Ok...
Builds run under:
Linux dominion 2.1.33 #28 Sat Apr 12 03:30:10 EDT 1997 i586
Tyan Tomcat III, BIOS 4.00E (12/12/96), Dual P200, 512k cache, 128M 60ns EDO,
AHA2940UW + Quantum XP34300W (L915),
PPP link up w/ ping -i 120 active.
GNU Make version 3.74
GCC Version 2.7.2.1
make -j1: 100% Adj
563.070u 25.260s 12:07.04 80.9% 0+0k 0+0io 143261pf+0w -> 588.17536
550.470u 24.250s 11:44.79 81.5% 0+0k 0+0io 141999pf+0w -> 574.40385
549.930u 24.660s 11:45.99 81.3% 0+0k 0+0io 141806pf+0w -> 573.96987
553.870u 24.010s 11:30.67 83.6% 0+0k 0+0io 139640pf+0w -> 577.40012
538.460u 22.690s 11:26.27 81.7% 0+0k 0+0io 137234pf+0w -> 560.68259
-----------------
Total: 2874.63179
Avg: 11:42.95 81.8% 574.92635
make -j2: 100% Adj
624.220u 29.480s 7:15.00 150.2% 0+0k 0+0io 137534pf+0w -> 653.37000
614.540u 36.250s 7:21.49 147.4% 0+0k 0+0io 137608pf+0w -> 650.75626
628.970u 40.580s 7:37.47 146.3% 0+0k 0+0io 140654pf+0w -> 669.27861
630.950u 40.720s 7:37.73 146.7% 0+0k 0+0io 140550pf+0w -> 671.48991
630.840u 39.180s 7:36.67 146.7% 0+0k 0+0io 140009pf+0w -> 669.93489
-----------------
Total: 3314.82967
Avg: 7:29.67 146.5% 662.96593
Avg Improvement: 04:13.28 or 1.56 x faster with both processors.
NOTE:
1) The sample size is too small to be statistically valid.
2) I'm too tired for the numbers to really make sense.
3) Every build gets: (syslog) kernel: Exception at [<c0137e5d>] (c01a610e)
^^^^^^^^^^^^ == getname
4) A dual PPro 200 (PR440fx mb) loads up to 184% CPU; and compiles it's
kernel in just over _3_ minutes.
5) The entire kernel source tree should -- theoretically -- have been in
memory cache during the compile(s).