP8 and different topologies. As well, a bit more clear about NVLink (as internal switch for connecting GPUs inside the same server or rack)

When it moved to the inter-server traffic, I started to understand a bit more things like “rail-optimized” (it is like having a “plane” for my old job where the leaf only connects to a spine instead of all spines, in this case each GPU connects to just one leaf. If you cluster is bigger then you need spines). I am not keen of modular chassis from operations point of view but it is mentioned as an option. Fat-tree CLOS, Dragon-Fly: reminds me to Infiniband. Like all RDMA.

And Fabric congestion it is a big topic with many different approaches: adaptive LB (IB again), several congestion control protocols and mention to Google (CSIG) and Amazon (SDR) implementations.

In general I liked the article because I dont really feel any bias (she works for Juniper) and it is very open with the solutions from different players.

LLM Inference – HW/SW Optimizations: It is interesting the explanation about LLM inferencing (doubt I can’t explain it though) and all different optimizations. The hw optimization (different custom hw solutions vs GPU) section was a bit more familiar. My summary is you dont need the same infrastructure (and cost) for doing inference and there is an interest for companies to own that as it should be better and cheaper than hosting with somebody else.

Network Acceleration for AI/ML workloads: Nice to have a summary of the different “collectives”. “collectives” refer to a set of operations involving communication among a group of processing nodes (like GPUs) to perform coordinated tasks. For example, NCCL (Nvidia Collective Communication Library) efficiently implements the collective operations designed for their GPU architecture. When a model is partitioned across a set of GPUs, NCCL manages all communication between them. Network switches can help offload some or all of the collective operations. Nvidia supports this in their InfiniBand and NVLink switches using SHARP (Scalable Hierarchical Aggregation and Reduction Protocol – proprietary). This is call “in-network computing”. For Ethernet, there are no standards yet. The Ultra Ethernet Consortium is working on it but will take years until something is seen in production. And Juniper has the programmable architecture Trio (MX routers – paper) that can do this offloading (You need to program it though – language similar to C). Still this is not a perfect solution (using a switches). The usage of collectives in inference is less common than their extensive use during the training phase of deep learning models. This is primarily because inference tasks can often be executed on a single GPU

From a different topics:

Learning at Cambridge: Spend less hours studying, dont take notes (that’s hard for me), go wild with active learning (work in exercises until you fully understand them)

British Library CyberAttack: blog and public learning lesson. I know this is happening to often for many different institutions but this one caught my eye 🙁 I think is a recurrent theme in most government institutions were upgrading is expensive (because it is not done often), tight budgets and IT experts.

“Our major software systems cannot be brought back in their pre-attack form, either because they are no longer supported by the vendor or because they will not function on the new secure infrastructure that is currently being rolled out”

However, the first detected unauthorised access to our network was identified at the Terminal Services server. Likely a compromised account.

Personally, I wonder what you can get from “stealing” in a library ???

Google Networking, AI Cooling, MATx

OpenFlow at Google – 2012: Openflow to manage to network, to simulate your network. 2 backbones: first for customer traffic and second for inter-DC traffic

UKNOF32 – Google Datacenter networking 2015: Evolution until Jupiter. Moving from chassis based solutions to pizza boxes. Smaller blast radius than a chassis. This switches have small buffers but Google uses ECN (QoS) for dealing with it.

Google DC Network via Optical Circuit 2022: (other video paper google post) Adding optical circuit switches, no more Clos network !!! Full mesh connection of aggregation blocks. Spines are expensive and bottlenecks. Traffic flows are predictable at large scale. Not building for worse scenario. Drawback: complex topology and routing control! Shortest path routing is insufficient. TE: variable hedging allows operation on different points along the continuum to tradeoff optimality under correct prediction vs robustness under misprediction -> no more spikes. Hitless topology reconfig. It seems it has been running already for 5y…. To be honest, It goes a bit… beyond my knowledge.

Google TPUv4 + Optical reconfigurable AI Network 2023: Based on the above but for AI at scale. Although there is already TPUv5. From this page, the pictures help to get a view of the connectivity. Still complex though.

Open Computer Project 2023: AI Datacenter – Mainly about how to cool down the AI infra with some much requirement of GPU/power.

MATx: A new company to design hw for AI models

AI will save the world, Nutanix kernel upgrade, GPU Programming

AI will save the world: Positive view of the AI development. Interesting the attack to China/Karl Marx at the end. In general I feel confident this will be good.

Nutanix kernel upgrade story: This is a bit hardcore for me (and looks a bit old from 2021) but still quite interesting how they did the troubleshooting.

GPU programming: I have never read about how to code for a GPU and this looks interesting and quite different from what I would do in a CPU. From the “Execution Model of the GPU” I started to lose track. Still is nice to see a summary at the end and resources/books.