Jump to content

looking4awayout

Member
 View Profile  See their activity

  • Posts

    214

  • Joined

  • Last visited

  • Days Won

    2

  • Donations

    0.00 USD 

  • Country

    Italy

 Content Type 

Everything posted by looking4awayout

  1. looking4awayout
    I use the UOC Patch myself and I can bookmark sites just fine. Are you using DecentralEyes? Older versions interfere with the behaviour of the browser.
  2. looking4awayout
    Neat. I'm not a programmer at all but it's interesting to see that such a thing might be feasible on Windows.
  3. looking4awayout
    I've found out about it by reading a thread on VOGONS, where a user mentioned his RAM patch for Windows 98. While I never used his patches, being an XP user, it is a big loss for the entire community and the people who still RDD on Windows 98. Here's his obituary: https://kraussfuneralhome.com/tribute/details/858/Rudolph-Loew/obituary.html Rest in peace, @rloew. He was 67 years old and passed away the 11th of this month.
  4. looking4awayout
    So that means it could be possible to compile Serpent and New Moon 28 for SSE only CPUs in Windows? I'm not a programmer but would be interesting to see such a thing see the light of day.
  5. looking4awayout
    Now that's interesting... I wonder if it could be possible to do the same with the Windows version? I never had the opportunity to check Serpent out, as y'all know, my PC has an SSE only CPU. I'm quite curious to find out how an hypothetical SSE Windows version would perform with the UOC Patch and the other essential add-ons.
  6. looking4awayout
    I suppose the 9600 Pro is too slow for those tweaks to apply. At what resolution are you running the system? What Catalyst version are you using? I'm using an X1950 Pro, so that might be a factor.
  7. looking4awayout
    Intel and Nvidia cards seem to be optimized better for single tile accelerated window. ATI cards on the other hand, seem to run better with multiple tiles, based from what I have discovered today. Guess it's just the way Gecko and Goanna behave, or better to say, how the graphics card deals with those rendering engines. I look forward to read the report about the trick applied to older ATI cards.
  8. looking4awayout
    That's cool. After enabling those tweaks I mentioned in the UOC Patch thread, now scrolling performance with the X1950 Pro is almost on par with NVidia cards, I think it might be even faster than Nvidia cards now. But it's also true I'm using an X1950 Pro. I wonder if a 9600/9800 Pro will get the same benefits.
  9. looking4awayout
    Great! I look forward to read if it improves scrolling on your system too, if I remember correctly your RDD uses an ATI card, am I right?
  10. looking4awayout
    Just tested the latest test build of Firefox 45 ESR SSE. It runs much faster than the previous versions. While testing it, I also managed to finally fix the choppy scrolling issue that happens on Mozilla-based browsers above 800x600 if you use an ATI card. You can find the solution in the UOC Patch thread, check out Update II in the main post, or just check the link below: It will also increase scrolling performance at lower resolutions. But remember, it only works with ATI cards. Nvidia cards don't have this issue.
  11. looking4awayout
    After all this time, I'm finally happy to say that I have fixed the choppy framerate issue with ATI cards* on Mozilla based browsers (including @roytam1's) on resolutions above 800x600. In order to have an almost NVidia-like smooth scrolling experience, you need to set these three parameters to True in about:config: layers.enable-tiles - True layers.tiles.adjust - True devtools.performance.ui.enable-framerate - True (this one might be already set to True if you use the UOC Patch) This dramatically improves the framerate if you use an ATI card, so if you have experienced choppy scrolling above 800x600, you might have to check out if setting these three values to True in the about:config, will bring some benefits to your system. You might also have to tweak the framerate of your browser too, if you experience tearing. You can easily do that by editing the two values below: layers.offmainthreadcomposition.frame-rate layout.frame_rate The UOC Patch sets them to 120, but you might get a better performance at 60 FPS or 240, or 480, your mileage might vary according to your system. The sites that use their own scrolling routines, such as this forum, are still choppy unfortunately, but less than before. *= DO NOT, and I say do not, attempt to enable tiles with NVidia cards. Or else you will get extremely choppy framerate, rendering your browser unusable unless you edit your prefs.js file to set them back to false. This tweak only works on ATI cards since NVidia cards already run smooth without additional tweaking. EDIT: There is another extra value that you can set to True in order to get better scrolling performance. But it only works on Firefox 45 ESR SSE. I don't have a computer where I can test FF52 ESR based versions and beyond, so use this extra tweak below only on FF45 ESR SSE. Set the value below in this way: layers.async-pan-zoom.enabled - True DO NOT set this value to True in New Moon 27 and K-Meleon 76 Goanna, or else you will get unexpected 100% CPU spikes and the scrollbars will no longer be visible unless you revert the value to False. I do not guarantee that this extra value will work correctly on your system. If you notice worse performance with layers.async-pan-zoom.enabled, just leave it set to False as per default settings.
  12. looking4awayout
    I have left that parameter to be freely customizable by the end user. While I don't use the awesome bar at all, so I keep it set to 0, some might use it. So it's up to the user, whether he wants to use that feature or not. You can set it to 12 or to 1, or even disabling it completely by setting it to 0. Up to you, if you need it or not.
  13. looking4awayout
    Those three parameters are automatically enabled with the UOC Patch. It's just that ATI cards perform poorly on Mozilla based browsers at 1024x768 and beyond.
  14. looking4awayout
    I suppose Mozilla developers conceived them to have a sort of "one size fits all" configuration. So, instead of having potential instabilities with some machines, they preferred to set the browsers' behaviour to be slow but stable on most of the platforms (think of it as a "failsafe mode") for better compatibility, and then it would be up to the end user to tweak the browser to suit his needs. And thanks God it can be heavily tweaked, or else web browsing on old machines would be impossible: Chrome and its derivatives on my overclocked Tualatin runs awfully compared to Firefox 45 ESR SSE and the other Roytam1's browsers. Now to make things better we would need someone who can mod ATI drivers to get better performance out of Firefox and derivatives... Unfortunately a "UOC Patch" for ATI Catalyst drivers is something I can't do. 😛
  15. looking4awayout
    I have tested the new versions of Firefox 45 ESR SSE, New Moon 27 and K-Meleon 76 Goanna on the RDD and they work much better than the previous ones. Unfortunately I've been noticing that the X1950 Pro seems to struggle at accelerating Mozilla based browsers beyond 800x600, while scrolling is perfectly fine with my 6800GT. But I suspect it's an ATI driver issue, rather than a browser one. And this with Catalyst 6.12, the fastest ones for the X1950 Pro. NVidia cards are much, much better for Mozilla based browsers.
  16. looking4awayout
    You're most welcome. I'm glad the UOC Patch has helped making your web browsing experience on your computer better.
  17. looking4awayout
    So I have been trying to delve deeper into the "stuttering scrolling" issue, and I noticed some peculiar aspects on the two machines I've tested (the Tualatin and the Athlon 64 X2 at work). The issue seems to mostly affect Firefox 45 ESR SSE: if you scroll a page, in the beginning it's fine, but then it starts to stutter, even badly in some sites. Ironically enough, the site that suffers the most of this issue is Google, regardless of the user agent. But every site is affected by this, in a way or another. I have checked the browser even at stock configuration and without hardware acceleration enabled, and does the same thing, so it's a browser issue, or at least a problem caused by the way the browser renders the page. On the other hand, the issue is mitigated drastically in New Moon 27, to the point it's almost unnoticeable, and is inexistent in K-Meleon 76 Goanna. Now, I could have used New Moon 27 as a main browser, but unfortunately I can't since it doesn't support the earlier versions of YT 2 Player, an add-on that automatically parses Youtube links to VLC, only the latest version, which is much more bloated than the earlier ones, works. I spent all the night sleepless in order to figure out a way to mitigate this issue, but I haven't found anything that could fix it. There might be something that could help reduce the stuttering a little bit, but I need to test it thoroughly before I can implement it in the UOC Patch. As I'm not a programmer and I can't in any way contribute to improve the source code of FF45 ESR SSE due to my lack of programming skills, I'm doing my best. Will keep you all updated. EDIT: Nothing. None of the possible solutions I tried works. Looks like I'll have to abandon Firefox 45 ESR SSE and switch to New Moon 27, but only if I find a way to make the earliest version of YT 2 Player work on it. The issue lies in the rendering engine of the browser, so no patch can fix that, unless the code of the browser itself is modified.
  18. looking4awayout
    Sure I do! Once I have enough time I will put it in the main post, alongside a little tip for people who experience choppy frame rates at 120 FPS, usually with an ATI Radeon X1950 Pro. While I developed the UOC Patch with an NVidia card (my overclocked 6800GT), now I got a Radeon X1950 Pro in my Tualatin, and I noticed that at 120 FPS, scrolling is quite choppy, almost as if the browser is running at 30 FPS. I will post the workaround in the next days, alongside "sticking" Mathwiz's workaround with Mediafire and disabled ciphers. Work is keeping me quite busy lately so I had little time to post on the forum, and I apologize for that. EDIT: I've added @Mathwiz's tip in the main post. If you experience choppy scrolling (which is unrelated to the stuttering issue I mentioned in the New Moon thread), you can try to set these two values in the about:config, in the way showed below: layers.offmainthreadcomposition.frame-rate - 60 layout.frame_rate - 60 60 FPS seems to be the sweet spot on my Tualatin RDD with the X1950 Pro, while the 6800GT runs silky smooth at 120 FPS. I'm curious to see how much the upcoming Geforce 7800GS will fare. Hopefully, it should have way less overhead than the X1950 Pro since it uses Forceware 81.98, just like the 6800GT, but overclocks way more, from what I've read online.
  19. looking4awayout
    I think it might be something in the rendering engine of the browser. I've rolled back to an old version of Firefox 45 ESR SSE and New Moon SSE and the issue is still there, but less than before, but annoying enough. However, as stated previously, the stutter is less on New Moon than on Firefox. Happens even at stock settings, with a fresh profile and no extensions/Patch. If that thing gets sorted out, those browsers would run even better than they already do. No wonder I use them to browser the web on my oldtimer!
  20. looking4awayout
    @roytam1 Smooth scrolling is already disabled. @Mathwiz It happens both with a single tab and with more tabs. There's no additional slowdown with more tabs, but this stuttering during scrolling is less noticeable on New Moon, rather than on Firefox. It also happens when hardware acceleration is disabled. I don't think it's a dated architecture issue, because it also happens on my Athlon 64 at work, which handles Quantum just fine, although I prefer to use 45 ESR SSE as a testbed for the UOC Patch on more modern systems.
  21. looking4awayout
    After using both Firefox 45 ESR SSE and New Moon 27 SSE, I have to report something that affects both browsers, although New Moon 27 is affected in a minor way. Basically, if you scroll a webpage up and down, even without hardware acceleration and regardless of the platform, the scrolling begins to stutter, and in some cases even badly. I have done some tests both on my overclocked Tualatin and on an Athlon 64 X2 machine with a Geforce GT430, a machine fast enough to run Chrome and Internet Explorer without issue, but the same issue happens on New Moon 27 and Firefox 45 ESR SSE. It's not caused by the UOC Patch, because it happens regardless of the patch or the extensions installed in the browser. I hope @roytam1 can fix this issue, or at least mitigate it. On NM27 SSE the stutter is very minimum but is there, but on FF45 ESR SSE it's much more noticeable. Changing the framerate doesn't make things better either.
  22. looking4awayout
    It does work on Arctic Fox. I use it on my Mac Mini 1,1 and it does speed up the browser. I don't have any PowerPC Macintosh though, so I cannot test it in TenFourFox, as it doesn't run even with Rosetta installed. Compatibility wise, the Macintosh version should run without issue on 45 ESR based browsers too. Would be interesting to see how it performs on high end G4 and G5s. I had a guy test the patch on G3 and low end G4 Power Macintoshes (including a G4 Powerbook), but the performance was very disappointing, mostly due to the slow graphics cards used, an ATI Rage 128 on the G3 and a Geforce 2 MX on the Powerbook, and the meager amount of RAM (128MB on the G3 and 256MB on the G4). A fast graphics card that is optimized for OpenGL is essential on the Macintosh, as well as on Linux, as well as having at least 768MB of RAM upwards.
  23. looking4awayout
    UPDATE! After getting a semi-vintage Mac Mini from 2006 running System 10.6.8 "Snow Leopard", I've been able to port the UOC Patch to the Macintosh. Installing the patch is as simple as doing it on Windows. You just unzip the patch and put the UOC_Patch_Mac.js file into the following folder: Applications\[Browser folder]\Contents\Resources\Defaults\Pref The UOC Enforcer for 38 ESR browsers is the version that must be used with the Macintosh one. You can install the Enforcer on the Macintosh by putting the user.js file in the following directory: Macintosh HDD\Users\[Username]\Library\Application Support\[Browser Name]\Profiles\[Profile Name]\ In this way, you will have the UOC Patch + the Enforcer fully working even on your old Macintosh. Just like the Windows version, please test it on as many platforms as possible! I'm particularly interested to know how it performs on a PowerPC Macintosh, G4 and G5, and I'm curious to know how it would run on a G3 Power Macintosh (both the original iMac, the eMac and the beige G3 models), maybe it could even work with Classilla. The patch is offered AS IS, and to the people who still haven't updated the Windows version, please do it now and do not forget to try the UOC Enforcer alongside the Patch, for better performance!
  24. looking4awayout
    Oof! I'll fix them in a jiffy. EDIT: Fixed! Please, remember that the UOC Enforcer is just an add-on for the UOC Patch, and you must download both in order to fully optimize the browser. So while the Enforcer is not strictly required (although I recommend it), the UOC Patch is necessary, since it has the core settings.
  25. looking4awayout
    UPDATE! After some weeks of intense development, I have finally released the new version of the UOC Patch, build number is N2F. Compared to the previous version, this new one has been reworked almost entirely: many entries that point to my FF45 ESR SSE installation have been removed, shaving the size down to around 25KBs, making it even easier to fit in a floppy disk (3.5" or 5.25", your choice, for easier portability), but I've also reworked the loading routine of webpages, disabling the asynchronous panning and zoom feature, which instead of offloading the CPU, it is somehow bugged even on 45 ESR based browsers, causing an excessive spike in the CPU usage. Also, scrolling webpages while they are loading images should be smoother now, at least it is on my Tualatin RDD. I've also increased the memory cap of Javascript processes to 96MB, which seems the sweet spot between usability and performance, at least on my machine. But the biggest news is the introduction of the UOC Enforcer. The UOC Enforcer is a custom "user.js" file that must be placed into the browser's profile folder and that tells the browser to change the "stubborn" entries in the about:config that are hardcoded by the developer, making the UOC Patch work even better. You must remember though that the UOC Enforcer doesn't let you change the edits from the about:config, so if you want to change a "stubborn" entry, you must edit the user.js file directly (you can do it with Notepad). Therefore, I strongly suggest you to open the UOC Enforcer file and check out what are the stubborn settings, in case you want to change them in future. Most likely, you won't need to do that. While the Enforcer is optional, I strongly recommend it in order to get a better experience with the UOC Patch, as it tweaks the browser further. You can download the updated version of the UOC Patch and the UOC Enforcer in the main post. Please test the patch on as many machines as possible, and remember that it's offered AS IS, since I develop it in my free time, in order to squeeze some more power out of old computers and making them still useful online just like when they were new. Thanks!
×
×
  • Create New...