paper numbers - compile times

[Stephen Weeks]

Here are the compile time numbers for everthing except for the kit.
There are two versions of each benchmark: the first is with -O2, the
second with -O1 (of course this only expains the difference in the gcc
column -- all other columns should remain unchanged).

		NJ	smlc	gcc	convert	flow
barnes-hut	1.55	3.32	1.95	0.280	0.010
barnes-hut	1.63	3.35	1.53	0.300	0.010
count-graphs	1.19	2.54	1.96	0.230	0.0
count-graphs	1.23	2.56	1.52	0.250	0.010
fft		0.94	2.12	1.47	0.200	0.0
fft		0.95	2.17	0.93	0.210	0.0
knuth-bendix	2.78	3.46	3.91	0.320	0.020
knuth-bendix	2.65	3.46	3.05	0.320	0.010
lexgen		4.66	9.19	10.32	0.760	0.030
lexgen		4.49	9.18	7.98	0.790	0.030
life		0.89	1.94	1.61	0.160	0.010
life		0.9	1.95	1.29	0.160	0.0
logic		2.12	6.25	279.17	0.350	0.010
logic		2.07	6.35	203.27	0.340	0.010
mandelbrot	0.23	0.87	0.33	0.060	0.0
mandelbrot	0.23	0.86	0.32	0.060	0.0
matrix-multiply	0.26	0.74	0.37	0.070	0.0
matrix-multiply	0.26	0.72	0.35	0.070	0.0
mlyacc		29.89	92.43	75.94	2.270	0.090
mlyacc		30.13	90.06	54.76	2.340	0.100
nucleic		7.02	7.62	2.05	0.400	0.010
nucleic		7.23	7.64	1.73	0.430	0.020
ratio-regions	2.11	4.03	4.09	0.340	0.010
ratio-regions	2.2	4.01	2.12	0.350	0.010
simple		4.73	15.72	23.07	0.850	0.040
simple		4.66	15.63	18.51	0.860	0.040
tsp		0.82	2.48	1.53	0.210	0.0
tsp		0.83	2.48	1.06	0.220	0.010
vliw		16.31	25.69	53.95	1.600	0.060
vliw		16.65	25.74	43.4	1.640	0.060
zern		0.89	2.15	0.79	0.160	0.0
zern		0.89	2.13	0.65	0.150	0.010
smlc		620.11	321.67	1540.33	33.490	2.010
smlc		637.52	321.92	1350.15	33.840	1.940