[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
more on workstation speed.
Date: Fri, 14 Aug 87 10:15 EDT
From: Daniel L. Weinreb <DLW@ALDERAAN.SCRC.Symbolics.COM>
Date: Wed, 12 Aug 87 17:22 EDT
From: Jeffrey Del Papa <dp@JASPER.PALLADIAN.COM>
I propagated an error when giving compile times for the various processors. It turns
out that the compile time I mentioned for the apollo was only for the common "core"
system of our products. If you compile the whole thing (which the other times were
for) you spend 38 hours. Also for comparison purposes it takes 7.2 hours for an 8mb
400k lines of code, times in hours (a summary for those who missed the prev)
explorerII 16mb 1.5
symbolics 3620 16mb 4.5
explorerI 8mb 7.2
apollo 3000 8mb, lucid 36.
I'd like to make sure everyone understands that these numbers should NOT
be considered comparisons of processor speed. The Symbolics compiler,
the TI compiler, and the Apollo compiler are completely unrelated and
very different programs. These numbers reflect not only processor
speed, but the speed of the compilers as well.
I admit that these are not a direct comparison of raw machine speeds. The compilers
are different. However it is a measure of machine utility, if a machine lets me do
the things I do most, quickly, I don't completely care if it is because the one has a
hot processor, and the other a blindingly fast compiler, since I spend my time
developing code, (and don't run it much after it is tested) compiler wall clock is
DLW protested privately that the symbolics compiler has not been tweaked for
performance lately. It is still faster than compiler that TI uses (the compiler was
esentially re-implemented as part of the 3600 project), as the time difference for
the explorerI - 3620 comparison are slower than the processor speed differences would
by the way no one complained about the typo on the apollo times, the correct number
is 38 hours.