Puppy Linux Discussion Forum Forum Index Puppy Linux Discussion Forum
Puppy HOME page : puppylinux.com
"THE" alternative forum : puppylinux.info
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

The time now is Sun 08 Dec 2019, 18:57
All times are UTC - 4
 Forum index » Advanced Topics » Puppy Projects
Xenialpup64 CE 7.5 / 25 Nov 2017
Moderators: Flash, JohnMurga
Post new topic   Reply to topic View previous topic :: View next topic
Page 5 of 56 [829 Posts]   Goto page: Previous 1, 2, 3, 4, 5, 6, 7, ..., 54, 55, 56 Next
Author Message
Sailor Enceladus

Joined: 22 Feb 2016
Posts: 1565

PostPosted: Sat 30 Jul 2016, 00:06    Post subject:  

@sszindian: Something to consider. Newer versions of htop show higher memory usage. Here's a screenshot showing:

- htop 1.0.3 on the top (shows 101MB used)
- htop 2.0.2 in the middle (shows 250MB used)
- busybox 1.23.1 top on the bottom (shows 635MB used)

All running at the same time.
Screenshot.png
 Description   
 Filesize   68.43 KB
 Viewed   1829 Time(s)

Screenshot.png

Back to top
View user's profile Send private message 
LateAdopter

Joined: 27 May 2011
Posts: 316
Location: Reading UK

PostPosted: Fri 05 Aug 2016, 05:47    Post subject:  

LateAdopter wrote:

It looks like it's a kernel 4.6 regression. I just built 4.6.3 using the same DOTconfig as for 4.4.4 and it performs the same as your 4.6.3. So maybe MSR doesn't matter after all.


Hello 666philb
Xenialpup64 is working very well and I have installed tvheadend, mpv, gnome-mpv, Kodi17 alpha2. No problems from Xenialpup64..

I have investigated the media player performance problem with the 4.6.3 kernel. It looks as though the Intel p-state governor is more broken than usual, at least with Braswell. It does not clock up to maximum turbo before the media player starts noticeable skipping. I have tested kernel 4.7.0 and that has the same problem, but 4.4.4 works properly.

My workaround is to disable pstate and use cpufreq ondemand. However, with its default settings, ondemand is worse than pstate, but they are adjustable. Increasing sampling_down_factor to 80 makes it work properly.

So I have
intel_pstate=disable as a kernel boot parameter, and
echo 80 > /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor in /etc/rc.d/rc.local

Note: Modern Intel processors with hardware power management (Sandy Bridge onwards) make up their own mind about core clock frequencies. The numbers reported by cpufreq are quite different from those reported by turbostat which reads back from the CPU.

Thanks for Xenialpup64
Back to top
View user's profile Send private message 
live

Joined: 10 Feb 2010
Posts: 223

PostPosted: Fri 05 Aug 2016, 17:03    Post subject:  

@sszindian
All 3 shows:
* different RAM usage
* different CPU usage
* different nb of processes

Have you tried to monitor which each of them separetly, as maybe they are accessing the same data at the same time and this could lead to wrong measurements.
Back to top
View user's profile Send private message 
sszindian


Joined: 24 Apr 2010
Posts: 799
Location: Pennsylvania U.S.

PostPosted: Fri 05 Aug 2016, 17:26    Post subject: Ram Usage  

@live

Those graphics are from Sailor Enceladus, not me. I don't know what configuration he used or what all was running at the same time?

At any rate, doing some research, Distrowatch has a recent article about this subject and it's their belief that having as much in ram as possible is 'better' contrary to what I believed over the years. Makes one wonder?

Sure is a fast/responsive program and operates without major issues so far.

>>>---Indian------>

_________________
Cloud Computing For Every Puppy (a .pet)
http://murga-linux.com/puppy/viewtopic.php?t=69192
Back to top
View user's profile Send private message 
rufwoof


Joined: 24 Feb 2014
Posts: 3671

PostPosted: Fri 05 Aug 2016, 18:47    Post subject:  

Sailor Enceladus wrote:
@sszindian: Something to consider. Newer versions of htop show higher memory usage. Here's a screenshot showing:

- htop 1.0.3 on the top (shows 101MB used)
- htop 2.0.2 in the middle (shows 250MB used)
- busybox 1.23.1 top on the bottom (shows 635MB used)

All running at the same time.

Reporting memory utilisation is subjective. Looks like the different versions/apps chose to report in different ways. htop I believe counts cached and buffers as being 'free' i.e. they can be dumped at any time by the system to make space for other things (being used, but free to be used elsewhere).

Debian Jessie KDE 64 Frugal, running kinfocenter alongside htop (see attached).

htop mem coloured bars correspond to used, buffers, cache in different bar colours. Which in my case seems to conflict with what kinfocenter is suggesting

# free
provides other source of information.

Its all a relative type thing. If application data is near filling up memory and there's little free/cache/buffer allocations, then its at a stage where swap is being needed and the system isn't running efficiently. i.e. in htop that would be if the mem green bars were running right across and the numeric figure was low.
mem.jpg
 Description   
 Filesize   80.16 KB
 Viewed   1598 Time(s)

mem.jpg

Back to top
View user's profile Send private message 
mcewanw

Joined: 16 Aug 2007
Posts: 3200

PostPosted: Sat 06 Aug 2016, 02:16    Post subject:  

Actually, decisions were made higher up in the LInux hierarchy regarding how memory statistics should more correctly be reported. Fred and I both noticed the different reporting of not only free versus htop but also some versions of htop versus others. In particular, the reporting mechanism was noticably different between the htop used in MintPup and that used in the newer XenialDog.

SInce a statistic about memory usage, system performance is only as good as the tools used to measure it, I considered this an important matter to get to the bottom of. I ended up reading the free and htop source code just to see what the differences were in the newer htop reports and documented it back in May here:

http://murga-linux.com/puppy/viewtopic.php?p=904344#904344

The differences are all very technical, but absolutely not imagined or accidental. The summary is that the measurements are now being calculated differently in newer versions of these utilities because the inner-circle kernel developer guys believe the new reporting mechanism is more realistic in terms of how much Mem really is free and so on. The end result is that we cannot use htop to reliable indicate performance/resource-usage of different distributions since the measurement tools themselves are now reporting different things.

William

_________________
github mcewanw
Back to top
View user's profile Send private message Visit poster's website 
Belham

Joined: 07 Jan 2016
Posts: 133

PostPosted: Sat 06 Aug 2016, 08:57    Post subject:  

mcewanw wrote:
Actually, decisions were made higher up in the LInux hierarchy regarding how memory statistics should more correctly be reported. Fred and I both noticed the different reporting of not only free versus htop but also some versions of htop versus others. In particular, the reporting mechanism was noticably different between the htop used in MintPup and that used in the newer XenialDog.

SInce a statistic about memory usage, system performance is only as good as the tools used to measure it, I considered this an important matter to get to the bottom of. I ended up reading the free and htop source code just to see what the differences were in the newer htop reports and documented it back in May here:

http://murga-linux.com/puppy/viewtopic.php?p=904344#904344

The differences are all very technical, but absolutely not imagined or accidental. The summary is that the measurements are now being calculated differently in newer versions of these utilities because the inner-circle kernel developer guys believe the new reporting mechanism is more realistic in terms of how much Mem really is free and so on. The end result is that we cannot use htop to reliable indicate performance/resource-usage of different distributions since the measurement tools themselves are now reporting different things.

William



Hi William,

As long as htop keeps providing the correct PID (only ever open it to find the pid when I need to kill something), I guess I'll keep on using it...haha


P.S. I see in your murga-linux sig, it says "New Zealand'. You have any knowledge of the LinuxLite crew in NZ doing their developments? Some months back I downloaded their latest LinuxLite (3.0). I saw that the Firefox they included still had everything set (in its about:config file) to antipodian browser search urls/settings and such. Have been pleasantly surprised with their distro, as it's clean and everything works, which is high praise with all these recent distros popping up that leave users floundering.
Back to top
View user's profile Send private message 
greengeek


Joined: 20 Jul 2010
Posts: 5624
Location: Republic of Novo Zelande

PostPosted: Mon 22 Aug 2016, 01:53    Post subject:  

Booted a Toshiba M700 laptop successfully (from CD) but I have two issues so far:

1) If i click an mp3 file it does not play at all.

2) If i rightclick an mp3 file and select "send to" deadbeef it plays fine but when I click the "close" cross the deadbeef gui disappears but leaves the audio still playing. Have to use pprocess to kill it.
Back to top
View user's profile Send private message 
Sailor Enceladus

Joined: 22 Feb 2016
Posts: 1565

PostPosted: Mon 22 Aug 2016, 14:09    Post subject:  

greengeek wrote:
2) If i rightclick an mp3 file and select "send to" deadbeef it plays fine but when I click the "close" cross the deadbeef gui disappears but leaves the audio still playing. Have to use pprocess to kill it.

Heheh that also happens to me in pupradio. I have to remember to press stop instead of close first. I wonder if that's by design (in case you want it to keep playing without having the program staying open)?
Back to top
View user's profile Send private message 
greengeek


Joined: 20 Jul 2010
Posts: 5624
Location: Republic of Novo Zelande

PostPosted: Mon 22 Aug 2016, 15:22    Post subject:  

Belham wrote:
Some months back I downloaded their latest LinuxLite (3.0)
Based on your comment I had a look at LinuxLite and the wikipedia page says the following:
Quote:
Software developed by Linux Lite:
- Lite Info: an application that retrieves hardware information from the users computer and shares it with the community
Do you have any idea if that utility can be disabled? Or is it a function that runs automatically and shares your hardware info invisibly in the background?
Back to top
View user's profile Send private message 
Moat


Joined: 16 Jul 2013
Posts: 967
Location: Mid-mitten

PostPosted: Tue 23 Aug 2016, 00:24    Post subject:  

Sailor Enceladus wrote:
I wonder if that's by design (in case you want it to keep playing without having the program staying open)?


Likely so, and I do believe Deadbeef's preferences allow setting it to either minimize to tray or close - and usually a Deadbeef install defaults to minimizing, not closing. I like having the tray feature with the GUI hidden, myself, and right-clicking the tray icon for basic play controls.

IOW - give the preferences a peek...

Bob
Back to top
View user's profile Send private message 
ravi23681

Joined: 24 Aug 2016
Posts: 1

PostPosted: Wed 24 Aug 2016, 03:09    Post subject: need support  

need support to my usb wifi ralink wifi and usb bluetooth dongle to work...
Back to top
View user's profile Send private message 
jd7654

Joined: 06 Apr 2015
Posts: 297

PostPosted: Wed 24 Aug 2016, 15:45    Post subject:  

There appears to be some problem with the suspend with Xenial pup on my machine. Once suspend is initiated, initially it seems normal as hard drive churns, screen goes blank and power led goes off, but only for a split second before it immediately wakes up again.
Actually, this same problem carries over from Tahr Pup too. But I also have Ubuntu Tahr installed on this machine, in the form of elementary OS, and suspend works just fine there.
Suspend does work fine with Slacko64 and Fatdog too so I know its not a hardware problem.

Also, noticed some BIOS beep in the bootup process. Can't see exactly where as the screen is blank during that time, but its somewhere late in the process, between the login prompt and the desktop appearance. Don't know if this is some kind diagnostic beep for alpha version?



▶—— BASE SYSTEM ——◀

PC Manufacturer: Hewlett-Packard
Product Name: HP Pavilion dm1 Notebook PC

Motherboard Vendor: Hewlett-Packard
Product Name: 1611

BIOS Vendor: Hewlett-Packard
Version: F.12
Release Date: 04/26/2011

AMD E-350 Processor
Max Speed: 1600 MHz
Current Speed of Core 0:1600 MHz, 1:1600 MHz
Core Count: 2
Thread Count: 2

Frequency governor : ondemand
Freq. scaling driver : acpi-cpufreq


Linux Kernel: 4.6.3 (x86_64)
Kernel Version: #1 SMP PREEMPT Mon Jun 27 19:59:25 BST 2016

Distro: xenialpup64 7.0.7
Window Manager: JWM v2.3.5
Desktop Start: xwin jwm
Back to top
View user's profile Send private message 
LateAdopter

Joined: 27 May 2011
Posts: 316
Location: Reading UK

PostPosted: Wed 07 Sep 2016, 09:04    Post subject: run-as-spot  

As a refugee from Opera 12.16 I'm trying out Vivaldi, but I want to get it to run as spot before I let it loose on the internet.

If I try run-as-spot it fails. In a terminal I get this:
Code:
root# run-as-spot vivaldi-stable
su: can't execute '/bin/sh': Permission denied

But sh is a symlink to bash which has execute enabled for everybody.

Any ideas what the real problem is? Thanks.

Vivaldi will run OK as root, as long as I specify a user directory on the command line.

EDIT:
Code:
su spot
gets the same error. So its nothing to do with Vivaldi.

EDIT2: I was wrong. It was installing Vivaldi that broke SPOT.

EDIT3: It seems the chromium post install script breaks SPOT. Extracting the vivaldi files and moving to them to /opt they work without problems.

Last edited by LateAdopter on Wed 14 Sep 2016, 08:35; edited 3 times in total
Back to top
View user's profile Send private message 
Sailor Enceladus

Joined: 22 Feb 2016
Posts: 1565

PostPosted: Wed 07 Sep 2016, 09:22    Post subject:  

LateAdopter wrote:
Vivaldi will run OK as root, as long as I specify a user directory on the command line.

Oh yeah, that was the issue with root. I had to add --user-data-dir to the end. It's all coming back to me now - Celine Dion Laughing
Back to top
View user's profile Send private message 
Display posts from previous:   Sort by:   
Page 5 of 56 [829 Posts]   Goto page: Previous 1, 2, 3, 4, 5, 6, 7, ..., 54, 55, 56 Next
Post new topic   Reply to topic View previous topic :: View next topic
 Forum index » Advanced Topics » Puppy Projects
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum


Powered by phpBB © 2001, 2005 phpBB Group
[ Time: 0.2484s ][ Queries: 12 (0.0477s) ][ GZIP on ]