Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited


This user doesn't have any awards

About ryao

  • Title

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Read the context of what you wrote. It was a reply to something and you said nobody would say that. I replied to the contrary.
  2. @GabenJr Your remark about the subpixels seems wrong to me. Here are the raw numbers: iPhone 6/6s/7/8 screen: 1334*750*3 = 3001500 subpixels iPhone 6/6s/7/8 Plus screen: 1920*1080*3 = 6220800 subpixels iPhone 11 screen: 1792*828*3 = 4451328 subpixels iPhone 11 Pro screen: 2436*1125*2 = 5481000 subpixels iPhone 11 Pro Max screen: 2688*1242*2 = 6676992 subpixels The actual numbers on the iPhone 11, 11 Pro and 11 Pro Max screens are smaller due to the notch. We had true 1080p on the plus series for years before Apple jumped on the pentile bandwagon. You need the 11 Pro Max to get a similar level of subpixels and the arrangement is off enough that you probably won’t be getting the best image quality. Maybe the typical human eye won’t notice, but I am not a fan of this.
  3. I would not say that Windows is immune from needing fiddling. People fiddle with a bunch of settings (especially CFG) on Windows to try to fix stutter in recent games. Getting older games like Zoo Tycoon or Shattered Union working requires more fiddling than you need on Linux. They both need fiddling for various things. There will always be things that one does better than another. Linux has achieved at least “good enough” status in a number of key gaming metrics though. However, the news that it only has a 2-3% performance hit in a recent AAA Windows game has put it beyond good enough territory into effectively the same performance territory in that metric. The difference for that game is like the difference between Intel and AMD for gaming. Anyway, I would say otherwise as evidenced in the preceding paragraphs. Linux does not need to be better at being Windows than Windows to be a better option for many people.
  4. Did you read it? They successfully switched to Linux, ran everything on Linux for over a decade and were happy with it. No one seems to have been clamoring for them to run Windows other than Microsoft. Microsoft had been. After years of lobbying, Microsoft succeeded in making a deal with the most recent mayor. They are spending 90 million euros that they did not need to spend to migrate to Windows. They have not fully switched to Windows yet. The migration finishes next year. Until then, they are still using Linux. As for Windows making it easier for them to source compatible applications, that is certainly true. It will be easy for them to run cryptolocker ransomware now. Sometimes, being compatible can be a bad thing.
  5. Thus would suggest otherwise: https://en.wikipedia.org/wiki/LiMux#History Anyway, Linux desktop adoption is slowly increasing much like subscriber counts on Linus Tech Tips.
  6. https://www.stirlingcryogenics.eu/en/products/custom-closed-loop-cooling-systems/closed-loop-liquid-nitrogen-cooling-systems Linus should get in touch with those guys to build the world’s first publicly known closed loop LN2 CPU cooling system. That would go a step beyond this: It is not actually closed loop because it lacks the ability to reliquidify the evaporated nitrogen. A system that could would be cool.
  7. As one of the developers, I would be happy to answer any questions that they have about that.
  8. I am a Linux distribution developer: https://www.gentoo.org/inside-gentoo/developers/ Here is my list off the top of my head: Public issue trackers. Have you ever had a problem and you wanted to talk to the developers about getting it fixed? Have fun doing that with Microsoft and while Apple has a form that you can fill out, there is no possible way to see the status of the issue or talk about it. There is no way for the developers to ask for more information. Software transparency. There is far less need to worry about whether you were given software that has undocumented backdoors or telemetry because you get the source code. Everything is distributed by HTTPS mirrors and checksummed/signed to ensure that you really are getting the same updates as everyone else. There are no system identifiers being requested/provided to mirrors. The mirrors distributing updates have no opportunity to give you some “special” version based on one. Updates occur when you decide to do them and are never forced on you. If support gets dropped, you are not limited by your OS developer’s support decisions and you can fork or go elsewhere easily. Nearly all of the software is installed through the package manager (excluding games installed through steam), so system updates update basically everything. You do not have updaters running in the background wasting resources and you do not need to go hunt for updates when you want to do them. Security issues in shared libraries like libpng, OpenSSL and others do not put you at the mercy of every developer whose software you use when it comes to vulnerabilities in them. If those developers are long gone you still get the security updates. Each securityupdate is only done once and applies to everything using a shared library, unlike on Windows and Mac OS X where it gets to be done per application. You get a large number of choices for customization that others either do not provide or are a pain on them. For example, you can have different desktop environments to completely change the look and feel of the system. Replacing the desktop environment on Windows just is not done and it is a huge pain to do on Mac OS X. You can have different filesystem options. If you do not like the poor data integrity of the default filesystem option, you can use ZFSOnLinux (which I also help develop) and have some assurance that it will keep working through updates, unlike on Windows and Mac OS X. ZFS ports exist for those two, but you have no guarantee that it will work after doing Microsoft’s 6 month upgrade/update or Apple’s yearly upgrade/update. You could have tried using ReFS on Windows, but Microsoft actually took that away in an update a few years ago, so it definitely stopped working after that update. You do not need an antivirus. Malware designed for Linux is exceedingly rare and the ingress mechanisms are all patched fairly quickly. You would need to go out of your way to download and execute malicious code to get it. On other systems, downloading and executing code is normal, so you are not doing anything special when your system gets infected. On Linux, you get software through the package manager, so getting malware generally does require doing something special. The result is no antivirus in the background slowing your system down. You do not need a firewall for the most part. Linux distributions often do not start network daemons (and if they do in the case of say cups, you can configure them to not talk to the network). Without any open ports to target, a firewall just is not needed. Many ship them anyway, but being locked down by default goes a long way toward improving security. You do not need a defragmenter. Linux filesystems are designed to minimize the performance loss caused by defragmentation by allocating from different regions of the disk as opposed to Windows which tries to write all new data linearly to maximize sequential write performance (with the trade off that horrible slowdowns later happen when reading what was written). This means that you do not see extreme the slowdowns from file fragmentation that you see on Windows and you do not need a defragmenter wasting resources in the background to try to fix it. No UAC dialogs. With the exception of things that really do require root access (like software updates), you do not get dialogs saying a program is trying to do something. Those dialogues are so routine that even if they by chance caught something, an end user would click yes anyway, so there is not much point beyond annoying users. The restriction of asking for user permission to things that truly require elevated privileges makes being asked for root permissions much more useful in the rare event that something is doing something malicious. The predictability of these dialogs (being mostly only for software updates and adjusting system settings) means that you really can tell when it is for something that is not routine.
  9. Is there anything that prevents vulnerable files from iOS 12.4 from being put in place of the files from newer versions? While Apple will not sign the 12.4 download anymore, the files themselves should have valid signatures. Maybe downgrades to 12.4 would become possible through this. If just grabbing the few vulnerable files needed to enable an untethered jailbreak turns out to work, then that is another way to go.
  10. You can not prevent it when crossing an international border either. The scope of physical access is wider than people would think.
  11. @GabenJrThese processors can do up to 16 operations per cycle with AVX2. At a 2.7GHz all core clock speed, you would see 2764.8 GFLOPS. Memory bandwidth is reportedly 264GBps. This is around the level of a GeForce GTX 670, which probably could do crisis at maximum settings at 4K. If you reached out to the guys doing swiftshader development at Google, they could probably tweak it to perform better. Just being able to use all 64 cores would be a substantial boost, to say nothing about using AVX2. Getting its LLVM backend to target AVX2 should be another boost, if it does not do it already. My guess is that it does not. Also, swiftshader claims to support Vulkan. If you got it to take full advantage of the Rome processor, you might be able to use DXVK to do Direct3D10/11 to Vulkan to get it to render most modern titles on the CPU.
  12. You could use POSIX ACLs. NFSv4 ACLs are better, but they are not as well supported on Linux. The kernel support for them is fairly crippled. Samba translates NFSv4 ACLs to POSIX ACLs, but the translation is not possible to do perfectly. It works for most people though.
  13. Would you make a video explaining why none of AMD’s AIB partners have used their semi-custom silicon business to get better GPUs? https://www.amd.com/en/products/semi-custom-solutions They have all of these pseudo brands like gaming X or ROG, but they never do the one thing that would make them mean something by getting custom GPUs for them. It would be more risk, but they could for example get custom Navi GPUs that have more than twice the compute power at costs that are presumably just royalties and fabrication costs. Maybe the ones doing both nvidia and AMD cards would not do that for fear of retaliation from Nvidia, but I do not see what AMD only brands like XFX stand to lose by doing it. Driver support does not seem particularly difficult given that AMD seems to provide their semi-custom silicon customers with the sources for their drivers. Intel seems to be doing driver releases based on them: https://www.digitaltrends.com/computing/new-graphics-driver-for-intel-module-chips-arrives/ I imagine that that you guys would be in a position to talk to XFX and others to find out why they did not go this route. Is it just that they did not think about it?
  14. Review gaming motherboards made by an enterprise hardware ODM: https://www.supermicro.com/en/products/motherboards/desktop-gaming-boards Supermicro is basically the #1 manufacturer of server equipment in terms of quality. It would be interesting to see their gaming lineup reviewed, even if it lacks ECC. They seem to have given up because nobody in their target market seems to have noticed. The most recent motherboard that they made was for the 8th generation: https://www.supermicro.com/en/products/motherboard/C7Z370-CG-L It still would be nice to see a review, if only because of how unusual it was for them to make these.