Open Ravenant1234 opened 2 months ago
it speeds up A LOT the browser
Really, it almost DOUBLE the speed on my pc
By developers of Chromium the "Quick Intensive Throttling" flag affects on power saving. By design it blocks Javascript execution on background tabs earlier, then without it. And saves up to ~10% of CPU time. https://chromestatus.com/feature/5580139453743104
May be speedup of the browser are side effect of blocking such scripts. It must be checked by ProcessExplorer. Because its strange if there are so much effect as you say.
I think using of NoScript
extension or something like this has more effect in such situation.
It must be checked by ProcessExplorer. Because its strange if there are so much effect as you say. What thing must search en ProcessExplorer?
What i must check?
By developers of Chromium the "Quick Intensive Throttling" flag affects on power saving. By design it blocks Javascript execution on background tabs earlier, then without it. And saves up to ~10% of CPU time. https://chromestatus.com/feature/5580139453743104
May be speedup of the browser are side effect of blocking such scripts. It must be checked by ProcessExplorer. Because its strange if there are so much effect as you say. I think using of
NoScript
extension or something like this has more effect in such situation.
Well, my pc is not latest gen i9 or even i3, we are talkin about an MUCH slower system (Pentium dual core e5200, 2gb ram DDR2!!! Windows 7) sooo maybe Anithing that i do will speedup such turtle
What I must check ?
You should enable check "CPU Usage History" and watch it then. Also determine sites, eating so much CPU.
I recommend use ˋ--process-per-siteˋ switch for this.
To more accuracy, such sites will be in separate renderer processes.
Also open internal "Chrome Task Manager" - Shift-ESC
, and do watch on CPU and memory.
And compare results with scripts enabled and disabled on that sites (by script blockers like NoScript). And results by using such "throttling "flag.
Also watch memory allocated by process (scripts can eating it also). And disk operations, if you are not using SSD drive here.
All this will affect on performance.
Will do and report it
Also determine sites, eating so much CPU
Well, there is no site in particular..... maybe youtube and some other video playback sites like twitch i cant imagine any site that eat cpu most than that maybe a huge page
Yoy can see CPU eating per site, if running in --site-per-rpcess mode, and watch CPU colum in Chrome Taskmanager for it (it will be separate per site), Then do tests with enabled and disabled scripts, (above sites will not function without scripts of course), and with "throtling" flag then. Also try separate ad blocker extensions.
Well this is weird i have discovered something When the flag off and Ublock off the browser does this
Flag off UBlock On
Flag On Ublock Off
Flag on ublock on 2
FLAG OFF UBLOCK OFF
I will make more tests
Notice the high network spikes in the last and first (and it didnt stop after some minutes) and the constant cpu usage
will do the tests in process-per-site
I think using of
NoScript
extension or something like this has more effect in such situation.
I have tried "NoScript" and alternatives but there are tough to configure (bc some requires configuration site-by-site), i prefer this chrome flag instead bc is global
What I must check ?
You should enable check "CPU Usage History" and watch it then. Also determine sites, eating so much CPU.
I recommend use ˋ--process-per-siteˋ switch for this.
To more accuracy, such sites will be in separate renderer processes.
Also open internal "Chrome Task Manager" -
Shift-ESC
, and do watch on CPU and memory.And compare results with scripts enabled and disabled on that sites (by script blockers like NoScript). And results by using such "throttling "flag.
Also watch memory allocated by process (scripts can eating it also). And disk operations, if you are not using SSD drive here.
All this will affect on performance.
will do
So if its really has such effect and do job for blocking of scripts on background pages, -- its cool of course. (If this feature not enabled by default). I will use it also.
So if its really has such effect and do job for blocking of scripts on background pages, -- its cool of course. (If this feature not enabled by default). I will use it also.
Yes, you should try it
So if its really has such effect and do job for blocking of scripts on background pages, -- its cool of course. (If this feature not enabled by default). I will use it also.
Can u restore it to chrome://flags ?
Its a question to guy called "win32ss", which do such coding here, restoring of old Chromium commits, and building of the browser.
@win32ss ?
yes,
Yoy can see CPU eating per site, if running in --site-per-rpcess mode, and watch CPU colum in Chrome Taskmanager for it (it will be separate per site), Then do tests with enabled and disabled scripts, (above sites will not function without scripts of course), and with "throtling" flag then. Also try separate ad blocker extensions.
yes but i cant see a graph for "Historical" regarding the cpu usage, and in Process Explorer yes
maybe for reporting it might be more useful to use process explorer? because u can actually see the "peak"
well actually i have done the test again comparing the flag with a scrpt blocker in process per site (More specifically this: https://chromewebstore.google.com/detail/script-blocker-ultimate-n/onedokhafbakmhkblnokgjcliehfognj ) and yes, the flag based on the cpu graph on procexp is similar to an script blocker in reddit
but in youtube does the job even better
so.... maybe is better? idk please be free of doing tests by yourself
if u restore the flag it would be awesome even if it will be perma-flag i dont care because is actually useful and if is a flag it will not do any harm because it will be disabled by default and if is enabled didnt create more overhead to the browser than an extension
I think to show more differences, you shoult open some more pages, say 10 from one site. Or 20 per site. Or more.
So other pages will be on background tabs with their scripts running.
And they will be handled be one process. And CPU usage should be increased to be more visible.
The CPU history I recommend to watch not for the whole system, but for target process, whose Process ID
you can see in "Chrome Task Manager".
(If you interested to do such experiment of course).
This is example, which I see for 1 youtube page on my XP notebook without Hardware acceleration.
I think to show more differences, you shoult open some more pages, say 10 from one site. Or 20 per site. Or more.
My pc will explode if i do such Dont forget, i have a 2gb ram ddr2 667mhz meme-pc
and is my most powerful setup right now
What should i do instead?
My pc will explode if i do such
:) Not so much in such case. Less. 5-10 pages. Youtube pages should not be in "playing video" state, but paused. It should open 10 pages on your computer at least. Without explosion )).
ok :)
My pc will explode if i do such
:) Not so much in such case. Less. 5-10 pages. Youtube pages should not be in "playing video" state, but paused. It should open 10 pages on your computer at least. Without explosion )).
it will help if i install some extension to disable autoplay?
some extension to disable autoplay
Not all of them do their job fine. But I'm currently use this: https://chromewebstore.google.com/detail/cmedhionkhpnakcndndgjdbohmhepckk https://chromewebstore.google.com/detail/autoplaystopper/ejddcgojdblidajhngkogefpkknnebdh https://chromewebstore.google.com/detail/disable-html5-autoplay-re/cafckninonjkogajnihihlnnimmkndgf
May be another will work.
btw i forgot to ask..... how are u with the supermium 123 stabilization? is close? maybe.... less than 1 month or less than 2 weeks? or u prefer to wait more to fix some bugs ?
I know nothing about new versions. I do not build it. I personally prefer to wait and to get stable version with fixed bugs (which must be listed as I described earlier #421 ). Or you can use Thorium builds (which refreshes more frequently, but mostly has same bugs as here, and no one wants to fix them there :),
I got used to living with the version that exists now. (in other words, with having workarounds for issues). Even if next never be release, Life is going. I will fix issues by myself (which I can). And its about all software. Its my philosophy.
Otherwise I wouldn't be using XP and sometime relatively old hardware. )), And would go participate in upgrade races.
I know nothing about new versions. I do not build it. I personally prefer to wait and to get stable version with fixed bugs (which must be listed as I described earlier #421 ). Or you can use Thorium builds (which refreshes more frequently, but mostly has same bugs as here, and no one wants to fix them there :),
oh ok
And they will be handled be one process.
U mean process-per-site?
It could be returned to the flags table, but I don't see any major difference with this flag myself.
btw i forgot to ask..... how are u with the supermium 123 stabilization? is close? maybe.... less than 1 month or less than 2 weeks? or u prefer to wait more to fix some bugs ?
A Windows 7 user reported freezes on YouTube pages. So I am waiting for him to report that my proposed fix (since this one appears to be render-specific, to disable the recent modification to D3DImageBackingFactory::CreateSharedImage to add WebGPU on Windows 7) will have resolved that problem. And then there is the password manager bug I am going to fix on XP.
It could be returned to the flags table, but I don't see any major difference with this flag myself.
btw i forgot to ask..... how are u with the supermium 123 stabilization? is close? maybe.... less than 1 month or less than 2 weeks? or u prefer to wait more to fix some bugs ?
A Windows 7 user reported freezes on YouTube pages. So I am waiting for him to report that my proposed fix (since this one appears to be render-specific, to disable the recent modification to D3DImageBackingFactory::CreateSharedImage to add WebGPU on Windows 7) will have resolved that problem. And then there is the password manager bug I am going to fix on XP.
Yay! Awesome, u have solved a lot of bugs since the last one, hope that this version will be even more better than the current one (wich is really good for me at least)
Regarding the flag, yes, please return it
I am still alive...... hahahahah
The flag is not present in chrome://flags but if i set it manually IT SEEMS to work (pages load faster)
proposal: restore the flag to be present in chrome://flags