Alpha 500 vs Bi Ppro.

Yann DUPONT (dupont@unantes.univ-nantes.fr)
Thu, 15 May 1997 19:02:08 +0200


Hello. We are testing some new machines, and I'd like to hear
experiences from others ...

Today we were comparing
a Bi Ppro 200 (128 Mo Ram) & A No-name Alpha 500 Mhz (21164 Ev5) 256 Mo Ram
1Mo cache.

Both were loaded with redHat 4.2. (Linux 2.0.30)

To our great disapointment, it seems that the pentium Pro is faster
than the Alpha. We weren't expecting that.

Ok, maybe we didn't test it the right way, and we will give it another
try with a more serious benchmark. Anyway, today we tested:

Compilation :

For exemple we compiled Python 1.4
(make -j)

on bip (Bi pentium pro) => 26 sec.

on Alpha => 57 sec

Ok, maybe the code generation on alpha is longer ?

Something we don't understand :

python pystone.py (a little loop program, not very serious abot real
performance)

1) Pic (1 processor only) 1600 pystones
2) Alpha 500 under linux 1500 pystones
3) Alpha 250 under Digital Unix 1300 pystones

I expected 2) to be somewhere in the 2000/2500 range...

Load :

We compiled bogomips (bogo-1.2), and launched 256 in the same time,
then launched xdvi & ghostscript, and compilations in the same time .

-> load jump over 100 on the 2 stations, but the system is still
responsive on the Bip & on the Alpha. With a load of 100 !!! I think
it's really good (there wasn't disk activities at this time)

compilations time are in the same range (the bip is still winner)
The bip also finish the huge bogomips test a little before the alpha.

When the load is at it's MAX, trying to refresh xdvi (with a
postscript in it) is quite fast & smooth on the ppro, much slower &
less smooth (that is, refresh is done little bit by little bit)
on the ALPHA...

Also tried the Unix Byte benchmark. The results are better on the
Alpha on most cases (2x to 10x !!!! improvement).
But in real use, This win seems not to be perceptible...

Those machines are aimed to be used by computer science students, and
with probably a constant load of 10,20 or more students... with a lot
doing C...

We didn't test floating point yet, it's not a priority for us. Maybe
this is a point where Alpha can blast the Ppro ?

I'd like to know if the alpha results are Ok, of if we have problems
with the configuration (I have to say that i'm not a guru on Alpha...)

In any case, I have to say that I've been really impressed by those 2
machines running linux !

(imagine : Compiling 2.1.37 on The Bip : make -j zImage => 2:40 !!!!)

Yann.

--
 _ _  ___      _________________________________  
| | || . \    | Fac des sciences de Nantes.     |
\   /| | | _  | Tel : 40373069 / 40374945       |
 |_| |___/<_> | ydupont@sciences.univ-nantes.fr |
	       ---------------------------------