On Friday 25 June 2004 7:09 am, Alexandr Malusek wrote:

linuxjim writes:

...

I upgraded one of my 8.2 boxes to 9.1, and expected to find 9.1 a bit faster than the older version. So far that ain't happening. I did the upgrade, not fresh install. I had to clear out some older programs to eliminate dependencies, but other than that all seem to go fine.

It boots in about the same amount of time, till it gets to the login screen. After that the login process is much slower, when KDE gets to the desktop, it takes several minutes before it is usable.

Here are some numbers from my 550 MHz PIII with a 5400 rpm IDE disk:

- boot to runlevel 5 with kdm: 1 min 36 s - start of a KDE session: 1 min 5 s (it includes 15 s for applications like xmms, xosview, ...) - Install or remove software module of YaST2: 40 s

...

The desktop graphic shows as to the desktop icons, but the menu bar does not show for several minutes.

Try to create a new user to eliminate the effect of old ~/.kde and other files.

My experience is the same as the one already reported here: the boot time is approximately the same as for 9.0 but KDE is a bit less responsive probably due to all those jumping icons, ... It takes 3-4 s to open a shell (konsole) window.

-- A.M. Hi Alexander,

On my machine it takes about 37 seconds for the shell to come up. No kidding. It wasn't fast in 8.2, but it was faster than that. Not at all sure what is going on. Ill try a new user to see what that does. Loggin in to root takes about the same time as existing user. Jim Flanagan

Fred Miller wrote:

On Friday June 25 2004 11:12 am, Philip Amadeo Saeli wrote:

[snip]

...

Anyone have any comparisons between 9.0 and 9.1? Was the slowness introduced in 9.0 and then carried over into 9.1, or is 9.1 faster or slower than 9.0? 8.2 to 9.0 definitely brought a huge hit.

9.1 is slower.....install 9.0 or stay with 8.2.

Well, that's certainly not true for all hardware. I've got a 300MHz PII laptop that ran fairly well with 8.2. Its performance under 9.0 was, to say the least, abysmal. I was beginning to think it had finally reached the end of its usefullness as my desktop for work, unless I wanted to stick with 8.2. Then I installed 9.1 on it and, boy howdee, it runs great again. In fact, it feels faster than 8.2. Steve

On Tue, 29 Jun 2004 03:40 pm, Fred Miller wrote:

Some have great success and some don't....it's one of the reasons I think it has something to do with the 2.6 kernel and some or all of the modules.

Good point, could the people who are getting good results post some more details? Processor, RAM, chipset, and output of lsmod This setup DOESN'T work well: suse:/etc # lsmod Module Size Used by raw 39840 0 udf 85380 0 edd 9368 0 joydev 10304 0 sg 35616 0 st 39068 0 sd_mod 20224 0 sr_mod 16292 0 scsi_mod 108748 4 sg,st,sd_mod,sr_mod ide_cd 36740 0 cdrom 36764 2 sr_mod,ide_cd nvram 8456 0 snd_seq_oss 31232 0 snd_seq_midi_event 7680 1 snd_seq_oss snd_seq 54928 5 snd_seq_oss,snd_seq_midi_event snd_pcm_oss 57512 0 snd_mixer_oss 18944 2 snd_pcm_oss snd_cmipci 34232 5 snd_pcm 97032 2 snd_pcm_oss,snd_cmipci snd_page_alloc 11528 1 snd_pcm snd_opl3_lib 11008 1 snd_cmipci snd_timer 25860 3 snd_seq,snd_pcm,snd_opl3_lib snd_hwdep 9860 1 snd_opl3_lib gameport 4736 1 snd_cmipci snd_mpu401_uart 8064 1 snd_cmipci snd_rawmidi 25508 1 snd_mpu401_uart snd_seq_device 8456 4 snd_seq_oss,snd_seq,snd_opl3_lib,snd_rawmidi snd 61444 22 snd_seq_oss,snd_seq_midi_event,snd_seq,snd_pcm_oss,snd_mixer_oss,snd_cmipci,snd_pcm,snd_opl3_lib,snd_timer,snd_hwdep,snd_mpu401_uart,snd_rawmidi,snd_seq_device soundcore 9056 2 snd hid 40896 0 ohci_hcd 19332 0 sis_agp 7424 1 agpgart 30888 1 sis_agp vmnet 24720 12 parport_pc 35520 0 parport 37832 1 parport_pc vmmon 41656 0 ipt_TCPMSS 4480 1 ipt_TOS 2688 9 ipt_state 2176 88 ipt_LOG 5760 74 usbcore 103516 4 hid,ohci_hcd evdev 9856 0 ip6t_LOG 5632 2 ip6t_REJECT 7296 3 ip6table_mangle 2816 0 ipt_REJECT 6656 3 iptable_mangle 3072 1 iptable_filter 2944 1 ip_nat_ftp 4848 0 iptable_nat 22060 1 ip_nat_ftp ip_conntrack_ftp 72112 1 ip_nat_ftp ip_conntrack 31280 4 ipt_state,ip_nat_ftp,iptable_nat,ip_conntrack_ftp ip_tables 17280 8 ipt_TCPMSS,ipt_TOS,ipt_state,ipt_LOG,ipt_REJECT,iptable_mangle,iptable_filter,iptable_nat ip6table_filter 2816 1 ip6_tables 18704 4 ip6t_LOG,ip6t_REJECT,ip6table_mangle,ip6table_filter ipv6 236800 18 ip6t_REJECT binfmt_misc 10120 1 subfs 7424 3 nls_utf8 2304 1 nls_cp437 6016 1 vfat 14208 1 fat 43584 1 vfat dm_mod 50300 0 e1000 82436 0 reiserfs 241104 3 What a lot of stuff, how do I cut this down a bit? eg: I could live without ipv6 but can't find it in modprobe.conf michaelj -- Michael James michael.james@csiro.au System Administrator voice: 02 6246 5040 CSIRO Bioinformatics Facility fax: 02 6246 5166

cincaipatron@gmx.net wrote:

...

On Tue, 29 Jun 2004 03:40 pm, Fred Miller wrote: Good point, could the people who are getting good results post some more details?

Processor, RAM, chipset, and output of lsmod

Want to add one more thing: the IO scheduler. I remember vaguely that on low-end machines (< P4), the CFQ IO scheduler causes KDE applications to startup slower then the default (anticipatory) IO scheduler.

There was one earlier kernel 2.6.x-mm? from kernel.org that was very slow, can't remember which, there was frantic disk thrashing and top showed very high I/O wait. I reported it to the kernel mailing list and Andrew Morton had it fixed in the next update. See if there is anything like excessive memory or CPU usage in top, also I/O wait. I/O Wait doesn't show up now. Tasks: 182 total, 1 running, 179 sleeping, 0 stopped, 2 zombie Cpu(s): 9.1% us, 4.5% sy, 0.0% ni, 85.8% id, 0.0% wa, 0.6% hi, 0.0% si Mem: 514692k total, 510424k used, 4268k free, 53744k buffers Swap: 5373724k total, 268772k used, 5104952k free, 155232k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 14543 root 15 0 183m 15m 138m S 7.5 3.2 189:44.29 X 14714 lancelot 15 0 31504 6168 25m S 1.9 1.2 0:20.47 kdeinit 15201 root 15 0 51476 6380 37m S 1.6 1.2 30:49.66 epiphany-bin 4732 root 17 0 1904 1004 1564 R 1.3 0.2 0:00.57 top 14716 lancelot 16 0 59636 8816 39m S 0.6 1.7 10:02.60 kdeinit 14705 lancelot 15 0 26760 2784 22m S 0.3 0.5 4:37.08 kdeinit 14815 lancelot 16 0 32324 2328 23m S 0.3 0.5 0:20.31 suseplugger 14837 lancelot 15 0 25524 10m 17m S 0.3 2.1 15:18.68 skype 14982 root 15 0 70872 39m 32m S 0.3 7.9 6:23.53 thunderbird-bin 15216 root 15 0 51476 6380 37m S 0.3 1.2 1:01.71 epiphany-bin 1 root 16 0 588 72 444 S 0.0 0.0 0:04.25 init 2 root 34 19 0 0 0 S 0.0 0.0 0:00.35 ksoftirqd/0 3 root 5 -10 0 0 0 S 0.0 0.0 0:04.38 events/0 4 root 5 -10 0 0 0 S 0.0 0.0 0:00.06 khelper 5 root 5 -10 0 0 0 S 0.0 0.0 0:00.00 kacpid 45 root 5 -10 0 0 0 S 0.0 0.0 1:19.82 kblockd/0 <SNIP> Regards Sid. -- Sid Boyce .... Hamradio G3VBV and keen Flyer ===== LINUX ONLY USED HERE =====

Bruce Marshall wrote:

On Tuesday 29 June 2004 01:40 am, Fred Miller wrote:

...

On Tuesday June 29 2004 1:05 am, Steve wrote:

...

Well, that's certainly not true for all hardware. I've got a 300MHz PII laptop that ran fairly well with 8.2. Its performance under 9.0 was, to say the least, abysmal. I was beginning to think it had finally reached the end of its usefullness as my desktop for work, unless I wanted to stick with 8.2. Then I installed 9.1 on it and, boy howdee, it runs great again. In fact, it feels faster than 8.2.

Same here with my P-II 333 64M laptop, admittedly I could do with more memory as it's a tad slow in waking up from sleep.

...

Some have great success and some don't....it's one of the reasons I think it has something to do with the 2.6 kernel and some or all of the modules.

If so, then it must be something that SuSE did to it. I've been running the 2.6.x kernels on 9.0 and now 9.1 (vanilla from kernel.org) and have had no problems whatever.

I would like to see the output of top as that would show what's soaking up resources, something definitely is and killing it if possible speeds things up. Regards Sid. -- Sid Boyce .... Hamradio G3VBV and keen Flyer ===== LINUX ONLY USED HERE =====

Goto the archives and look at the thread IPv6. CWSIV On Fri, 2004-06-25 at 08:12, Philip Amadeo Saeli wrote:

* linuxjim [2004.06.24 23:35]:

...

Greetings all,

I upgraded one of my 8.2 boxes to 9.1, and expected to find 9.1 a bit faster than the older version. So far that ain't happening. I did the upgrade, not fresh install. I had to clear out some older programs to eliminate dependencies, but other than that all seem to go fine. ...

I have not yet loaded 9.1 on my system (a 750 MHz PIII with 544 MB RAM). However, I found KDE on 9.0 to be significantly slower that on 8.2. I have a multi-boot system and have both 8.2 and 9.0 loaded, so can quickly compare them and, e.g., 8.2 is 3-4 times faster in starting up konsoles (and even opening up sub-shells in konsoles!) than 9.0.

Anyone have any comparisons between 9.0 and 9.1? Was the slowness introduced in 9.0 and then carried over into 9.1, or is 9.1 faster or slower than 9.0? 8.2 to 9.0 definitely brought a huge hit.

BTW, I am using the exact same KDE setup for both 8.2 and 9.0. Also, I did full installs of each version. Additionally, I've kept 9.0 updated via YOU to the latest security & recommended updates.

Phil

-- Philip Amadeo Saeli SuSE Linux 9.0 psaeli@zorodyne.com

I tried reducing the resolution and color bits, and shut down suseplugger. No increase in speed. This thing is UNBEARABLY slow. Yast takes over 2 minutes to start up. Even logoff takes about a minute. I wonder what kind of machine it takes to make 9.1 rock? I guess I'll try a fresh install to see what that does. Its that or go back to 8.2 (I missed buying 9.0 as I had purchased 8.2 not long before 9.0 release. I was not familiar with the suse product cycle times then). Either way its a fresh install, so I'll try 9.1 again but fresh this time. Jim F On Friday 25 June 2004 10:34 pm, Fred Miller wrote:

On Thursday June 24 2004 7:34 pm, linuxjim wrote:

...

Greetings all,

I upgraded one of my 8.2 boxes to 9.1, and expected to find 9.1 a bit faster than the older version. So far that ain't happening. I did the upgrade, not fresh install. I had to clear out some older programs to eliminate dependencies, but other than that all seem to go fine.

[snip]

Correct. ALL of the installs I've done, 9.1 IS slower than 9.0, which is what I've upgraded all of them from. I don't pretend to know why, but some work needs to be done "somewhere" in the area of performance.

Fred

-- "Ballmer is no more designed for the art of persuasion than the Abrams tank is for delivering meals on wheels."

linuxjim writes:

I tried reducing the resolution and color bits, and shut down suseplugger. No increase in speed. This thing is UNBEARABLY slow.

1. KDE 3.2 performance on your machine will never be excellent. It should work but if productivity is an issue then use a "lightweight" window manager. 2. DMA seems to be turned off. If your disk is accessed via /dev/hda then check the value of using_dma: # hdparm -v /dev/hda /dev/hda: multcount = 16 (on) IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 1 (on) keepsettings = 0 (off) readonly = 0 (off) readahead = 256 (on) geometry = 65535/16/63, sectors = 120103200, start = 0 Performance differences between 8.2 and 9.1 are not very big (on a desktop), just KDE is a bit slower due to "visual enhancements". Huge differences that people report are often due to DMA disk access and similar options ... which are turned on by default in one distribution but not in the other. The 2.6 kernel is supposed to perform better on machines with many CPUs and where many threads are involved. On a desktop, the response should be more "balanced" but don't expect any significant speedup. -- A.M.

On Saturday June 26 2004 11:46 am, linuxjim wrote:

I tried reducing the resolution and color bits, and shut down suseplugger. No increase in speed. This thing is UNBEARABLY slow. Yast takes over 2 minutes to start up. Even logoff takes about a minute. I wonder what kind of machine it takes to make 9.1 rock?

I guess I'll try a fresh install to see what that does. Its that or go back to 8.2 (I missed buying 9.0 as I had purchased 8.2 not long before 9.0 release. I was not familiar with the suse product cycle times then). Either way its a fresh install, so I'll try 9.1 again but fresh this time.

I've done both and there' NO difference that I've found. Fred -- "Ballmer is no more designed for the art of persuasion than the Abrams tank is for delivering meals on wheels."