[ixp1200] RE:

[Michael Kounavis]

THe uof file does include information about the
microcode associated with a project.

If I remember correctly this is either the .list file
or the "C" structure characterizing the encapsulated
image, all in ASCII format. This is why the size is much
larger  than the actual code encapsulated.

A set of libraries that parse the .uof file format
can also be found as part of the NetBind source code
distribution at:

www.comet.columbia.edu/genesis/netbind

--
Michael



On Tue, 16 Jul 2002, Johnson, Erik J wrote:

> The large size is most likely due to the inclusion of debugging information
> (the -g option).  Of course the uof file format will have some overhead for
> magic numbers and headers, even without debugging information, but most of
> the extra space comes from the symbols (i.e., strings) included for
> debugging.  I am not a uof file expert, but this is my best guess.
>
> Both the assembler and compiler indicate how many instructions were
> generated with a line like the following:
>
> Number of uwords compiled: 620
>
> erik
>
> -----Original Message-----
> From: ravi modgekar [mailto:ravilm at yahoo.com]
> Sent: Tuesday, July 16, 2002 7:15 AM
> To: ixp mailing list
> Subject:
>
>
> Hi,
>    Sorry this question is not related to VERA ,but I
> thought I could get my answer here.
>
> When I compile any of the INTELS SDK 2.01 reference
> projects ,then the size of the final .uof file is more
> than
> the capacity of the control store of the 6
> microengines For eg:For the L3Fwd16 reference
> project the final uof file size is 346KB.Now my
> questions are
> 1)How does this 346KB uof file gets loaded into 2kB of
> control store on all microengines.
> 2)How do I know when my uof file will exceed the 2kB
> limit of the control store.
> 3)Does the Linker throw any exception.
>
>  Any INTEL guys out there please help me out.
>  Thanks
>  Ravi
>
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Autos - Get free new car price quotes http://autos.yahoo.com
>

-- 
--
Michael Kounavis
http://www.comet.columbia.edu/~mk