Benchmark (was: Re: Junior kernel hacking)

Matthew Dillon dillon at apollo.backplane.com
Wed Jan 14 19:23:31 PST 2004


:I don't have any of those options. Here is my config file. I use more
:or less the same config file in all cases.
:
:Jean-Marc
:
:machine		i386
:cpu		I486_CPU
:cpu		I586_CPU
:cpu		I686_CPU

    Get rid of I486_CPU (though I don't think it would make a huge
    difference).

:ident		WORM
:maxusers	32

    Set maxusers to 0 so it autoconfigs.  This could be why the 5.x I/O
    throughput tests are so bad.  Then again, 5.x is far more conservative
    on its buffer cache / KVM use then we are anyway (but it's their own
    fault.  We can afford to be more liberal because we no longer split
    KVM into two kernel VM maps).

:# To make an SMP kernel, the next two are needed
:options 	SMP			# Symmetric MultiProcessor Kernel
:device		apic			# Symmetric (APIC) I/O
:...
:-- 
:Jean-Marc Zucconi -- PGP Key: finger jmz at xxxxxxxxxxx [KeyID: 400B38E9]

    I was about to say try turning off SMP, but then I went back to your
    original posting and you are indeed running on an SMP box, so SMP 
    certainly applies.

    You might want to try 5.x's SCHED_ULE instead of SCHED_4BSD.  I don't
    expect it to make a huge difference but ULE is supposed to deal with
    affinity better so who knows.

    * Make sure the filesystem partitions for DFly and 5.x, if different, are
      using the same block and fragment sizes.  'tunefs -p partition' ought 
      to tell you.

    * Check the IDE setup using 'atacontrol info 0', (1, 2, 3, etc..) under 
      both OS's and 'atacontrol mode 0', (1, 2, 3, etc..) under both OS's.

					-Matt
					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>





More information about the Kernel mailing list