FreeBSD on dual core Opterons - stupid buildworld test

[Attila Nagy]

Hello,

I did some buildworld tests on an HP BL45p with 4 2,2 GHz Opterons.

Sequential time make buildworld -jX times (if there is a missing 
iteration, then the build has failed):

4BSD:
BUILDING WITH -j1
1725.073u 645.453s 37:50.03 104.4%      4084+2935k 1725+3896io 385pf+0w
BUILDING WITH -j2
1764.483u 1020.054s 17:36.07 263.6%     3987+2766k 1735+3748io 385pf+0w
BUILDING WITH -j3
1806.182u 1564.830s 13:38.53 411.8%     3926+2605k 1741+3939io 385pf+0w
BUILDING WITH -j4
1827.213u 1781.092s 12:31.84 479.9%     3900+2559k 1736+3847io 385pf+0w
BUILDING WITH -j5
1832.213u 1942.488s 12:13.05 514.9%     3862+2522k 1731+3838io 385pf+0w
BUILDING WITH -j6
1844.549u 2054.711s 11:56.27 544.3%     3855+2489k 1727+3868io 385pf+0w
BUILDING WITH -j7
1849.804u 2156.657s 11:45.74 567.6%     3828+2460k 1751+3831io 385pf+0w
BUILDING WITH -j8
1859.690u 2208.954s 11:49.85 573.1%     3832+2447k 1755+3866io 385pf+0w
BUILDING WITH -j9
1855.253u 2245.942s 11:55.85 572.9%     3825+2428k 1755+3845io 385pf+0w
BUILDING WITH -j10
1867.634u 2258.330s 12:00.68 572.5%     3818+2441k 1743+3943io 385pf+0w
BUILDING WITH -j11
1858.855u 2302.655s 12:06.67 572.6%     3806+2418k 1740+4094io 456pf+0w
BUILDING WITH -j13
1863.750u 2319.476s 12:06.90 575.4%     3794+2411k 1753+3999io 391pf+0w
BUILDING WITH -j14
1859.940u 2318.790s 12:07.43 574.4%     3795+2409k 1740+4112io 385pf+0w
BUILDING WITH -j15
1862.859u 2329.179s 12:10.75 573.6%     3789+2396k 1744+4261io 385pf+0w

ULE:
BUILDING WITH -j1
1685.692u 532.736s 36:36.58 100.9%      4200+3011k 1983+3727io 385pf+0w
BUILDING WITH -j2
1741.276u 834.259s 17:58.30 238.8%      4124+2871k 1831+4111io 385pf+0w
BUILDING WITH -j3
1769.405u 1113.314s 14:19.19 335.5%     4087+2767k 1850+4065io 385pf+0w
BUILDING WITH -j4
1788.789u 1243.970s 13:16.23 380.8%     4063+2731k 1759+4202io 385pf+0w
BUILDING WITH -j5
1797.793u 1363.709s 12:49.00 411.1%     4037+2688k 1762+4169io 385pf+0w

http://people.fsn.hu/~bra/bl45p/acpidump
http://people.fsn.hu/~bra/bl45p/dmesg.boot
http://people.fsn.hu/~bra/bl45p/pciconf

It is interesting that the CPU usage couldn't exceed 575 percent... But 
HP says that dual core increases performance about 75 percent, so given 
that this machine had 4 CPUs it is seems OK :)

-- 
Attila Nagy                                   e-mail: Attila.Nagy at fsn.hu
Adopt a directory on our free software   phone @work: +361 371 3536
server! http://www.fsn.hu/?f=brick             cell.: +3630 306 6758