• 1 Post
  • 74 Comments
Joined 1 year ago
cake
Cake day: June 15th, 2023

help-circle





  • Lol, especially because I work in IT. I’m tired of companies only issuing laptops to employees when SFF systems get you more for less. The company where I work currently sends out a laptop, two monitors, a mouse, and keyboard for all hires.

    They were giving out 8GB systems to developers (mostly running Windows and Visual Studio). It was a massive issue. I made a big deal out of it and advocated for new systems with 32GB. They bought new systems, sent them out, and it turns out they all had 16GB.

    The person doing the order missed that detail and thought they were getting a deal for the company. Which resulted in a complicated process of ordering everyone an extra stick of RAM and then trying to instruct everyone how to open a laptop to install the RAM.

    A SFF system would have solved much of these problems. Cheaper than laptops, usually better specs and thermals, and far easier to upgrade.

    So to me, there’s a massive difference between desktops and laptops.



  • So most of the entire world is using windows/mac if they want to do something serious other than web browsing.

    Absolutely not. Depending on what you want to do, Linux enables you to do way more than Windows.

    And most of the linux desktop usage is contributed by enterprise and office pcs using linux.

    Do you have actual numbers for this, or is it just entirely your own anecdotal observation?

    Even if developers follow these numbers and start supporting linux, they will soon realise it is not worth their time because linux usage is mostly due to enterprise running chrome.

    Garbage argument. It’s like the whole induced demand nonsense city project planners use. No one ever says “I’m only using Linux because I don’t need the extra stuff Windows can offer”, instead what you often hear is “I’d switch to Linux if this one specific application wasn’t Windows only”.

    The largest thing that has held Linux adoption is application compatibility.

    And one fact that I know that questions your “it’s only cheap enterprises” argument is that Linux is huge in the academic sector in India (and the world also). More than half of the AI and ML demo videos on YouTube are from Indian accounts.



  • These numbers are inflated due to our population and government and health sector office pc using linux

    So just like Windows numbers being massively inflated because of corporate computer fleets?

    These office pcs just require a chrome browser and all the work is done on the browser Nobody here cares what os they use in their office pc.

    Right, so again, the mostly the same with Windows for both office and personal use.

    I don’t see anyone here switching to linux on their personal pc other than the IT students who are forced to install kali linux.

    What are you expecting exactly? Is the choice of each person supposed to be formally announced? Are we supposed to real into a populated areas and declare like Micheal Scott “I declare: I’M USING LINUUUUUX!”?

    People here buy desktops only for gaming/content creation, which means most households here doesn’t need/require a desktop.

    You just described the entire world. This is far from unique to India. Most people I know don’t have a desktop and maybe have a laptop, and I live in North America.

    Not to be conceded, but I’m guessing this post is in response to my comments from a couple days ago?

    I really don’t understand your point. It’s like you’re saying “the users in India don’t count because they’re not using Linux the way I do”.

    Does that mean that all the workstations at CERN don’t count? Or that the systems up on the ISS don’t count?

    To me (and I’m certain most people in general would agree) the ISS story is very important, because they were originally running Windows on those systems, but it kept crashing. They switched to Linux to get more stability out of those systems and have been using Linux ever since.

    Also, does the story of the City of Munich switching to Linux not count either? It’s supposed to be a major win, btw. A city government switching away from Windows and choosing to go with Linux is huge. I see it the same way with India. The more often people are Linux in the wild, the more normalized it is and the more mind share it generates. And mind share is huge in getting people to make a certain choice. It’s the reason why product ads are everywhere. The more often you see a product/brand, the more likely you are to say to yourself “that’s the thing I’ll buy”.

    Before anyone says Munich switched back to Windows, they didn’t. Microsoft made an under-table deal with some officials with the at-the-time in power government to switch back to Windows if they set up a Microsoft office in Munich. Then a new government was voted in a few months later and said “hell no, we’re continuing with the Linux rollout” and that’s where we are today. The City of Munich is a Linux success story.

    Ultimately your post was just stating some facts and then waffling on about how it doesn’t count.




  • FAILED] Failed to mount /boot.

    Something you did made your boot partition disappear.

    Then it’s asking to give root password or press control-D, which I’ve dealt with before but this time my keyboard just doesn’t work

    It dropped you into an emergency console. Which makes sense because something has gotten terribly messed up.

    It would be helpful if you have some extra details like system specs. Partition layout (is boot, root, and home all on the same drive?). Which FS are you using, etc.

    It’s likely you’re only going to get back to a working system with a live environment. You’ll have to manually mount the important partitions (root and /boot), chroot to those mounts, and then repair the bootloader.




  • A lot of what you said is true.

    Since the TPU is a matrix processor instead of a general purpose processor, it removes the memory access problem that slows down GPUs and CPUs and requires them to use more processing power.

    Just no. Flat out no. Just so much wrong. How does the TPU process data? How does the data get there? It needs to be shuttled back and forth over the bus. Doing this for a 1080p image with of data several times a second is fine. An uncompressed 1080p image is about 8MB. Entirely manageable.

    Edit: it’s not even 1080p, because the image would get resized to the input size. So again, 300x300x3 for the past model I could find.

    /Edit

    Look at this repo. You need to convert the models using the TFLite framework (Tensorflow Lite) which is designed for resource constrained edge devices. The max resolution for input size is 224x224x3. I would imagine it can’t handle anything larger.

    https://github.com/jveitchmichaelis/edgetpu-yolo/tree/main/data

    Now look at the official model zoo on the Google Coral website.

    https://coral.ai/models/

    Not a single model is larger than 40MB. Whereas LLMs start at well over a big for even smaller (and inaccurate) models. The good ones start at about 4GB and I frequently run models at about 20GB. The size in parameters really makes a huge difference.

    You likely/technically could run an LLM on a Coral, but you’re going to wait on the order of double-digit minutes for a basic response, of not way longer.

    It’s just not going to happen.


  • when comparing apples to apples.

    But this isn’t really easy to do, and impossible in some cases.

    Historically, Nvidia has done better than AMD in gaming performance because there’s just so much game specific optimizations in the Nvidia drivers, whereas AMD didn’t.

    On the other hand, AMD historically had better raw performance in scientific calculation tasks (pre-deeplearning trend).

    Nvidia has had a stranglehold on the AI market entirely because of their CUDA dominance. But hopefully AMD has finally bucked that tend with their new ROCm release that is a drop-in replacement for CUDA (meaning you can just run CUDA compiled applications on AMD with no changes).

    Also, AMD’s new MI300X AI processor is (supposedly) wiping the floor with Nvidia’s H100 cards. I say “supposedly” because I don’t have $50k USD to buy both cards and compare myself.




  • And you can add as many TPUs as you want to push it to whatever level you want

    No you can’t. You’re going to be limited by the number of PCI lanes. But putting that aside, those Coral TPUs don’t have any memory. Which means for each operation you need to shuffle the relevant data over the bus to the device for processing, and then back and forth again. You’re going to be doing this thousands of times per second (likely much more) and I can tell you from personal experience that running AI like is painfully slow (if you can get it to even work that way in the first place).

    You’re talking about the equivalent of buying hundreds of dollars of groceries, and then getting everything home 10km away by walking with whatever you can put in your pockets, and then doing multiple trips.

    What you’re suggesting can’t work.