CentOS 5 Hardware :: CPU Throttling - Frequency Variation
Jan 12, 2011I have de Dell desktop (optiliex 755) and i have instal CentOS 5.5 on it. Some informations about the system are:
View 2 RepliesI have de Dell desktop (optiliex 755) and i have instal CentOS 5.5 on it. Some informations about the system are:
View 2 RepliesIs there a way of throttling a process resources, something akin to limits but for processes not users?ie I want processX to be restricted in the amount of memory it can consume. For process cpu I guess I can simply nice the process, but total memory consumption is my primary concern.
View 1 Replies View RelatedSystem recognizes that CPU is capable of throttling, but there is no module to govern it. So Xeon always runs at full frequency.This is not convenient as my server is not always in full use, so I would like it to lower power consumption, especially for AC, with frequency scaling enabled.Can you help me is this some bug, misconfiguration in default kernel or guys who writes kernel think that scaling is not necessary for servers processors?
From my analysis so far, I think that the issue lies with the fact that module acpi-cpufreq does not support Xeons and that there is no dedicated module for this processor family (e.g. p4-clockmod) module in the default kernel.
would ask how to fix this warning that comes when i start my centos 5 virtual machine that run under windows 2003 std, previously it run smoothly with linux centos 5.
currently
Centos 5 Virtual Machine
____________________
VMware workstation 6 for windows
____________________________
Windows 2003 STD
previously
Centos 5 Virtual Machine
______________________
VMware workstation 6 for linux
___________________________
Linux Centos 5
I de-crappified a new laptop last week for a friend and set him up with OpenOffice.I had earlier salvaged files from his old PC HD, saved them to my Box. Most of his office files were actually the dreaded M$ Works (.wps)But I tried them and OpenOffice had no trouble, least in Linux that is.I burned his files to a DVD-RW and copied them to his new Laptop. OpenOffice for windows wouldn't have anything to do with the .wps filesStrange or what !
View 6 Replies View RelatedI needed to find the time it takes to execute a for loop... I decided to use the gettimeofday() function and wrote the following code:
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <sched.h> // for use of sched_setaffinity
[code]....
so i know that i can create a live usb stick, but i dont want to do that. i want a full install of the system on the stick. at the moment i have a 2GB usb stick, and when i try to install kubuntu, i get an not enough space error. thats understandable. but if there is an ubuntu variation that i can install on less than 2GB of space (and optionally run the KDE environment), then please provide a link. if there isnt an ubuntu variation, then please provide something other then Damn Small Linux.
View 6 Replies View RelatedI decided to install Fedora today ( from Debian ) so I downloaded the 64 bit variation and then burnt it and booted from it. My partition table before looked like this ( the XXXs of values I do not know ) :
Code:
Number Start End Size Type File system Flags
1 1049kB 10.5GB 10.5GB primary fat32 diag
2 10.5GB 110GB 100GB primary ntfs
3 110GB 111GB 98.7MB primary ext2 boot
4 111GB 320GB 209GB extended
5 111GB 315GB XXXXXX logical ext3 XXXX
6 315GB 320GB 5124MB logical linux-swap(v1)
I wanted to seperate my /home and / (root) partitions in my new system so that in the future, distro-hopping would be easier. I deleted everything in my fifth partition except for my home folder, from which I moved my home folder ( /home/whales ) to the top layer of the partition ( now /whales ).
I then did a yum install gparted so I could shrink partition 5 (now going to be /home) and add another 20gb one for / (root). Unfortunately Gparted spat out an error about the kernel not reading the partition table until reboot ( but first it succeeded in shrinking the FS to 286GB ). I panicked and tried to format p5 to ext3 ( it was ext3 ) but then realised this was dumb before it could 'create the ext3 filesystem' and I cancelled the operation.
I then rebooted, realising I should have done what the error said to in the first place.Nope. The partition was 'empty' apart from lost+found ( which is empty ). Nuts I thought to myself. The filesystem must have been overwritten . I tried a 'rescue' on parted on the command line, but that did not work. So I then removed the parition reference and then tried a rescue, but nothing happened. No output other than progress, and now I am stuck with this layout:
[Code]...
I know you can use grep to find keywords in filenames, but is there a variation that can enable you to search the content of the files themselves?
View 9 Replies View RelatedAfter being forced to rebuild my computer (failed drive) I finally got current and installed F11 (was using F9). I was watching GKrellm last night and noticed something odd. The CPU frequency on my processor (Phenom 9600) was scaled to the slower speed and was rarely being pushed to full speed.
I verified that cpuspeed was in ondemand mode and ran some processor intense tasks (huge image loads, large file cats, large yum installs, etc...). Only when I was able to sustain near 100% load would the system throttle to the full clock speed. It would then drop back to the slower speed very quickly even though the process was still running. As a rough benchmark, I ran these tasks again in performance mode to see what impact this was having. Most of these tasks were taking as much as 40-50% longer to complete in ondemand mode.
Digging further, I found that the default up_threshold in F11 is set to 95%!! This is verified by cating /sys/devices/system/cpu/cpu0/cpufreq/ondemand/up_threshold. This means that the system would not throttle to full speed unless a 95% load was sustained for multiple samples (36 miliseconds each by default).
I overrode these settings in /etc/sysconfig/cpuspeed. I change up_threshold to 60 and down_threshold to 30. I am at work right now, so I can't benchmark this change until tonight. I guess I could have just set it to performance mode and left it, but I'd rather save the power when the machine is idle.
Does anyone have any thoughts on why 95% is the default and if there is any problem/benefit to changing this to a much lower value? It seems that anyone with a CPU running SpeedStep or CoolnQuiet would suffer the same severe performance impacts I saw with the default values.
I want to install ATLAS on my laptop to do this i have to turn of cpu throttling.
View 4 Replies View RelatedI'm trying to get speedstep to work on a intel ss4200 NAS box with a core 2 duo e4600 cpu.
I'm running Fedora 13; Kernel 2.6.34.7-56.SS4200.fc13.x86_64, which is custom with a couple of patches to properly enumerate the hard drive controler.
For some reason the cpu frequency driver it is picking is "p4-clockmod", how do i force the kernel to load the correct driver?
Code:
I'm trying to install ATLAS which requires disabling cpu throttling. Normally one can do it in the BIOS, while there isn't such an option of my dell inspiron 6400. Actually there is a SpeedStep option in the BIOS, however, you cannot get the highest but the lowest performance by disabling it (dell!!!).
After googled a lot, I found in some distros, there is a /usr/bin/cpufreq-selector through which one can disable throttling. But it doesn't exists in slackware.
I known that by appending kernel option "acpi=off", one can disable the whole acpi and thus the throttling control in slackware, but it seems so dirty.
Anyone known a better way to do it?
ps: With a intel T2050 cpu, i didn't find the directory /sys/devices/system/cpu1, but only the /sys/devices/system/cpu0. Similar case in /etc/acpi. it seems that slackware treats my cpu as a single core one, while it is not.
I am trying to limit bandwidth of certain ip addresses on my server. I have been doing hours of reading and not getting very far...
So far I believe the iptables command is:
And now I just need the tc command to read those marks and limit bandwidth, I have a gigabit connection and would like to limit each of these ip addresses to 10mbit in and out.
If I read /proc/acpi/processor/CPU0/throttling my processor is at T2 and I can not move it down to T0.
View 10 Replies View RelatedThere seems to be many different ways of controlling bandwidth usage of downloads of content from Apache2. Does anyone know which is the standard module deployed/deployable in OpenSuse?
View 2 Replies View RelatedHow to do bandwidth throttling in Squid server? i wan to set particular bandwidth for particular ip, how to do it in Squid server?
View 1 Replies View RelatedI'm going to set up a new linux router for a company, and have to set up bandwidth throttling. They have an unlimited ADSL internet connection which will be shared between 2 businesses, one being them. I will need to set it up so their connection will never be slowed down by the other business. They will both be connected to the same NIC, but will be on different subnets. How would I go about doing this?
View 2 Replies View RelatedEvery time I use Transmission for torrent download it throttles my Internet connection although I set speed limits way below my bandwidth.
View 2 Replies View RelatedI'm not getting file transfers that utilize the max bandwidth of my ADSL broadband. I have a 512 kbps line (I know that's not broadband in most places, but such is the option available to me), and I usually get download speeds of 55-60 kbps. Since the past couple of days, this has dropped to an abysmal 2-3 kbps. So I used [url] to test my download speeds, and they were as expected (i.e. the usual 55-60 kbps). I also used the speed test page of my ISP, and that too, gave the usual results. There has been no throttling from the ISP, as I confirmed from their help desk.
Also, web pages open fine. It's only when I'm downloading a file (Yeah, web pages are also files, but what I mean is compressed archived files; .rar, .tar.?g* et.al.) that I can't get the desired speed. I haven't changed the resolv.conf file, not have I made any other changes that might cause this. I use ppp to dial-up to my ISP, and use the pon and poff scripts that ppp provides. I have a peers file configured for my ISP, which again, I haven't edited.
I am uploading the incremental backups using duply/duplicity using the sftp-module. As the initial upload is pretty big and runs several days (more than 50GB over a 1Mbps-line) I am confronted with the problem that other users in the network experience slowdowns when I upload.
I would like to run a script every n minutes which pings a host in the internet (second hop of the traceroute for example). If the response time is less than a value (150ms), the script throttles the upload for one specific host and protocol. Traffic to the local net (Samba mainly) should be unaffected. I cannot use the QoS of the firewall/router. Also I would like the penalty to be removed if the ping is quicker (loess than 70ms for example) I looked at trickle, and some other out-of-the-box shaping tools but they do not give me the possibility to change the rate while the upload is running.
I would now write a script in perl which uses [URL]some wrapper for iptables combined with some ping module [URL] Also I was trying to get the proof of concept before I start coding: (I haven't verified if this works yet)
sudo tc qdisc add dev eth0 root handle 11: cbq bandwidth 100Mbit avpkt 1000 mpu 64
sudo tc class add dev eth0 parent 11:0 classid 11:1 cbq rate 100kbit allot 1514 prio 1 avpkt 1000 bounded
sudo tc filter add dev eth0 parent 11:0 protocol ip prio 16 u32 match ip dst MyserverIP flowid 1:1
I'm thinking about some ways to limit access to my web-server. It runs Nginx and php in FCGI. The server contains a large amount of information. The data is freely available and no authentication is required but other companies might like to mirror it and use on their own servers.
The requests could be limited on different levels: IP, TCP, HTTP (by nginx) or by the php application. I found some solutions (like Nginx's limit_req_zone directive), but they do not solve the second part of the problem: there's no way to define a whitelist of clients who are allowed to use the data.
I thought about an intellectual firewall that would limit the requests on IP basis, but I'm yet to find such device. Another way was to hack some scripts that would parse the log file every minute and modify the iptables to ban suspicious IPs. It would take days and I doubt this system will survive, say, 1000 requests per second.
Perhaps, some HTTP proxy, like Squid, could do this?
I just update the system, and the cpu frequency applet stop working. I can no longer lock the cpu frequency. Once I click the cpu applet icon, The whole applet just freeze right there. Does not matter if what I do with it. Well, at least I hope some one just offer me a bit help on how to lock the cpu frequency in terminal. Other wise I can not run Folding@home due to cooling issue.
Post added at 05:08 AM CST Previous post was Yesterday at 11:58 PM CST Well, it seems like the program take of authentication is mess up. I have the same problem when I tried to config my firewall. In the Firewall case, once I type in my password, the whole window just went blank. For the cpu applet case, nothing happen, the window ask for password does not show up. nothing. The system log did not have any entry...neither does the crash log. Which there are lot of entry there since I install F12
I am not entirely convinced that my CPU is actually changing frequency as it is meant to. It sometimes changes frequency, but most of the time it is stuck on 800MHz even when doing cpu intensive tasks. Here is information that may or may not be of help:
cat /proc/cpuinfo
processor: 0
vendor_id: GenuineIntel
[code]....
I notice here that it says between 800 and 800, but does that mean it is currently at 800, or can only be 800.
how to lock the CPU frequency under Fedora 15. I try to run cpufreq comand but the terminal tells me the comand is bad and ask me if I want to install the program. When I click yes. It says the program has already been installed. While I try to use software manager under GUI, same thing happend. Software manager clearly tells me I have not install that program. But once I choose to install, the system tells me all package has been install?
View 2 Replies View RelatedMy CPU has 2.19GHz (2190.000kHz)
but linux runs to fast if i watch the time 1 minute (real time) = 5 minutes (linux)
i checked the cpu freq with
Code:
cat /proc/cpuinfo |grep MHz
and the output was
cpufreq 2194.800 kHz
the cpu is overclocked!!!!
[Code]....
Can anyone point me in the direction of how to set the screen to 60 as I find I get sore eyes. Sorry I don't know the tech jargon.
View 2 Replies View RelatedI have a n2pap-lite motherboard with a AMD Sempron 2800+. My host clock is at 133mhz when it needs to at least be at 166mhz to be a AMD Sempron 2800+... I installed the system on 100mhz clock not thinking that it mattered.
I tried the following
sudo apt-get install rcconf
sudo rcconf
*Disabled "ondemand"
It says on guides that it will ask me to reboot but it doesn't, so I manually did it and put the CPU Frequency Monitor on my taskbar. It does not have the options like it shows in the screenshots to set my CPU frequency..
I'm looking for a CPU frequency applet which monitors the frequency of ALL CPUs at once.
Does it exist such applet?
I wanted to be able to see the frequency my cpu is running at
Quote:
Originally Posted by xfce4-cpufreq-plugin
CPUFreq support is not present. The CPUFreq Monitor will not function this session.
Code:
~$ cpufreq-info
cpufrequtils 006: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to cpufreq@vger.kernel.org, please.
analyzing CPU 0:
[Code]....