Why does Carolite-1.2 run so slowly?

[Mike7]

Hi, starhawk.

two things are bogging your system down, Mike7.

Only two? <grin>

One, XFCE is quite heavy. Like "GNOME2 minus a pound or two" heavy. No, really. Look here. MATE (in case you don't know) is basically the "we continued it because they wouldn't" edition of GNOME2.

According to that article, XFCE is one of the lighter desktop environments, just behind LXDE.

The other issue is that your RAM is only a single gig. You should have two.

I agree, and I will do as you suggest and install a 2Gb RAM stick when I'm back in the States in the fall. That is, if it can be done. I'm not sure the 1000HA can take more than 1Gb.

If you have a little money you can also put an SSD in there.

I wouldn't know how to install it, and I don't think I want to start messing around with the insides of this netbook. Stuff's pretty tightly packed in there. But it's a good thought.

M.

[Mike7]

mikeslr-

That's a lot of food for thought. While I'm mulling it over, I'll reply to a few points.

[quote]Even if you only had the current firefox pet installed into your SaveFile, if you were to entirely load that SaveFile on boot –as apparently you do using the “copy

[rg66]

Looking at the specs for your Asus eeePC 1000HA, the processor is probably part of the problem. "Intel® Atom™ Processor N270 (512K Cache, 1.60 GHz, 533 MHz FSB)". If you have a few apps running such as firefox, mplayer etc, what is the "CPU usage: %" in taskmanager?

[Mike7]

starhawk-

I know for a fact that XFCE is heavier than eg JWM, IceWM

Sure, but those are the fastest environments there are.

I put up with XFCE, though, because the slowdown is worth it to me. . . even on the netbook it's worth the speed penalty to me. There are just some things that I can't stand about JWM/ROX.

I wouldn't go as far as that. . . but I do like Carolite's XFCE look.

That's ultimately the question -- what's important to you? If speed is all you care about, to the point that you don't mind being stuck with old apps -- go with an older Pup

.
Well, from what I've seen, there's a high percentage of older apps on all Puppies, even the newer ones. Much of the updating of apps is purely superficial and cosmetic.

If you care about having a computer that doesn't look "stuck in the 90s" and you're willing to give up some speed... LXPup or one of rg66's X-Pups is for you (X-Precise, X-Slacko, X-Tahr)

I might try them out. I'm getting a little dizzy, though, from all these many Puppy versions.

if you want both speed and looks you're kinda screwed

I really can't see why they should be mutually exclusive.

M.

[Mike7]

Hi, rg66.

If you have a few apps running such as firefox, mplayer etc, what is the "CPU usage: %" in taskmanager?

I have this Firefox window open and Mplayer running,, and the CPU reading in Taskmanager is around 45%. If I shut off Mplayer it drops down to around 25%. Still high.

M.

[rg66]

I have this Firefox window open and Mplayer running,, and the CPU reading in Taskmanager is around 45%. If I shut off Mplayer it drops down to around 25%. Still high.

M.

With a dual core 3Ghz CPU and taskmanager open I started firefox and it jumped up to 33% then dropped to 7%. Yours is probably closer to 100% and would explain why apps take longer to open.

[Mike7]

• DISCREPANCIES IN REVISED CPU AND RAM USAGE

Running: this Firefox browser window, Top in a Terminal, Free in another Terminal, and Task Manager

Top
Mem: 828904K used, 195484K free, 0K shrd, 28224K buff, 616728K cached
CPU: 22% usr 6% sys 3% nic 67% idle 0% io 0% irq 0% sirq

TaskMan
CPU usage 33% Memory 186MB of 1000MB used

Code: Select all

# free
             total         used         free       shared      buffers
Mem:       1024388       827888       196500            0        28224
#

Such a large discrepanciy between Top and Taskman in RAM usage makes the data appear unreliable. Is Top right, with 829Mb used memory, or Taskman, with 186Mb? Or are they showing different things?

Free seems to confirm the correctness of Top as opposed to Taskman.

Maybe Taskman has it backwards and is showing used memory as free and vice-versa. That would approximate the three memory readings. Is it possible that such a mistake has been made in TaskMan? This appears to be the only reasonable explanation.

There is no discrepancy in CPU usage. TaskMan reads more frequently than Top, that's all. Which further persuades me that an error has been made in used memory reporting in one of these applications.

So, I must assume that even with so few applications open, the major part of my RAM is in use, leaving little for other chores. This alone can probably account for the sluggishness of this OS.

M.

[Philh]

You could try with
pfix=nocopy to free up some more ram.
And try without a savefile pfix=ram to make sure its fast enough then and not some other issue.

[mikeslr]

Hi again Mike7,

For an accurate explanation we'll need someone with actual technical knowledge to chime in. The following is my “layman's

[8Geee]

Mike7... EDITTED... wrong question, sorry.

Brand of the USB stick? And is it an older one? Say 3 yrs+ old.
Also I presume, that USB stick is USB2 and not USB 1.1.

Regarding personal expeeriences here. My new stubby Sandisk Cruzer Fit is 50% faster than my Patriot X-mini. My startups Power Button to desktop has gone from 90 seconds to 55-60 seconds. A marked improvement.

Menu --> System --> PupSysInfo --> Devices --> Drive Storage --> Drive Test

Click on the Drive Test for Read Speeds.

[Mike7]

Hi, Philh.

You could try with
pfix=nocopy to free up some more ram.
And try without a savefile pfix=ram to make sure its fast enough then and not some other issue.

Thank you for these interesting suggestions. I'll try them out as soon as I have time, hopefully this weekend.

Mike

[Mike7]

8Geee-

Brand of the USB stick?

I'm using a Sandisk Cruxer fit, too (for Carolite).

Menu --> System --> PupSysInfo --> Devices --> Drive Storage --> Drive Test

That menu item doesn't exist in Carolite-1.2, and I couldn't find anything called "PupSysInfo" or "Drive Test" in Carolite.

Mike

[Mike7]

Philh-

I added pfix=nocopy to the boot menu by doing "e" at the boot menu screen. Then the first thing I did after it booted up was a RAM check with the "free" command in a terminal. Free memory was around 8xxMb, much better than the usual 5xxMb:

Code: Select all

# free
             total         used         free       shared      buffers
Mem:       1024388       216664       807724            0        24224
-/+ buffers:             192440       831948
Swap:            0            0            0
# 

However, after opening and closing a couple of apps, I checked again and it was back down to 500:

Code: Select all

# free
             total         used         free       shared      buffers
Mem:       1024388       612356       412032            0        34212
-/+ buffers:             578144       446244
Swap:            0            0            0
# 

So, while pfix=nocopy may initially provide more free RAM, that memory apparently gets used quickly. OTOH, applications did seem to be opening and functioning more rapidly, which I can't see any reason for if they're not running in RAM (or are they?).

All very confusing.

I have no idea what the pfix=nocopy actually did (or prevented), and I don't know how to find out (i.e. which log file, if any, to look at).

At this point, I'm asking for some help from the people in the forum who know how all this actually works.

Anyone out there who can straighten this out?

Mike

[Philh]

I think stuff gets left in ram in case you open applications again.
Then if the ram is needed it gets flushed out.
It should only be a problem and slow down if you were getting close to zero.

[Mike7]

I think stuff gets left in ram in case you open applications again.Then if the ram is needed it gets flushed out. It should only be a problem and slow down if you were getting close to zero.

That makes sense. But why would applications run faster after a pfix=nocopy boot? It's hard to believe that apps in flash memory run faster than apps in RAM.

M.

[Mike7]

Hi, mikeslr.

Got a little hung up by the server switch, but it gave me time to think over your post from last Monday.

[quote]The following is my “layman's

[sheldonisaac]

those who built opera 12.16 constructed it so that its configuration and cache would be in the opera folder – you don't have to move them out of the SaveFile.

I used to symlink /root/.opera to /mnt/home/RealOpera. Then savefolders were invented, so I thought why bother with that. Don't understand how these things work, but I've gone back to the symbolic linking.
Since I don't have nor understand your eeePC(?), I can't say whether what works for me would work for you.

Maybe Opera is the way to go...
I'll buy a 2Gb RAM card ASAP, and I've put 50Mb as the cache limit in the FF preferences. (Could I make it zero?)

I'm using Opera 12.16, and have disk cache set to zero, as well as 'empty on exit' (does that make any sense?)
The 'memory cache' is set to Automatic; other choices range from 'off' to 400MB.

[Mike7]

sheldonisaac-

I'm using Opera 12.16, and have disk cache set to zero, as well as 'empty on exit' (does that make any sense?)

Not really, but it's what I had for a while. Just in case. . .

The 'memory cache' is set to Automatic; other choices range from 'off' to 400MB.

I'm unclear as to what and where these different caches are. In my FF preferences I find "web content cache" and "application cache". Are these what you are referring to as "disk cache" and "memory cache"? What about "other choices"?

Mike

[sheldonisaac]

sheldonisaac-

I'm using Opera 12.16, and have disk cache set to zero, as well as 'empty on exit' (does that make any sense?)

Not really, but it's what I had for a while. Just in case. . .

The 'memory cache' is set to Automatic; other choices range from 'off' to 400MB.

I'm unclear as to what and where these different caches are. In my FF preferences I find "web content cache" and "application cache". Are these what you are referring to as "disk cache" and "memory cache"? What about "other choices"?

Mike

Don't know.
How does one find experts?

[Mike7]

sheldonisaac-

How does one find experts?

They seem to be everywhere <grin>.

M.