buildworld times

Bob Willcox bob at immure.com
Sun Feb 29 11:34:21 PST 2004


On Sun, Feb 29, 2004 at 07:52:31PM +0200, Evren Yurtesen wrote:
> Sorry, my mistake...I guess it was taking longer than 15 minutes... :(
> 41m51.44s real          24m52.18s user          14m56.58s sys
> 
> I dont know how could I get 15 last time, if I forgot to do make clean 
> or something, or a lot of stuff was cached in the memory maybe? I think 
> I did it 2 times in a row :) I guess doing from a memory disk would 
> enhance things a lot.

I've never experienced _that_ big (3-fold) of difference when running
buildworlds back-to-back though.

> 
> Actually now I am running squid proxy on this machine with 8-16% cpu 
> usage all the time. Maybe that might be the problem also.

Could be, but the differences still seems excessive.

> 
> Good question, how do I know my kernel debug options? I think I didnt 
> compile them in. I thought the version of the freebsd could be seen from 
> dmesg output. It is cvsupped freebsd-current at about feb 4-5...

If you are running the stock GENERIC -current kernel then I would expect
the following (standard) debug options to be set:

# Debugging for use in -current
options         DDB                     # Enable the kernel debugger
options         INVARIANTS              # Enable calls of extra sanity checking
options         INVARIANT_SUPPORT       # Extra sanity checks of internal structures, required by INVARIANTS
options         WITNESS                 # Enable checks to detect deadlocks and cycles
options         WITNESS_SKIPSPIN        # Don't run witness on spinlocks for speed

In my experience, WITNESS increases my buildworld times by a factor of
about 2 (this is on an AMD Opteron 146 system). I see a big increase in
system time (as expected). For example here are two buildworlds done
back-to-back on this system (running -current as of yesterday) with the
only change being the removal of INVARIANTS and WITNESS:

logs/qui-gon-3:     4183.40 real      1625.23 user      2462.64 sys
logs/qui-gon-4:     1909.12 real      1478.65 user       355.25 sys

Bob

> 
> 
> 
> Bob Willcox wrote:
> >On Sun, Feb 29, 2004 at 12:02:41PM +0200, Evren Yurtesen wrote:
> >
> >>It takes 15 minutes on this one with make buildworld...
> >
> >
> >Interesting, but you don't mention what FreeBSD version or what your
> >kernel debug options are. Mind sharing those?
> >
> >Bob
> >
> >
> >>Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
> >>       The Regents of the University of California. All rights reserved.
> >>FreeBSD 5.2-CURRENT #0: Fri Feb  6 13:59:01 EET 2004
> >>   toor at proxy.ispro.net.tr:/usr/obj/usr/src/sys/PROXY
> >>Preloaded elf kernel "/boot/kernel/kernel" at 0xc075a000.
> >>Preloaded elf module "/boot/kernel/splash_bmp.ko" at 0xc075a21c.
> >>ACPI APIC Table: <A M I  OEMAPIC >
> >>Timecounter "i8254" frequency 1193182 Hz quality 0
> >>CPU: Intel(R) Pentium(R) 4 CPU 3.06GHz (3059.02-MHz 686-class CPU)
> >> Origin = "GenuineIntel"  Id = 0xf29  Stepping = 9
> >>
> >>Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE>
> >> Hyperthreading: 2 logical CPUs
> >>real memory  = 2142437376 (2043 MB)
> >>avail memory = 2094530560 (1997 MB)
> >>FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
> >>cpu0 (BSP): APIC ID:  0
> >>cpu1 (AP): APIC ID:  1
> >
> >

-- 
Bob Willcox                  A person is just about as big as the things
bob at immure.com               that make him angry.
Austin, TX


More information about the freebsd-current mailing list