
Speed of Oberon on DOS and SPARC Systems
Recent discussion has said various things about Oberon performance on
DOS machines. So I thought it might be useful to mention my experience.
There is a set of number-crunching oriented benchmarks that do the sort
of mathematical stuff that I am interested in, and which I have run on
some tens of machines in the PC and non-PC world under various languages
during the last 14 years.
Oberon-2 DOS, running these benchmarks on a 486DX2/66, has performance slightly
less (~ 5%) than that of the same benchmarks in Oberon-2 on a Sun Sparc2.
The timings for both are in between the times of these benchmarks (in C and
Professional Pascal) on a Sun IPC and a Sparc2.
I therefore see no major performance penalty in Oberon-2 on either system.
The readme.txt information from ETHZ on performance of SPARC Oberon-2 seems
basically correct (very good, comparable to -O4 optimized C code), and the DOS
version comes out (on average) the same on a Gateway 2000 486DX2/66V.
For those with 386 machines, the same Oberon-2 benchmarks are a factor of 10-15
slower than the 486DX2/66 on a Grid 1750 386&387DX/20.
So, I think you can crunch away in Oberon-2 or Oberon System3 if you wish.
============================================================================
National Radio Astronomy Observatory | | \/ o\
P.O. Box O | Telephone: 505-835-7273 | /\__/
(1003 Lopezville Road - UPS etc. only)| |
Socorro, NM 87801-0387 USA | FAX: 505-835-7027 | AMDG
============================================================================