Do you know where can I find information about XBENCH 3.0 for
multivalue BASIC ?
I found a reference to this benchmark on the Univision site (see link
below) but there are no details and somehow I don't trust the numbers
as they are presented in their report.

http://www.fastfreenet.com/support/u...h_linuxAS4.htm

In my experience the fact that a particular instruction in a
particular situation runs faster or slower is not indicative of the
average program performance.

Lucian

I've never seen XBENCH for MV, but the fact that they are making
reference to UV8 should give you a pretty solid indication that even
if the figures were accurate at the time, they a probably meaningless
now ..... unless of course you were looking at deploying over RH 7.1.

I share your mistrust, and have always found it worthwhile to write my
own little test rigs to perform benchmarks, and also to answer the
perennial question of "what is the fastest method to <insert problem
of interest today>", and have found the "right" aswer may vary with
platform (*nix tends to be faster in my experince), resources (more is
always good) and version (sometimes newer versions may speed things
up, introduce a new method of solving a problem, OR they may have
stuffed things up and the previous option 2 is now faster)

On Feb 25, 1:44*am, Ross Ferris <ro... (AT) stamina (DOT) com.au> wrote:
If it didn't look good it would not be published. But if you really
believe benchmarks then I've got a bridge you might like to buy.
BobJ

I don't know, may be this XBENCH for multivalue BASIC has merits but
as they are presented the results don't make much sense to me. For
example adding 2 numbers. What kind of ADD are we talking about ?
R = A + 45
R = A<1> + B<3>
R = A<1000> + B<450>
R = V(15)<43,2> + 14.27
These are vastly different in scope and execution times.
For example if I add A<1000> to something we have a dynamic array
search plus a string conversion and then finally the actual add takes
place. In my experience, very seldom I have R = 13 + 14 but mostly R
= A<13> + B<14> hence for me the speed of the first type of ADD is
irrelevant.
Without the pertinent details those XBENCH results look like
misleading advertising.
And yes, I am very interested to buy the Brooklyn bridge if we can
bargain it down to a couple of thousands. I am going to contact you
later this week because over the weekend I am flying to Paris to sign
the papers for the Eiffel tower. I just made a huge down payment for
it to a Nigerian bank ;-)

Lucian

On Feb 25, 6:00*pm, x <lucian_p... (AT) yahoo (DOT) com> wrote:
You can leave the down payment on the bridge with the same Nigerian
Bank
BobJcates

On Feb 25, 1:08*am, x <lucian_p... (AT) yahoo (DOT) com> wrote:
I agree with the limitations of such a benchmark. A better one, if
you can find it, was the CUBS benchmark. This simulated typical
application processing with multiple users. The number of users
simulated is a parameter you can set when you run it so you can review
different scenarios.

On Mar 7, 10:09*am, Rich Taylor <richtaylor0... (AT) gmail (DOT) com> wrote:
If you do find the CUBS benchmark, please let me know. I would be
very interested in that as well.

Eugene

I haven't found any multivalue centered benchmarks available for
public consumption (download).
Based on a few articles mentioning CUBS benchmark, looks like the
benchmark in question is tailored to CUBS software therefore may or
may not be relevant to greyware (run of the mill software).

Lucian