Yes, I know, Linux is not Windows …
but as not long ago I turned to openSUSE again (after a long time).
I installed it dual-boot to my beloved Win7 (as long as I’m not familiar with openSUSE).
Both OS are on (different partitions of) a SSD.
One of the things I noticed was that applications I use quite often
seem to take more time to start than in Windows.
So much actually that it thwarts my workflow. Not joking.
In Windows I never noticed having to wait for Firefox to start,
in openSUSE I have to watch the hopping icon. Hop-hop-hop…
OK, I did the test now (in seconds, after starting time has stabilized):
Firefox
Win (FF 115.x): 2.4
openSUSE (FF 128.11): 5 (!)
LibreOffice Writer
Win: 1.1
openSUSE: 1.8
Texteditor
Win (notepad): 0
openSUSE (Kate): 1.5
As I said… this is not hair-splitting, but it kind of thwarts.
Moreover I can’t imagine this is the normal behaviour of openSUSE,
there must be something wrong?
It is difficult to give you a final answer regarding your question. But i would like to note following points.
You are comparing completely different applications:
MS Notepad is a barebone minimal texteditor whilst Kate is a fully fleged editor
Regarding Firefox:
you are comparing completely different versions
Regarding LO Writer:
not much difference.
Which raises the last question:
How did you meassure the start time?
As i do not have a MS Windows OS on any of my private machines since more than 25 years, i can only compare with my company laptop which runs Windows 10. The private linux machines (even the ones with lower HW spec than the company laptop), are running way faster and snappier than MS Windows. I did never meassure the time, but MS Windows applications often need way more time to open than any application on linux. Not to speak from the real startup times of the machines until you can use the desktop.
Yes, the difference between Kate and MS Notepad is a point. But I wonder if a comparable editor in openSUSE would do as fast (?).
Different Firefox version:
I can’t compare. But, to be honest I really doubt that the newer version will take double as long. There would have been an outcry in the community.
LO Writer
You’re right; not much difference. Actually even in Win7 it bothered me.
Maybe someone can give some more information (esp. about Firefox).
Or about the DE’s influence…
I never cared about Superfetch. But first starting an app takes always about twice as much. I thought it was a cache effect.
2. to 4. No.
How much takes it usually for Firefox to start for other users? (Have to watch hopping too?)
How are you timing startup?
A stop-watch on the screen or your smartphone?
Counting out-loud, “one thousand one, one thousand two, …”
“time” app at CLI?
xdotool app at CLI?
Something else?
An external stop-watch is sufficient. Doing few tests the difference between them is 0.1 s max…
Is it a normal starting time for Firefox (from SSD): 5s ?
And if yes ( ), are there ways to speed it up (change DE etc.)?
BTW:
I checked KWrite instead of Kate: It takes about the same time,
having to watch the hopping. OpenHop.
(A good tool is characterized by being invisible - be it a screw driver or SW).
To see something like that I have to boot up a 2007 Core2-Duo, 5400 rpm spinning rust, 2 GB RAM (barely usable Leap 15.6 Plasma).
On a 2015 laptop with SSD, Tumbleweed Gnome I see less than 1 s startup (then actually loading pages might depend on site and network), about 2 s on TW KDE in a VirtualBox VM.
Make sure nothing else is clogging your system when you perform such tests…
Asking other users to provide timings is not a proper comparison. Why? Mostly because of hardware differences. A user might have (or not) a faster motherboard and processor, a lot of RAM (or less), and a faster SSD (or an HDD).
Anyway:
FF = less than two seconds
LO Writer = some fraction over one second
Kate = a second or slightly less.
Never seen a hopping icon here.
BTW, you wrote:
“Win (notepad): 0”.
That’s nonsensical, because ‘0’ means “no time at all” … so it’s obviously already running
As there are many aspects so far, I would like to summarize some observations:
As the animated (hopping) application icons while starting didn’t seem to be noticed by anyone,
that points to some existing effect at my constellation. (Maybe there’s an animation setting at Plasma?).
I mentioned the HW already (At the same HW (dual boot) is Win7 doing quicker, boot and app start).
As a newbie I haven’t manipulated the installed 15.6 that might clog the system, only few simple things.
I did a closer look at Firefox, as the effect is largest and most bothering there.
It seems there’s a difference when starting at with an internet address.
v.115.x @ Win7 seems to show the pane at first place, before any loading etc.
v.128.11 @ Leap 15.6 doesn’t show the pane then, but only if some info from the i-address has arrived.
If I start Firefox @ Leap without address, and with empty start page, it will take abt. 2s.
Enough time for some hopping however
Not really.
With ‘0’ I didn’t mean “no time at all”, but something around the Planck time, or the smallest value larger than 0
In other words: “Not noticeable, immediately”.
As I already mentioned: The way it should be for a tool: invisible.
The startup indication aka hopping cursor is a standard Plasma setting. It is absolutely normal and should be known by real KDE/Plasma users…as it exists since ages.
But mostly you only see one or two hops and not more except an app is failing badly.
Different operating systems start applications differently. They are structured differently, load needed libraries in different ways (and potentially not always at startup), and binary formats have different characteristics.
I don’t know that I’d worry about a half-second (or even 2 seconds) difference in loading times.
Comparisons between different systems will yield different results. I launch everything from SSDs on my main system. Comparing my launch times to a system that has Win7 installed on it (and thus probably HDDs) wouldn’t be a fair comparison. For that matter, comparing between a system with a 5400 RPM drive and a 7200 RPM drive might not be a fair comparison. Or one with a disk transfer speed of 100 MB/s and one with a disk transfer speed of 140 MB/s. Or one that’s optimized vs. one that’s not and has to seek all over the place for different blocks of data for a running app.
There are so many things that can cause a difference in speed, even within the same system.
Some users know the advantages of having the startup notification enabled. Dependend on the size of your desktop icons and many other variables, this indicator can tell you directly what is wrong:
The application fails to start or you missed the application icon whilst clicking or your mouse is misbehaving…
That makes more sense (I see about 1 s plus network response too), so maybe your network or sites you visit are not the fastest on the planet.
And yes, I see 2-4 hops on Plasma (as many as 15 on the 2007 box ).
I agree in what you wrote.
Still I once again want to point out that I’m not comparing systems just “out of principle”,
as a kind of hair-splitting.
The situation is that I’m using dual boot Win7 and Leap 15.6, both in the same constellation
(sytem at SSD, data on HDD).
And I am used that Firefox opens immediately (with or w/o address), and at Leap I have to stop
and wait for 5s every time.
Firefox has a quite large cache stored in your /home/<username> folder and if that is on the HDD that might explain the longer time compared to some of us.
), are there ways to speed it up (change DE etc.)?
).