Compile times and run times for benchmarks run with gcc-3_0-branch, 20010212, on 500 MHz Alpha 21264 using gambc40a4, with 20 GVM registers and 8 argument registers.

Run times are in milliseconds, compile times are in seconds. All run times that are no more than 10% higher than the minimum run times are in bold; compile times that are more than twice the minimum compile times are in bold. Other compile options were

-mieee -fPIC -fno-math-errno -mcpu=ev6

Benchmark name -O1 -O1 -fschedule-insns2 -O2 -fno-gcse -O2
boyer 1.50 0.34 0.06 0.00
browse 0.96 1.64 0.07 0.00
cpstak 0.61 0.52 0.66 0.49
ctak 0.62 0.57 0.71 0.57
dderiv 0.67 2.18 0.05 0.00
deriv 0.63 1.93 0.74 2.13
destruc 0.66 0.71 0.04 0.00
diviter 0.62 0.49 0.65 0.59
divrec 0.59 0.91 0.06 0.00
fft 0.57 1.13 0.69 1.16
puzzle 0.78 0.58 0.07 0.00
tak 0.59 0.63 0.03 0.00
takl 0.60 0.66 0.04 0.00
trav1 0.90 0.39 1.13 0.42
trav2 0.83 1.96 1.15 2.02
triangl 0.70 1.43 0.03 0.00
ack 0.42 0.73 0.61 0.72
array1 0.61 2.78 0.63 3.12
cat 0.54 2.19 0.65 2.45
string 0.57 1.98 0.69 1.97
sum1 0.61 2.26 0.66 2.12
sumloop 0.52 0.38 0.61 0.38
tail 0.62 0.34 0.66 0.25
wc 0.57 1.09 0.03 0.00
conform 2.03 0.92 0.13 0.00
earley 1.69 0.73 0.20 0.00
fib 0.50 1.41 0.05 0.00
fibfp 0.53 0.35 0.02 0.00
maze 1.74 0.41 0.12 0.00
mazefun 1.21 1.04 0.06 0.00
mbrot 0.59 0.40 0.67 0.39
nucleic 4.61 0.34 0.51 0.00
paraffins 1.05 2.09 0.06 0.00
peval 2.39 1.30 0.12 0.00
pi 1.22 1.32 1.63 1.32
pnpoly 0.64 0.38 0.04 0.00
ray 1.02 0.54 0.07 0.00
scheme 5.08 0.46 0.19 0.00
simplex 1.00 0.30 0.08 0.00
slatex 6.91 0.86 0.32 0.00
sum 0.51 1.48 0.63 1.47
sumfp 0.54 3.67 0.57 4.43
tfib 0.51 0.94 0.56 0.93