%CPU column in autotrace - 10gR2

Hi all,
Can anyone shed some light about the %CPU column in the output of an autotrace?
For one of my application queries, I tuned it by adding a few indexes - no big deal.
The cost went down to 95 from 492.
Here are the top lines from each autotrace:
Before indexes:
Id Operation Name Rows Bytes Cost (%CPU) Time
0 SELECT STATEMENT 215 78045 492 (2) 00:00:06
After indexes:
Id Operation Name Rows Bytes Cost (%CPU) Time
0 SELECT STATEMENT 215 78045 95 (4) 00:00:02
What I'm wondering about is the signficance about the %CPU column - although cost went down, it appears that %CPU went up from 2 to 4.
What exactly does that column mean?
Thanks in advance!

Hi Centinul,
thanks for the link! After doing some tests i agree with your idea that TIME in autotrace trace is only a 'guess' like it is using explain. I have been looking at v$sql_plan_statistics_all, there does exist a column named elapsed_time, but it's only filled when you change STATISTICS_LEVEL from TYPICAL to ALL. But the difference in between these two settings is unbelievable. When switching to ALL the elapsed time is growing from 0.5 secs to 30 secs. But it is at least to be seen in v$sql_plan_statistics ;-) without changing the TIME reported in the result of autotrace report.
Thanks Achim

Similar Messages

"find" process User =nobody [?], using up to 90% of CPU = G5Quad [??? !]

This one is for the PPC G5 Quad pros:
Had been working for hours on my Mac Pro (different machine, I know) with my networked G5 Quad supposedly idling in the background.
The only open applications on the G5 Quad are Adobe Photoshop CS4 and Adobe Bridge CS4, the main reason for my keeping the G5 at this point, because of vital plug-ins that are not Intel-compatible and are absolutely irreplaceable and mandatory. (I do have Ps CS5 and CS6 on the Mac Pro. Both aforementioned applications had no open documents. I have not actually used then in about six weeks, so the chance of Bridge building or rebuilding caches id very low.
Alerted by a sudden, very substantial increase of fan and liquid pump activity in my essentially idle G5 Quad, I switched the shared monitors inputs to it and was not surprised by the temperatures, almost all showing no more than 49ºC/121º.
Launched the Activity Monitor, which was showing a Process named "find" owned by a User named "nobody" using up to 90º% of the CPU column.
Terminated said process, only to see it become active again a short time (few minutes) later.
While typing this post, with interruptions, (20 minutes or so) the "nobody" user disappeared, as did the "find" process. some 15 minutes later, the fan and pump activity ended, returning to the G5's normal vigorous hum.
It has been running normally for about 30 minutes now, with very quiet sound levels.
Obviously, I'm mystified because all of this is above my pay grade.
I have a bunch of questions:
1.— Why on Earth is/was there a user named "nobody" working my machine so hard? What or who can it be?
2.— What is a "find" process, especially on a supposedly idle machine?
2.— Any chance my computer has been hacked and overtaken by a "third party"?
3.— Should I be worried? I am.
4.— What steps would you recommend I take at this point?
Thanks in advance to all. I look forward to hearing from old Apple forum friends. 
2.5 GHz Power Mac (PPC) G5-Quad; 16GB RAM; mutant, flashed 550MHz nVidia GeForce 7800GTX 1,700MHz 512MB VRAM; ATTO ExpressPCI UL5D LP SCSI card; Mac OS X Tiger 10.4.11 and Leopard 10.5.8 boot drives; intrusive Spotblight, Dashboard and Time Machine permanently disabled; dual 22"CRT/22"LCD monitors; USB wireless 'N' available but connected to the Internet via wired Ethernet; 2 SCSI scanners (one tabloid-size transparency scanner and a film scanner); various internal & external HDs; FW Epson 2200 and Ethernet Samsung ML-2850ND printers; 2 X Back-UPS RS 1500 XS units.
Intel Xeon Mac_Pro1,1 [late 2006] 2.66 GHz Quad; 16 GB RAM; mutant/flashed, factory-overclocked 875/1225 "Apple PowerColor" ATI Radeon HD 5770 with 1 GB of GDDR5VRAM driving side-by-side dual 22" CRT/24"LCD monitors in Extended-Desktop mode; Mac OS X Lion 10.7.5 and Snow Leopard 10.6.8 boot drives; intrusive Spotlight and frivolous Dashboard permanently disabled; USB wireless 'n' available but connected to the Internet and networked via wired Ethernet.
These two computers share keyboard and mouse through a simple, manually activated USB switchbox.

Gee… I should have done my due diligence frst.
Niel had already answered the very same question here:
https://discussions.apple.com/message/21746515#21746515
My bad.
Thank you, Niel. Please say at least "Hi !" here so I can give you your well deserve points here as I cannot do it in the other thread.

Finder and loginwindow use 100% cpu with 10.6.8

I recently installed the patch to OS X 10.6.8 and now the finder is very unstable and uses 100% cpu and the loginwindow uses 100% cpu also. Yes I have Parallels 6 on this machine but I already "un-checked" the "show windows applications in doc" option which is the recommended solution. However, this did not fix the problem. I do not know how to determine what actual process is causing this finder / loginwindow issue. I very much appreciate if anyone has suggestion how to fix this problem.
This is on a mid-2011 MacBookPro.
thank you.

TBauer, I had the same problem with 10.6.8 using 100% of the CPU with Parallels 6 installed. The following steps solved this problem for me:
0. Take a screen shot of your Dock if you wish to remember it as currently set
1. Launch Activity Monitor and click on the CPU column so you can see Dock using the CPU
2. Move ///Users/ ... /Library/Preferences/com.apple.dock.plist to the trash
3. Click on Dock and click on Quit Process to stop the Dock
OS X will then rebuild the default Dock
4. Put the items back in the Dock that you want
That solved the problem for me yesterday morning. I've launched Parallels 6 several times now and moved the machine from my work network to home and the 100% CPU problem hasn't returned.
Hope that works for you, too!
I don't think using the combo updater will solve the problem. The problem appears to be a .plist that is generated with OS X 10.6.7 and earlier that contains icons of a larger than Apple documented size causes the CPU looping. Rebuilding the dock.plist under OS X 10.6.8 appears to solve the problem.
Parallels has issued a knowledge bulletin saying they are looking into the problem. On their web site they have another approach involving moving the offending icon out of the dock to solve the problem. I like my solution better but theirs works, too.

Excluding one or some columns in output of Get-process cmdlet doesn't work

hello
in PS 4.0 i need a command so that the output includes all columns except for, one or two columns ?
for example get-process cmdlet shows seven columns. i need a cmdlet to show all columns except for handles & cpu columns.
( instead of running Get-process | select -object NPM,PM,WS,VM,CPU,ID )
note: i tested the following command but it still shows handles & cpu columns in the output
( Get-process | select -excludeproperty handles,cpu )
thanks in advanced

Bill should have also pointed out that you can change the "Types" files to adjust output permanently. I am sure that this would save you much typing.
PowerShell is very flexible. We can select properties with wild cards...
get-process|select s*,c* -first 3|ft -auto
You can adjust this as needed.
¯\_(ツ)_/¯
hi jrv
i didn't understand you first post at all:
"have a better and more permanent solution.
Print out put to very large wide E or F size plotter in landscape mode....
Take scissors and cut out columns not needed. Tape bits together. NOw you have the results you like."
but i undrestood your 2nd post, that nice.
thank you very much,
i also found this workaround:
Get-Process | Select-Object -Property * -ExcludeProperty cpu,handles | format-table -Autosize
regards

Why is Firefox taking up a lot of CPU and running very slowly?

For the past week my Firefox (and also Chrome) have been running very slowly. Any time a page loads the CPU column in Task Manager shoots up to between 10-20 when it's usually at 00, and it sends my cpu fans into overdrive until the page is loading.
Pictures load very slowly, and busy pages are even worse, such as Weather channel pages or Gawker blog articles. I've run MSE, Malwarebytes, I've run Firefox in safe mode, nothing is working. I'm really beginning to get frustrated because it's making any kind of browsing at all very unenjoyable.

Safe Mode didn't make any difference?
Run all Windows Updates, install all needed service packs, etc.
Update your graphics driver (Firefox uses your graphics card for some rendering, and an out of date graphics card driver can cause problems. [https://support.mozilla.org/en-US/kb/how-do-i-upgrade-my-graphics-drivers https://support.mozilla.org/en-US/kb/how-do-i-upgrade-my-graphics-drivers].
Update all your plugins (Flash, Java, etc.): [http://www.mozilla.org/plugincheck/ http://www.mozilla.org/plugincheck/].

Why is the Dock application consuming CPU?

I got a new MacBook Pro 15" retina display, migrated my data from my old mac, and now the Dock application is consuming a ton of CPU on the system.
I have found many other threads in these forums, many of them quite old. They typically suggest a problem with VMWare fusion or Parallels, neither of which I have on my system. They also suggest removing the com.apple.dock.plist file and killing the application, which did not help, nor did removing all applications from the dock and quitting all other applications.
What else can I do to try and fix this problem?

Have you tried restarting the computer? Also,
Open Activity Monitor in the Utilities folder. Select All Processes from the Processes dropdown menu. Click twice on the CPU% column header to display in descending order. If you find a process using a large amount of CPU time, then select the process and click on the Quit icon in the toolbar. Click on the Force Quit button to kill the process. Be sure to note the name of the runaway process so you can track down the cause of the problem.
You can also kill the Dock process using Activity Monitor.

Check cpu usage

I noticed that while idle my CPU usage spikes to 20% every few seconds. However, activity monitor shows no process using a lot of CPU during those spikes (I have activity monitor set to show all processes). As there any way in Unix that I can track CPU usage over say 10s and find out which process causes this spike?

Hmm, I wonder if this is not related to a specific process, but rather something like a NIC thread or something similar which doesn't show up obviously in top output.
You could try the "ps" technique, though that might suffer from the same limitation that top does. No harm in trying though. From the command line, do something like:
user% for ((i=0; i<30; i++)); do
for> ps alx >> ps.out
for> echo "\n=====\n" >> ps.out
for> sleep 1
for> done
Then look through your results in the "ps.out" file at the CPU column. Maybe you'll see something in there. That column is the %cpu -- from the ps man page:
   %cpu   The CPU utilization of the process; this is a decaying average over up to a minute
     of previous (real) time. Because the time base over which this is computed varies
     (some processes may be very young), it is possible for the sum of all %cpu fields to
     exceed 100%.

%CPU Value on plan_table_output

Does anyone know how the %CPU value is calculated when using dbms_xplan.display in oracle 10g?
I'm sure it comes from plan_table (v$sql_plan) but I can't figure out how the value is calculated...
Thanks

user8873443 wrote:
Does anyone know how the %CPU value is calculated when using dbms_xplan.display in oracle 10g?
I'm sure it comes from plan_table (v$sql_plan) but I can't figure out how the value is calculated...
ThanksFrom the documentation:
http://download.oracle.com/docs/cd/B28359_01/server.111/b28274/ex_plan.htm
"COST: Cost of the operation as estimated by the optimizer's query approach. Cost is not determined for table access operations. The value of this column does not have any particular unit of measurement; it is merely a weighted value used to compare costs of execution plans. The value of this column is a function of the CPU_COST and IO_COST columns.
CPU_COST: CPU cost of the operation as estimated by the query optimizer's approach. The value of this column is proportional to the number of machine cycles required for the operation. For statements that use the rule-based approach, this column is null.
IO_COST: I/O cost of the operation as estimated by the query optimizer's approach. The value of this column is proportional to the number of data blocks read by the operation. For statements that use the rule-based approach, this column is null."
You could try to perform a test like the following:
1. Create a dummy table named T1, in my case I am using a test table with 100,000 rows.
2. Experiment:
EXPLAIN PLAN FOR SELECT COUNT(*) FROM T1;
Explained.
SELECT
ID,
CPU_COST,
IO_COST,
COST
FROM
PLAN_TABLE;
ID   CPU_COST    IO_COST       COST
0   13331709         11         13
1
2   13331709         11         13
SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY);
Plan hash value: 2552807428
| Id | Operation             | Name         | Rows | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT      |              |     1 |    13 (16)| 00:00:01 |
|   1 | SORT AGGREGATE       |              |     1 |            |          |
|   2 |   INDEX FAST FULL SCAN| SYS_C0020571 |   100K|    13 (16)| 00:00:01 |
------------------------------------------------------------------------------From the above, the total cost of ID 0 is 13, of which 11 is the IO cost. Therefore, the CPU cost is 13-11 = 2. 2/13 = 15.38%. 15.38% rounded up is 16%. That might be the meaning of the %CPU column in the DBMS_XPLAN output.
Charles Hooper
Co-author of "Expert Oracle Practices: Oracle Database Administration from the Oak Table"
http://hoopercharles.wordpress.com/
IT Manager/Oracle DBA
K&M Machine-Fabricating, Inc.

Coreservicesd 100% cpu on 10.6 and quicksilver

I've been suffering from teh 100% coreservicesd race condition for a while. I've searched around for a long time before finding a discussion thread that hinted at the issue I'm having - quicksilver.
The problem was with the 'Dock' being in one of the 'Catalogue' items.
1. right click on quicksilver menu icon
2. select 'Catalog'
3. select 'User'
4. uncheck 'Dock'
As soon as I did this, coreservicesd cpu usage immediately dropped.

TBauer, I had the same problem with 10.6.8 using 100% of the CPU with Parallels 6 installed. The following steps solved this problem for me:
0. Take a screen shot of your Dock if you wish to remember it as currently set
1. Launch Activity Monitor and click on the CPU column so you can see Dock using the CPU
2. Move ///Users/ ... /Library/Preferences/com.apple.dock.plist to the trash
3. Click on Dock and click on Quit Process to stop the Dock
OS X will then rebuild the default Dock
4. Put the items back in the Dock that you want
That solved the problem for me yesterday morning. I've launched Parallels 6 several times now and moved the machine from my work network to home and the 100% CPU problem hasn't returned.
Hope that works for you, too!
I don't think using the combo updater will solve the problem. The problem appears to be a .plist that is generated with OS X 10.6.7 and earlier that contains icons of a larger than Apple documented size causes the CPU looping. Rebuilding the dock.plist under OS X 10.6.8 appears to solve the problem.
Parallels has issued a knowledge bulletin saying they are looking into the problem. On their web site they have another approach involving moving the offending icon out of the dock to solve the problem. I like my solution better but theirs works, too.

Not enough CPU power

Lately, my Mac has randomly been crashing. So, the first solution I tried so far, is to replace the RAM. So, I had upgraded to 16GB instead of the default 4GB, but if I still run an application (mainly videos, or a game), the CPU being consumed by the user will go up to 50% I'm not exactly sure if this normal. I had opened all the applications in the Adobe Master Collection CS5, and the CPU being consumed by the user went up to around 40%-55% and I thought this was fair, but watching a 420p video on YouTube (I'm using Google Chrome) goes ridiculously high just for a 420 pixel video, it takes up 40% to 60%
If this is normal, even with 16GB of RAM installed, how can I increase my CPU power? Also, if I keep my CPU usage under 90% then my Mac doesn't crash.

The only way to get a more powerful CPU is to sell your computer and buy a faster one. CPU power is not related to how much RAM is installed.
If what you described is correct then you might start with:
Open Activity Monitor in the Utilities folder. Select All Processes from the Processes dropdown menu. Click twice on the CPU% column header to display in descending order. If you find a process using a large amount of CPU time, then select the process and click on the Quit icon in the toolbar. Click on the Force Quit button to kill the process. See if that helps. Be sure to note the name of the runaway process so you can track down the cause of the problem.

What is CPU cost?

dear all
what is CPU cost in execution plan?
thanks
john

hi,
I like one of the finest explanation and demo was provided in Charles hooper blog. I would have explained the same, but it's own's by charles itself.
refer to his link.. http://hoopercharles.wordpress.com/2010/02/19/what-is-the-meaning-of-the-cpu-column-in-an-explain-plan/
I like the most.
HTH
- Pavan Kumar N

CPU Utilization check

Hello,
I have a situation where the oracle CPU spikes upto 99% ( avg about 50-60 %) during a specific period (about 3-4 hrs ) every day. As a result of this the entire application slows down. We are not running any additional processes (related to application) nor user load is incresed in this period. I would like to find out what processes are consuming so much of CPU.
Does anybody know how to find out more about processes/queries that are using more resources/CPUs?
TIA

Hi,
I would like to find out what processes are consuming so much of CPU. This "ps" command works great for me:
Display top CPU consumers with ps
The following command can be used to display the top CPU consumers on any UNIX server. Note that the sort +6 command displays the CPU column. This is because the ps columns are counted from left-to-right, with the first column being column zero.
root> ps -ef|sort +6|tail
root 512 1 0 Aug 10 ? 0:53 /usr/lib/osa/bin/arraymon
root 460 1 0 Aug 10 ? 1:22 /usr/sbin/cron
root 1 0 0 Aug 10 ? 2:26 /etc/init -
root 565 1 0 Aug 10 ? 6:12 /usr/bin/perl /pixlog.pl
root 471 1 0 Aug 10 ? 9:10 /usr/sbin/nscd
netsaint 765 1 0 Aug 10 ? 27:16 helpdesk/netsaint.cfg
root 742 572 0 Aug 10 ? 32:04 opt/VRTSvmsa
root 454 1 0 Aug 10 ? 44:09 /usr/sbin/syslogd
root 3 0 0 Aug 10 ? 363:12 fsflush
Here we see the tasks in order of their total CPU consumption, with the largest value at the end of the list. In cases where a process has been running for more than one day, the display format changes. In these cases, the time format changes and column 5 is the CPU column.
root> ps -ef|sort +5|tail
root 5440 2094 0 Nov 21 - 0:47 /usr/sbin/syslogd
root 9244 1 0 Nov 21 - 3:26 ./pdimapsvr.ip -D0
root 10782 1 0 Nov 21 - 4:41 ./pdiconsvr.ip -D0
root 5990 2094 0 Nov 21 - 5:33 /usr/sbin/snmpd
root 4312 1 0 Nov 21 - 7:14 /usr/sbin/cron
root 4448 2094 0 Nov 21 - 9:25 /usr/sbin/rwhod
root 1 0 0 Nov 21 - 198:59 /etc/init
root 2450 1 0 Nov 21 - 438:30 /usr/sbin/syncd 60
Using ps auxgw command in HP/UX and AIX
Another approach to finding top CPU consumers on your UNIX server is to use the ps auxgw command. The third column of this listing (%CPU) shows the percentage of CPU used. Hence, the following command will display the top CPU users:
root> ps auxgw|sort +2|tail
oracle 14922 0.6 1.0 8300 5720 - A 01:01:46 2:57 oracleprod
oracle 22424 0.6 1.0 8328 6076 - A 07:48:43 0:21 oracleprod
http://www.dba-oracle.com/linux/monitoring_memory_processor.htm
Hope this helps. . . .
Don Burleson
http://www.dba-oracle.com

CPU process over 100% usage

On my mountain Lion 10.8.4 version the CPU utilization as reported by "activity monitor" showed a result greater than 100%. The TOR Browser went up to 100.9 % CPU with still further applications running. I have enclosed a screen shot of the exceeding utilization below
This is a serious concern because this may mean a hidden process not being properly reported..
Can anyone give a reason for this error?

It's not hidden. The problem is the Tor browser. Select it then click on the Quit icon and then on the Force Quit button.
Open Activity Monitor in the Utilities folder. Select All Processes from the Processes dropdown menu. Click twice on the CPU% column header to display in descending order. If you find a process using a large amount of CPU time (>=70,) then select the process and click on the Quit icon in the toolbar. Click on the Force Quit button to kill the process. See if that helps. Be sure to note the name of the runaway process so you can track down the cause of the problem.

Running slow: Finder hogging CPU?

My machine has been running slower with many spinning beach balls even with only a couple apps. open. In Activity Monitor Finder is always in the top 2 of the CPU column. Is this normal or how can I fix?
-Thank you

Welcome to the Forums!
What % of CPU is Finder using? You might try trashing the file com.apple.finder.plist from the ~/Library/Preferences folder, then rebooting.
Hope this helps...

CPU Running 100%

Since upgrading to Lion, my CPU is now running at almost 100% all the time.
Many thanks

Open Activity Monitor in the Utilities folder. Select All Processes from the Processes dropdown menu. Click twice on the CPU% column header to display in descending order. If you find a process using a large amount of CPU time, then select the process and click on the Quit icon in the toolbar. Click on the Force Quit button to kill the process. See if that helps. Be sure to note the name of the runaway process so you can track down the cause of the problem.

%CPU column in autotrace - 10gR2

Similar Messages

Maybe you are looking for