Actually, we just had a reason to care. We just changed from raid 5 to raid 
1+0, and wanted to determine how and where we achieved performance increase. We 
weren't interested in knowing how we compared to a museum machine, but in how 
we compared to the weekend before. We did see quite a performance increase in 
disk related functions, but it's rather difficult to quantify just how much 
faster it is overall for any given user in a normal work day.

Charlie Noah
Inland Truck Parts

So, assuming the Spirit was a 1x machine (I seem to recall a small DEC 1400 
(?) being a 2x - sooo many years) if the spirit could complete 10,000 
"transactions" in a quanta of time, and the new machine finishes the same task 
(approximating real world environment) in some fraction of this time, then it should 
be fairly straight forward to work out the X rating.

I seem to recall that the "omnipresent" CUBS benchmark was trying to achieve 
the same thing ..... they may even have some old benchmarks from a known "X 
rating" machine, allowing an approximation of modern equipment to be made --> 
not that I think anyone really cares these days, as X tends to be sufficiently 
large !

Ross Ferris
Stamina Software
Visage â an Evolution in Software Development
u2-users mailing list

Reply via email to