[prev in list] [next in list] [prev in thread] [next in thread] 

List:       oprofile-list
Subject:    Re: op_time
From:       William Cohen <wcohen () redhat ! com>
Date:       2005-06-15 14:44:09
Message-ID: 42B03EB9.40508 () redhat ! com
[Download RAW message or body]

nem.me.viu nem.me.viu wrote:
> Hello all,
> 
> I installed oprofile 0.9 in a fedora core 3 box, I have followed the
> instructions in the docs.
> 
> I have compiled a kernel 2.6.11.2 from tarballs, including the
> oprofile module support and APIC enabled

It looks like OProfile is using the fall-back timer mechanism. This 
would be either because the processor can't be identified or the 
interrupt mechanism isn't available. Given this is a pentium III, I 
suspect that the problem is due to the interrupt mechanism. For a UP 
kernel Make sure you have the following in your config file:


CONFIG_X86_GOOD_APIC=y
CONFIG_X86_UP_APIC=y
# CONFIG_X86_UP_IOAPIC is not set
CONFIG_X86_LOCAL_APIC=y

Alternative just use an SMP kernel.


> when I am trying to analize data with opreport, I can't see the
> processor speedy, only 0 MHz as below.
> 
> opreport --merge=all
> 
> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
> Profiling through timer interrupt
>           TIMER:0|
>   samples|      %|
> ------------------
>   5748744 68.1128 no-vmlinux
>   1762196 20.8790 vmlinux
>    672089  7.9631 libbfd-2.15.92.0.2.so
>     84555  1.0018 Xorg
>     71489  0.8470 libc-2.3.3.so
>     13393  0.1587 libglib-2.0.so.0.400.7
>     12838  0.1521 libvte.so.4.4.0
>     11274  0.1336 libopcodes-2.15.92.0.2.so
>      9788  0.1160 libgobject-2.0.so.0.400.7
>      8926  0.1058 libgdk-x11-2.0.so.0.400.13
>      6124  0.0726 tiotest

It looks like you are mixing data from two different runs. See the 
"no-vmlinux" and "vmlinux" entriess.

> Is that correct or is that a specific processor problem?
> 
> Is there a way to tell oprofile to understand the speedy of my processor?

There isn't a way to tell OProfile the processor speed. OProfile tries 
to identify the processor speed when using the performance monitoring 
hardware. The fallback mechanism doesn't try to identify the processor 
speed.

> Another question I did look at some docs and there is a tool named
> "op_time", I can´t find this tool in the 0.9 version, is there a
> replacemente for op_time ?

optime (no space)

-Will


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id492&opĚk
_______________________________________________
oprofile-list mailing list
oprofile-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oprofile-list

[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic