This is such funny gatekeeping. I don't know what % of the U.S. population could even run 1 mile at all without stopping, but I'm certain it's well under 50%, much less do at least that every day for 10 years. This is an impressive feat. For real, shame on you for crapping on this person's very respectable achievement.
It's not "your" HN, HN doesn't do algorithmic/per-user ranking. (Ed.: Actually a refreshing breath of wide social cohesion on a platform, IMHO. We have enough platforms that create bubbles for you.)
It's top1 on everyone's HN because a sufficient number of people (including myself) thought it a nice writeup about fat ARM systems.
I haven’t been following hardware for a while, granted, but this is the first time I see a desktop build with an arm64 cpu. Didn’t know you can just… buy one.
For what it's worth, I've been using a Lenovo X13s for some 3 months now. It's not a desktop, and it took years for core components to be supported in mainline Linux, but I do use it as a daily driver now. The only thing that's still not working is the webcam.
Would you call Threadripper system "a normal build"? For many people they are normal builds because they need more computing power or more PCIe lanes than "normal user" desktop has.
On the other side you have those who pretend to use raspberry/pi 3 as "an Arm desktop" despite only 1GB of ram and 4 sluggish cores.
torchft can handle much larger scales but for public multi-day demonstration run this is what we had available. Point of this blog was to demonstrate correctness of the quorum algorithm and recovery with a stock PyTorch stack and not so much peak flops.
Stay tuned though -- planning on doing some much larger demos on B200s!
I was curious about this so I had o3 do a bit of research. Turns out 300 L40s have more compute than any supercomputer before 2013 (and arguably before 2016, depending on how you count reduced-precision FLOPs).
1. High-level, the post is all wrong. The point should be that you always need to make sure you can pull new cable. The poster illustrates this: single mode, multi mode, non-fiber, etc, etc. And if one "goes bad", you still can't run a new one, unless you have a pullstring.
2. The post cannot apply to fast/large networks - will be prohibitively expensive.
3. If running a few at home, I suggest to run MTP/MPO. It's basically a structured cable that can have around 12 fibers in them, plenty of future expansion.
Though I'll always run a large awg >>cat6 everywhere so it supports PoE++
That works for short runs, but for long runs, the only way to pull the cable without breaking it is to use wire pulling lubricant.
And there's no guarantee that the cable you are pulling is undamaged. Rats don't improve pull strength, and even an electrically good cable can have the tensile strength cord severed.
If you are in that situation, pulling the old cable with dried lube on it may snap it in two, especially if it's pulling the new strand.
That's why leaving a pulling tape in each conduit is a good idea.
This is correct - I have pull strings in my conduits at my house for that reason. Pulling fiber down my 1000ft driveway would be difficult to do if the pull string was another fiber.
You can pull a new pull string in empty conduit pretty easily with a shop vac and a ziploc bag.
I ran fiber when I built my house a few years ago. I have MANY 12 fiber runs going to my IDF/MDFs, Server room, the gate, the dmarc, etc. I also ran SM fiber to every AP location. Out of 20 miles of pulls about 1 mile was fiber.
I have both. 2 Cat6As plus a SM fiber to each AP location. The APs need power over POE. I just added in my first AP that was 10g (Ubiquiti XGS), and that uses POE over 10g.
> 3. If running a few at home, I suggest to run MTP/MPO. It's basically a structured cable that can have around 12 fibers in them, plenty of future expansion.
For a homelab honestly just pull pre-terminated cable with LC connectors, a 20 or 25mm hollow pipe is enough. That way you don't need a splicer machine, experience on how to operate it, or measuring equipment.
the author is a sysadmin, who can definitely change what is plugged in to the switches their employer has chosen to buy, but doesn't have the power to make physical changes between multiple buildings, easily authorise opex spending on cable pullers or retrain as a cable puller.
please actually consider what you're replying to before pushing the REPLY button.
> 2. The post cannot apply to fast/large networks - will be prohibitively expensive.
The expensive parts of fast/large networks are not the fiber strands.
* right of use / lease
* trenching and laying and covering
* amplifiers on long lines
* repair/maintenance
* endpoints
... which is why the first thing you upgrade are the endpoints, and the last you do is lay more fiber. Get the most you can afford (often physically) at the beginning.
Pull strings don’t give you a way to pull new cables. They give you a backup way to pull new cables. Every cable in a conduit can be used to pull its own replacement or 2. But if anything goes wrong, you have a second chance before you have to go find the fishing rod.
reply