Re: Aw: [vhdl-200x] Next Meeting: Thursday March 17 + posted 3/3 minutes

From: <hans@ht-lab>
Date: Fri Mar 04 2011 - 08:31:31 PST

Hi Gerhard,

-----Original Message-----
From: dk4xp@arcor.de
Sent: Friday, March 04, 2011 1:01 AM
To: vhdl-200x@eda.org
Subject: Aw: [vhdl-200x] Next Meeting: Thursday March 17 + posted 3/3
minutes

Is there anywhere a list of what has been proposed/rejected/approved
already?

[Hans] No not yet, there were some discussions on a Wiki but I am not sure
if anything has changed.

My pet topics:

int32, int64, int128, uint32, uint64, uint128, with 0x80000000 etc
guaranteed working.
2's complement assumed, these Cyber205s are dead for > 20 years and not
many people had one.
And probably no one doing VHDL.

Verifying a 64 bit multiplier is a r. pain if you cannot even write the
numbers you want to see.

[Hans] I agree, although as you probably know that verifying a 64bits
multiplier is not possible within the current space-time continuum. My
formal tool already dies with anything more than 14 bits.

Easy access to the bits in an integer. Yes, I'm a heretic.

[Hans] Me too.

overloading := and <= ... I'm already running, seeking shelter

The equivalent of the C function system()... Could help to get your
login name , date, time, machine name and many other things into a file...
Or to prepare some test bed functions in Matlab or C etc.

[Hans] This is easily done with a bit of Tcl and "forcing" the numbers into
your VHDL code. For communication with Matlab and other applications have a
look at sockets, again easily done with a bit of Tcl. The discussed VHDL-DPI
interface might make this even easier.

Maybe writing / reading unix-style pipes. Would ease co-simulation
with 3rd party soft/hardware.

[Hans] Also not difficult since named pipes are like files. I played with
this many years ago and from what I remembered it worked fine under Linux
but failed under Cygwin/msys/dosbox using a popular simulator ;-)

And, while I have not yet tried the 2008 type generics, my first impression
was that one has to explicitely give the type in the generic map section.
It would be nice if one could simply write
   signal intermediate_result: arg1'type; inside an entity
declaration.

Maybe that would require things like "arg1'type'high downto arg1'type'low"
to be really useful.

(Kann sein, das ich hier eine offene Tür einrenne.)

[Hans] at the moment all doors are still open :-)

Regards,
Hans.
www.ht-lab.com

regards, Gerhard

----- Original Nachricht ----
Von: Jim Lewis <Jim@synthworks.com>

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Fri Mar 4 08:35:40 2011

This archive was generated by hypermail 2.1.8 : Fri Mar 04 2011 - 08:35:46 PST