Working on a single-node job scheduler for Linux. Large HPC clusters use schedulers like SLURM or PBS to manage allocation of resources to users, but these systems are quite overkill when all you have is a single node shared by a few users.
I am trying to offload as much of the complex stuff to existing parts of the kernel, like using systemd/cgroups for resource limiting and UNIX sockets for authentication.
I also use a it this way, but instead I do 90/30 (work/break) as 90 minutes gives me enough time to finish most exercises and tests without having to hurry.
I do around 4 to 5 sessions per day which hits the sweet spot for me as I can endure this for multiple months without having to take days off. This comes at the expense of having to start a few weeks earlier to have enough time to prepare for tests compared to cramming 12 hour days the week before
I can tell you what I did, coming from a somewhat similar but different situation. I kind of over-promised for a small project, thinking that some electronics and little code could not possibly be too hard. After some time I realized that if I had to do the stuff around 50 times (project was a small sensor network), I would be stuck soldering for a few weekends. Then decided that maybe learning some ECAD along the way and getting PCBs manufactured would be cool, albeit totally overkill.
I cannot recommend Phil's Lab [1] enough. He has videos which take you from basically an empty sketch to ordering a finished board in around 2.5 to 3 hours. Some of the most worthwhile content I have found on YouTube to date.
Also, get a multimeter and an oscilloscope. Any oscilloscope will do probably, I have one from Amazon for 30$. For my project, it totally does the job, although you may want to invest a bit more here if you intend to use it for more advanced projects. You cannot believe how many times being able to look at a signal has helped me understand where my problem could be and what might be wrong. It has also helped me figure out what questions I had to ask. For simple projects, you are always going to get there by guessing I suppose, but why make it hard on yourself.
Aside from that, just try things. It is probably helpful to spend more time thinking when you are handling more expensive parts, so you do not burn a couple of them out because of carelessness, but also, don't overthink it.
I guess I understand where you are coming from, but they way I understand it is that RH submits a bulk order for a stock and gets a quote from the MM, which the MM thinks is appropriate. The MM then adjusts its own portfolio accordingly, buying/selling the underlying stock at its own pace.
So, more or less as described, but the order is a little different, right?
I was wondering if someone could explain the countermeasures for such an event. Obviously, as the article states, producers are being shut off in the regions with surplus, while drains, who can afford to shut off, are shut off in the deficit regions.
Is this an automatic process? Or is it more like someone from the company's energy provider calls them and tells them to shut off some devices? And is there not a potential problem, that if too many shut of at once, you now have a surplus again? Or is it coordinated by one single entity?
> In that case it could be interesting to plan for future reinforcements, and continue to run the simulations as time passes and the situation evolves.
While I agree that the idea is good, I imagine that at a point where you are planning to extend your dam by another 20m, you already have to implement the required changes in your foundation and general structure and additionally, turbines and emergency vents have to be upsized for the extended heights. So the only thing you are now missing during build is the extra 20m of dam, as the rest has to be built anyways. This leads back to my initial point of the right amount of over-/underprovisioning..
I am not sure if I understood the second point correctly, but you can obviously tell when a dam becomes too dangerous. This can either be because the foundation has set and you now have increased structural stress, the amount of water has increased beyond the maximum allowed levels or a multitude of other reasons, but they are actively monitored.
This is the question that has been going through my head after I wrote the parent comment.
I have never been in such a situation (at least not one where the stakes were relevant), so take this with a grain of salt. Blaming under these kind of circumstances might arise because the people who do it:
* Either do not have access to the information necessary in order to develop a sufficient understanding of the problem at hand
* Or are incapable of understanding the problem at hand to a sufficient degree (because the project is too big for a single person to grasp or whatever)
So, they resort to assuming some simplifications in order to make everything understandable to them, but these simplifications are likely to be wrong. Consequently, they start blaming people they think to be at fault under their own flawed model of the situation.
A blame-free culture? Explain the situation to such a degree that the decision making around the root cause is easily understandable for everyone. For example, if a bridge fails, there will likely be experts that identify what has caused the failure. If the decision makers for the bridge now say that due to circumstances they disregarded a certain load situation, because x and y and this is an acceptable argumentation (to whoever is judging the situation), there will be less/no blaming as it is now understandable why things have been done that way.
But at a cultural level? I think that this would require everyone to ultimately only judge situations once they have understood the problem at hand. At that point, why do we even talk to each other? Everything we say is based on incomplete information anyways..
Yeah, exactly. The thing is, where do you draw the line of what is statistically so unlikely that you can deem it improbable?
If we run multiple different climate models through some sort of Monte Carlo simulation, I suppose each model would output a different probability of such an event occuring. In the case where all models predict a very low probability, it may be easy to say that it is unlikely. But what if two predict a very low and one predicts a low probability? Is this now applicable and should we build to be able to sustain such an event?
These are hard questions and I currently do not see any way to get better data for the future as the different models still do not agree in many points
While I agree with the gist of this, the ones to finally call the shots on such projects are always in a bad spot.
They underprovision and an anomaly happens. Now they are the ones who did everything wrong and should be blamed.
The overprovision and nothing ever comes close to the theoretical limit of the structure. Now they have wasted huge amounts of money and again, should obviously be blamed.
The easiest way out: Plan exactly to what is considered standard. No one will blame you in either case, even if you know that it is insufficient or stupid.
This is something that will probably hold true in many different industries, the consequences for dams are just a little worse..
In Germany, we have two different institutions: Universities (identical to US university I guess) and something called Fachhochschule (I don't know if there even is a correct translation for that, lets call it FH).
University is focused on covering the theoretical aspect of computers with a lesser focus on practical applications - although they are thought to some extent - while FH is more focused on building applications and real-life systems with a smaller focus on theoretical aspects.
The only problem is that FH has long been considered second class and everyone wants to attend university. but in my opinion and based on my experience, many people would be a better fit for FH.
For someone deciding whether to attend Uni or FH, it is often not easy because you cannot guess what you are going to do in your work-life and whether you might enjoy research or you would rather go into industry directly.
Aside from that though, I think it is great that we have different institutions for different audiences with differing goals. Does something like that exist in the United States or somewhere else in the world?
This article surprised me since funding came from Softbank and not the state government as most does IIRC.
Where I see the problem with all these forms of energy storage is that there is no one efficient approach (yet!) and innovation will come mostly from government funding, so governments decide what they see as the best opportunity.
This seems like its pretty much hit-or-miss and betting on the wrong thing will arguably speed up the development of it but at the same time, we might end up at a point we could have reached quicker and cheaper had we used a different technology.
There are a bunch of other technologies as well. Huge flywheels [1], which are supposed to be used for very responsive power needs (i.e. grid stabilisation), trains loaded with weights which get pulled up a mountain and released [2], redox-flow batteries and so on.
Not such a big deal and something rather subjective: Due to Covid, my university started to upload recordings of all of my courses. This has been a godsend for me, as I did not attend lectures previously and now have the chance to watch them only when needed.
Also, where before I often had to go to campus for 1.5hrs, which often meant sitting in train for equally long. Now, I can watch stuff from home and don't have to waste time travelling.
While this is likely not low cost, the Anvil[0] seems to be a reusable intercept drone. I have yet to see it in action, but I could imagine it looks pretty cool.
I heard that the company is pretty controversial, but I find their tech to be very cool nevertheless.
At around 1:54:50, upon full landing thrust, the exhaust flames turn green. I know colored flames from chemistry classes, but I always assumed that was caused by certain salts.
Rocket fuel, to my knowledge, does not contain any of those ingredients. Can someone explain what happened there? Was this planned?
This has been done in my last exam phase and has worked well. They cranked the AC to 11, which made it quite an unpleasant envirommemt, but I would rather wear a jacket than take online exams
I am trying to offload as much of the complex stuff to existing parts of the kernel, like using systemd/cgroups for resource limiting and UNIX sockets for authentication.