New AI Breakthrough Has Big Potential to for Engineers

None of us were alive when Greek Mythology was in its heyday, but if you were you’d certainly now that Cerberus was the 3-headed dog that guarded the gates to Hades. Sure, there will be a few history buffs alive today that know of this mythological creature, but it really is a shame that generally speaking humans don’t read books like they used to. But enough about that, this blog is about anything and everything related to web hosting and / or computer development so fair to ask where we’re going with this.

No one at Cerebras Systems has more than one head and will never have been anywhere near any entrance to the Underworld. But Cerebras does have the distinction of being the maker of the world’s largest processor. And if you’re an engineer who works with development in AI then you may well think they’ve outdone themselves with their newest offering to the development world. Do a little research into AI and you’ll learn that the capacity it has is entirely based on being able to add more parameters to it.

20 billion is a big number, and so the reason this news around the CS-2 system and WSE-2 chip is so big is because that the newest offering from Cerebras is able to train AI models with up to 20 billion parameters. This is the ultimate in optimization at the software level, and that’s why it is something that a top Canadian web hosting provider like those of us here at 4GoodHosting are going to take an interest in. Like most, we have a nice front row seat for everything that major advances in computing technology have promise for doing for us.

So let’s use this week’s entry to have a much deeper look into these CS-2 devices. Long before AI technology becomes commonplace in darn near everything of course, but it’s good to be early rather than late.

No More Mass Partitioning

What this all promises to do is resolve one of the most frustrating problems for AI engineers. That being the need to partition large-scale models across thousands of GPUs in order to facilitate full cross-compatibility. What the CS-2 promises to do is drastically reduced the time it takes to develop and train new models.

Natural Language Processing has undeniable benefits, but the degree of functionality for them is entirely dependent on the number of these parameters that can be accommodated. To date the way it has worked is that the performance of the model correlates in a linear fashion with the number of parameters. Larger models provide better results, but these days the development of large-scale AI products traditionally requires a large number of GPUs or accelerators, with the models spread across them.

The wheels fall off when there are too many parameters to be housed within memory or compute performance is incapable of handing training workloads. Compounding the problem then is the way the process is unique to each network compute cluster pair, so each cluster has to be catered to individually and that makes the whole thing so much more of a drawn-out process. That’s time many of these engineers will be displeased to be losing.

Bottleneck Buster

It is true that the most complex models consist of many more than 20 billion parameters, but it may be possible now that the ability to train relatively large-scale AI models on a single CS-2 device will do away with these bottlenecks and majorly accelerate development for existing players, plus increasing access for those previously unable to participate in the space. More cooks in the kitchen is a plus when it comes to high tech development much of the time.

The consensus seems to be that the CS-2 may we have the ability to bring large language models to the masses in a cost-efficient way, and this greater ease of access may well be the stimulus needed for a new era in AI where big new steps are made and in the long run society benefits from it in many different ways.

There’s potentially more to this too, as according to Cerebras the CS-2 system may be able to accommodate even bigger models down the line, perhaps even capable of trillions of parameters. If that’s true, chaining together multiple CS-2 systems could start us down the path to AI networks that are more expansive than the human brain.

Tips for Protecting GPUs and CPUs from Heatwaves

If you’re from these parts of North America you won’t soon forget how the weather was about exactly a year ago from now. Late June of 2021 saw the West Coast hit with a heatwave that put every other one EVER before it to shame with just how incredibly hot it got. The mercury soared to well over 40 degrees in most places and one location not too far from here recorded the hottest temperature in Canadian history. To say it was sweltering hot would be an understatement, and unfortunately it seems that we’re all living in a much, much warmer world than we used to.

We are not going to segue into discussing climate change and / or global warming here, but what we will do is say that it is not just us humans and the animals around us that stand to be worse for wear because of extreme temperatures. Turns out they can be plenty harmful for the digital devices a lot of us rely on for work, and the same ones that nearly all of us will use for entertainment even if we don’t spend our workday in front of monitors and a keyboard. There a reason most GPUs come with the own radiator and cooling fan, and we’ve all heard our CPU or notebook whirring when it needs to cool itself down.

This is a topic of interest for us here at 4GoodHosting, and we are like any reliable Canadian web hosting provider in that just like many of you we are mix of both those scenarios and if it’s not the desktop or notebook that is being put through its paces it is the mobile. Such is the way with modern life for a lot of us, and anything that puts the operations of that in jeopardy is going to have a target on its back. There’s all sorts of buzz around malware and DDoS attacks and the likes these days, but what about temperature extremes?

Components Kept Cool

Processors and graphics cards are very sensitive to heat, which is why ensuring they have the ability to stay cool when the temperature around them rises drastically is important. This isn’t a new requirement reality, and most will come with some form of cooling solution and better ones depending how premium the product is. In high-strain environments you may also find water-cooling blocks and AIO cooling systems can require a fan and radiator to pull fan-cooled water over the hot surface. The thing is, these solutions will be dependent on having cool air around them if they’re to be effective.

As it is now the system generally only cools in relation to the ambient temperature of its surroundings, and this puts additional strain on our computers and laptops. But how hot is too hot here exactly? This isn’t our area of expertise, but we’ve dug up that typical safe GPU temperatures are usually between 150° to 185° (65° to 85° Celsius).

For CPUs the same values are around 140° and 160° Fahrenheit (60° to 70° Celsius), but anything even near that may be putting your system in jeopardy of overheating to the point of failure. That’s within your power to control, but what if the same reality exists for a CPU or GPU and there’s nothing you can do to remedy the situation?

Stay Cool Tips for Users

So here’s what you can do to safeguard your components if you’re aware that a heatwave is on its way.

  1. Clean Device and Make Sure Vents are Clean

The buildup of dust and fluff is likely occurring in every desktop PC much faster than the owner is aware of, and this may mean your system is already deprived of airflow. The smart move is to use a can of compressed air to gently dust away any buildup every once in a while. If you have a PC case that can be opened up by removing a panel then you should also give your case fans a blast of air to clean them, and make sure that the inside of your system is clear from any dust build-up.

  1. Overclock Less Often

Many PC enthusiasts aim to maximize performance with overclocking. This not only draws additional power, it also increases the average operating temperature and sometimes quite significantly. Unless you absolutely need your system to be overclocked at all times, consider pulling back on the days when the weather is hot and likely going to get hotter.

  1. Get Fanning

Case fans are integral to keeping your CPU and GPU cool. With some units you may have a controller that allows you to bump your fan speed up to 100%, but if not you will need to control this directly on the system by heading into the BIOS again. This is time consuming, so a good option is to install 3rd-party software like SpeedTemp, but be aware that some CPU fans won’t be compatible with it.

A case upgrade may be in order if you have those fans maxed out through SpeedTemp and the unit is still running too hot.

  1. Replace Thermal Paste / Thermal Pads

This one may be well beyond the means of a lot of people, and quite likely if you’ve never opened up a computer before in any capacity. If your build has a few years behind it, consider cleaning off the old thermal paste from the CPU and replacing it with a fresh batch. This may help to bring the heat down.

Some units may not have the best thermal pads to begin with either. We’ll let you look into this one further on your own as we don’t want to on too long here. There are good tutorial videos easily found on YouTube.

Machine Learning for Google Chrome Set to Make Web Browsing Much Better

\

It’s pretty rare to be entirely satisfied with something, and there’s usually always little things that irk you when it comes to whatever it is. And all that despite being overall fairly satisfied with the experience of it. When it comes to web browsing, it is likely that most people will say it’s the pop ups that annoy them most, and for people who have health concerns that have connections to optic nerve overstimulation they can be more than just a nuisance.

There’s no getting around the need for revenue, so unfortunately pop ups and the like aren’t going anywhere. But the web development giants of the world have made the efforts to improve the experience when using their web browsers over the years. Real improvements haven’t been fast to arrive, and in many ways that’s because the imperfections of browsers like Edge, Firefox, and Chrome are tolerable in the big picture of everything.

So with that understood we can assume that anything that does make news when it comes to an improved web browsing experience has to be something rather big and significant. And it is here with Google making it clear that they are now making real strides implanting machine learning into the function of their flagship Chrome browser. A browser that – as many of you will know – is far and away the one of-choice for most people around the world.

And here at 4GoodHosting we’re like most good Canadian web hosting providers in that we feel similarly, and we’re like anyone anywhere for whom a more enjoyable and tailored web browsing experience is going to sound plenty good. So let’s dig into this here today as it’s something that likely all of us are going to benefit from in the more long-term picture around this.

Specific Task Focuses

It was in a recent blog post that Google announced the rolling out of new features to Chrome via ‘on device’ machine learning. Improving a person’s browsing experience is the goal, by adding several new ML (machine learning) models that will focus on different tasks this will be made possible. For starters, ‘web notifications’ will be handled differently, and we may well be presented with a better and more adaptive tool bar too. Google states these new features will promote a ‘safer, more accessible and more personalized browsing experience’ and one that is better for all different types of users.

Another big plus is that having models run (and stay) on your device instead of in the cloud will promote better privacy, and maybe we can go so far as to think it may eventually keep certain people from feeling they have to go with Duck Duck Go or something similar.

But let’s stick with the indicated changes for web notifications first. What we’re going to see in upcoming versions of Chrome is an on-device machine learning structure that will examine how you interact with notifications. If and when it finds you are routinely denying permissions to certain types of notifications then it will put a stop to you receiving that type in the future. Nearly everyone will be thankful to not have to click ‘dismiss’ every time they pop up, and as a lot of us know they pop up way too often.

Good intentions, sure, but it will be nice if the browser is able to see the pattern and realize that permission is never going to be approved and then stop requesting for it. However, you’re still able to override Google’s prediction if in fact you don’t want to continue having the choice for yourself.

Responsive to Behaviour Patterns

Another aim that they’ve focused on where is to have Chrome change what the tool bar does based on your past behavior. One example would be where people like to use voice search in the morning while on public transit, or more likely when a person is prone to sharing a lot of links. For either scenario Chrome will then anticipate your needs and add either a microphone button or ‘share’ icon in the tool bar to simplify the process for the user.

Customizing it manually will be possible as well, but as of now whether or not this functionality will appear on other platforms is not known. We do know that any time proprietary interest aren’t involved that most other developers will follow suit if the changes to the technology end up being well received.

Google has also been quick to tout the work machine learning is already doing for Chrome users. Even with the current version when you arrive at a web page that page is scanned and compared to a database of known phishing or malicious sites. In the event of a match you are provided with a warning, and you may already be familiar with it – a full-page, all-red page block, that has been standard since March of this year. Google states that it is a reflection of how Chrome is able to now detect 2.5 times more malicious sites than it could before.

Some – Not All – In Next Release

Apparently the smart silencing of notifications will be part of the next release of Chrome, but we may have to wait for certain other offerings that are part of Google’s greater incorporation of machine learning into the flagship web browser. Another question people are asking is if these improvements will be for mobile only, and some people are wondering if overall browser performance might decline by this. Google has said that the adaptive toolbar is one feature that won’t be in the next roll out

This will be an interesting development to keep an eye on, and in large part because as mentioned when improvements are proven positive and well received other tech giants tend to follow suit with regards to their products. And given how much time nearly all of us spend web browsing it’s something that most will have a good level of interest in.

Power-Providing Clothing for Wearable Devices Soon to be a Thing

Most people won’t have heard of power dressing, or if they have they’ll assume it’s a term related to style choices for people who want to make the right impression in business. We will freely admit we’ve never heard of it in either context, but it seems as if the clothing that can power devices is in fact going to be joining the many rather stunning technological advances that have been made over recent years. Indeed, portable power banks in purses have been extremely helpful for people who tend to really put certain devices through their paces during the day, but can we think that they might be replaced en masse by garments?

Well, apparently they may be and what we are talking about here is a flexible waterproof fabric that is built to convert body movement into electrical energy for powering those devices. So if you’re wondering if we’re talking about a pair of pants that powers you up the more you walk – that’s exactly what we are talking about here! It really is quite the eye-opener to dig up all the technological advancements that are occurring in the world, and what we do here each week puts us directly in line to do that.

This is the sort of news that is intriguing for any good Canadian web hosting provider, and obviously that’s because it is going to have an immediate appeal for everyone and not just those who are especially digitally and web savvy, or even a person who relies on someone like us for reliable web hosting. If motion can mean recharging, who’s not going to see the upside in that?

Press on Polymer

If what we’ve read about this is correct, this really is some cool technology in the making and as the researchers in Singapore have noted it is set up to be particularly ideal for people with wearable smart devices that have more in the way of ongoing power needs and / or aren’t quite as easy to be recharging regularly.

The key ingredient in this powering fabric is a polymer that when pressed or squeezed takes vibrations produced by the smallest of motions and then is able to convert them into an electric charge. The material is made with a spandex base layer and reinforced with a rubber-like component. When conducting an experiment with it researchers found that when they tapped a 4cm piece of the material it immediately generated enough electrical energy to light up 100 LED bulbs.

Again, they are referring to this as Power Dressing, and the concept of it has actually been analyzed for more than 20 years before it picked up enough steam to be where it is today and getting closer to realization. One thing that they are finding, however, is that most electricity-producing fabrics don’t hold up to long-term use. They also don’t take very well to being cleaned in washing machines, and that is a hurdle that absolutely has to be overcome.

Work in Progress

So obviously the immediate focusing was getting over that roadblock, and developing something that does not degrade in function after being washed and maintains the same level of electrical output over time. Consider it done, as the development team in Singapore has now done just that and is reporting their revised conductive material doesn’t lose anything when washed, folded, or crumpled. What’s more, they are estimating that it will maintain stable electrical output for five months of wear, and keeping in mind no one’s going to be wearing these garments every day.

The prototype is set to woven into garments, but it may also eventually be compatible with footwear, which is where it would most likely get the most bang-for-buck with charging based on intensity of motion.

It is capable of producing a 2.34-watts-per-square-meter charge in one of 2 ways: either by pressing or squashing in the same way standard piezoelectricity is created., or when it comes into contact or is generating any measure of friction between other materials, including skin or certain other fabrics that would promote it. When attached to an arm, leg, hand, elbow, or even to the insole of a shoe it will be able to harness energy from a range of human movements, including running, playing sports, or roaring through the park trying to keep up with your dog.