Ada Timing Analysis

Ada Timing Analysis

[Bob McIntire]

Our office is preparing a proposal in which we are required to estimate the 
amount of CPU time the processes will take in the finished application.  We 
have done an estimate of the number of lines of Ada code we think the 
application will require.  Now, we are trying to "translate" that number to 
lines of generic machine language code.  From that, using spec marks, we can 
determine how much of the CPU will be used by each process. 

My question is, does any one have a reasonable number for the ratio of Ada 
source to machine language statements??

Bob McIntire
bmcintir@csc.com
Computer Sciences Corp.

Re: Ada Timing Analysis

[Robert Dewar]

Bob asked

"My question is, does any one have a reasonable number for the ratio of Ada
source to machine language statements??"

There is no such reasonable number. The reason is that it varies 
tremendously depending on the style of use of the language. If your code
is full of aggregates, protected objects, slices etc, the number will
be FAR higher than if you tend to use a relatively straightforward
subset of Ada.

One useful tool in getting a feel for this when using GNAT is to use the
-gnatdg switch to generate the intermediate Ada source code after
expansion, which is much less subject to this kind of variation, since
it is basically at the WYSIWYG level.

Re: Ada Timing Analysis

[Bob Kitzberger]

Bob McIntire (bmcintir@csc.com) wrote:

: My question is, does any one have a reasonable number for the ratio of Ada 
: source to machine language statements??

That's highly dependent on the CPU architecture, the compiler, and
any optimizations selected.  I suggest that you contact your
vendor (or potential vendors) for assistance.  Perhaps a small
sample program can be put through the paces to come up with
the baseline metrics you need.

--
Bob Kitzberger	      Rational Software Corporation       rlk@rational.com
http://www.rational.com http://www.rational.com/pst/products/testmate.html

Re: Ada Timing Analysis

[Philip Brashear]

In article <4t0iej$2c9@explorer.csc.com>,
Bob McIntire <bmcintir@csc.com> wrote:
>
>My question is, does any one have a reasonable number for the ratio of Ada 
>source to machine language statements??


This ratio varies tremendously with the kind of statements and the particular
compiler.  The Ada Compiler Evaluation System (ACES) produces massive amounts
of data regarding code expansion.

See http://sw-eng.falls-church.va.us/adaic/testing/aces

or contact me directly.

Phil Brashear
CTA INCORPORATED
brashear@sw-eng.falls-church.va.us
brashear.cta@juno.com
(513) 258-0831

Re: Ada Timing Analysis

[Marin David Condic, 407.796.8997, M/S 731-93]

Philip Brashear <brashear@NS1.SW-ENG.FALLS-CHURCH.VA.US> wrote:
>Bob McIntire <bmcintir@csc.com> wrote:
>>
>>My question is, does any one have a reasonable number for the ratio of Ada
>>source to machine language statements??
>
>
>This ratio varies tremendously with the kind of statements and the particular
>compiler.  The Ada Compiler Evaluation System (ACES) produces massive amounts
>of data regarding code expansion.
>
    You will also find an *enormous* variation depending on the target
    machine. For example: VAX Ada83 (which I like very much) will
    generate pretty massive files in comparison to it's cousin, the
    XD-Ada83 compiler targeting the 1750a. The reason is simple: VAX
    Ada doesn't need to be terribly concerned with image size so they
    include more elaborate "support" (debugging stuff, more powerful
    run-time support, "optional" features, etc.) whereas the poor,
    little 1750a with it's 64kwords of memory (and operating in
    "real-time" as well) hasn't got any room to spare.

    There's nothing inherent in Ada that demands grossly large code.
    Pick your target, decide how important size is to you, then look
    for the best implementation for the job.

    MDC

Marin David Condic, Senior Computer Engineer    ATT:        407.796.8997
M/S 731-96                                      Technet:    796.8997
Pratt & Whitney, GESP                           Fax:        407.796.4669
P.O. Box 109600                                 Internet:   CONDICMA@PWFL.COM
West Palm Beach, FL 33410-9600                  Internet:   CONDIC@FLINET.COM
===============================================================================
    "Government is not reason. It is not eloquence. It is a force.
    Like fire, a dangerous servant and a fearful master."

        --  George Washington
===============================================================================