Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search

Messages from 2650

Article: 2650
Subject: Re: [q][Reverse Engineering Protection]
From: michael.williams@armltd.co.uk (Michael Williams)
Date: 19 Jan 1996 10:57:56 GMT
Links: << >>  << T >>  << A >>
In article <DLCKD6.Lr@wombat.hanse.de>,
Bernd "Bernie" Meyer <root@wombat.hanse.de> wrote:
>hbaker@netcom.com (Henry Baker) writes:
>>Since the chips on the American cars are probably more sophisticated, they may
>>be doing more than their European counterparts.
>
>Do I detect a certain snobbishness here?

I think he's basing his opinion on the emission regulations in the states
being more stringent than those in the Europe. However I suspect some
European manufacturers (Audi/Porsche, BMW, Volvo, Jaguar all spring to
mind) might argue the point.

Mike.
       _____________________________________________________________________
\  x  / Michael Williams                      Advanced RISC Machines Limited
|\/|\/\ michael.williams@armltd.co.uk  Fulbourn Road, Cambridge, CB1 4JN, UK
|  |.__)"I might well think that Matti, ARM Ltd. couldn't possibly comment."


Article: 2651
Subject: PLD JDEC Files
From: swood@melpar.esys.com (S.G. Wood, Jr.)
Date: Fri, 19 Jan 96 13:45:15 GMT
Links: << >>  << T >>  << A >>
I'm searching for a tool (program) that will take a PLD JDEC file and 
reproduce the logic equations for the PLD.  We are currently trying to debug 
circuit cards with PLDs on them, however the documentation for the PLDs are 
out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that 
we are using are 22V10.  I would appreciate any pointers that people may have.

-Simon


Article: 2652
Subject: GRRR!!! Xilinx Makebits defaults changing
From: matthew@rd.bbc.co.uk (Matthew Marks)
Date: 19 Jan 1996 17:06:59 GMT
Links: << >>  << T >>  << A >>

: I have since been working on another old design where a microprocessor
: configures the chip (Peripheral Mode).  This ceased to function - and I had
: forgotton about the problem with the other design so wasted a similar amount
: of time diagnosing it!
: 
: Although it was easy to explain why the first design ceased to function, I am
: somewhat perplexed about the other.  It appears that the Xilinx chip is unable
: to produce the last (or it seems like the last two, when I hook a counter on
: it) RDY pulses to haul in all of the configuration.  Surely if the IOBs
: take over too early, this is a problem with the chip?
: 
: Incidentally, I managed to get the chip to configure without the -xa option
: by altering the length count - but I'm glad I've found out how to do it
: properly!
: 
: Matthew  matthew@rd.bbc.co.uk  My opinions, not Auntie's.

I think I may have had a similar problem to this just last week - a
3190A in peripheral mode would assert its DONE/PG_ pin but the IOBs 
remained in configuration mode (i.e. LDC low, HDC high and all others
high-Z). The solution, discovered at the end of a Xilinx app. note
after a day of probing around with the 'scope looking for dodgy bus
activity, turned out to be modifying the length count by a few bits using the 
"-lc=aligned_lc" option to makeprom. It has worked fine ever since.

Ian.

e-mail	ima@nemesys.co.uk			      Nemesys Research Ltd.
www	http://www.nemesys.co.uk	       		   14 Regent Street
'phone	+44 1223 566300			        	          Cambridge
fax  	+44 1223 566301			 	            England CB2 1DB
---


Article: 2653
Subject: Re: Programming Actels in circuit?
From: David Pashley <david@fpga.demon.co.uk>
Date: Fri, 19 Jan 96 18:01:47 GMT
Links: << >>  << T >>  << A >>
In article <hersman.821993240@aplcomm.jhuapl.edu>
           hersman@aplcomm.jhuapl.edu "Chris Hersman" writes:

"Does anyone know of a way to program an Actel FPGA in the circuit?
"
"I am aware that even if it is possible, it's not straight forward.
"
Apparently, the programming is a very complex and demanding matter. 
To my knowledge only three programmer vendors have managed to do it 
(BP Microsystems, Data I/O, SMS), so you're unlikely to be able to 
do it in circuit.

-- 
David Pashley 



Article: 2654
Subject: Re: GRRR!!! Xilinx Makebits defaults changing
From: ima10@donna.nemesys.co.uk (Ian Atkinson)
Date: 19 Jan 1996 18:25:27 GMT
Links: << >>  << T >>  << A >>

Article: 2655
Subject: Re: Emulation for a wireless chip
From: hutch@fpga.ee.byu.edu (Brad Hutchings)
Date: 19 Jan 1996 20:59:53 GMT
Links: << >>  << T >>  << A >>
No doubt I will regret this (I am already regretting it)... but here goes 

Jason Feinsmith at Xilinx wrote:

    JF> I have a study which was done by a researcher at University of
    JF> Toronto which does a full comparison of routing in the Xilinx
    JF> 4000 versus the Altera 8000 (Flex) family.  The summary you
    JF> may find interesting, and if you would like a faxed copy of
    JF> it, please send your fax number to me (my email is
    JF> jason.feinsmith@xilinx.com).  The conclusion implies that
    JF> Xilinx 4000 routes better than does the Altera Flex 8000.

    JF> I'll quote 2 summary items from the paper:

    JF> "3.1 Results for the Xilinx XC4000 FPGAs...  It is interesting
    JF> to note that for all circuits used, none of them becomes
    JF> un-routable even under the worst pin constraints.  This was
    JF> true even for the circuits that were very tightly packed, in
    JF> terms of percentage of available CLBs and I/O pins used."

    JF> and

    JF> "3.2 Results for the Altera FLEX 8000 FPGAs...  The Altera
    JF> FLEX 8000 FPGAs seem to be slightly susceptible to routing
    JF> failures uner random pin constraints in cases where the I/O
    JF> pin or logic element utilization is close to 100%. ... It
    JF> seems that system designers, when implementig a circuit using
    JF> FLEX 8000 FPGAs, should leave about 20% of the logic eements
    JF> and I/O pins free to avoid routability problems due to pin
    JF> constraints."


and then,

I (B. Hutchings) at Brigham Young flamed Jason. Rather than flaming
Jason, I should have said what I thought was wrong with the post. I
apologize for being harsh.  My main problem was that there was no
indication of where the paper had been published or who had wrote it.
That being said, I also disagree with the sweeping conclusion (which
might have been Jason's or the paper's, I could not tell). We do a lot
of work with a variety of FPGA devices here at BYU and while the
ability to route inspite of lots of pin constraints is important,
there are many other things to consider. For example, what happens to
your *timing* as you lock down pins and use more FPGA resources? That is
just one of several questions that the paper may even respond to (for
all I know) but wasn't addressed in Jason's post. Why not just post
the citation and let everyone draw their own conclusion?

Well, sorry to waste bandwidth...








-- 
            Brad L. Hutchings - (801) 378-2667 - hutch@ee.byu.edu 
Brigham Young University - Electrical Eng. Dept. - 459 CB - Provo, UT 84602
                       Reconfigurable Logic Laboratory



Article: 2656
Subject: Re: good interview questions ?
From: ccwest@ix.netcom.com (Bill Seiler)
Date: Sat, 20 Jan 1996 00:40:12 GMT
Links: << >>  << T >>  << A >>
Try this

//*************** Verilog Code *****************
module test ;

reg input ;
reg output ;

initial
  begin
    input = 1'b0 ;
    output = 1'b0 ;
    #10 input = 1'b1 ;
    #1 input = 1'b0 ;			
  end

always @(input)
  if (input)
    #10 output = 1'b1 ;
  else
    #5 output = 1'b0 ;

endmodule

//*************** Verilog Code *****************

/* I pulse the signal for 1 ns
   I expected the output to pulse 10ns later
  What happened is the ouput stayed high !
*/

Thanks
 
    ___  ___
   /   /\  /\           Bill Seiler
  /___/  \/__\          Circuit City / Patapsco West  
 /\   \  /   /\         3255-4 Scott Blvd, Suite 105
/__\___\/___/  \        Santa Clara, CA 95054
\  /   /\   \  /        408 982 5420 Direct
 \/___/  \___\/         408 982 5430 FAX
  \   \  /\  /\         ccwest@ix.netcom.com 
   \___\/__\/  \
        \   \  /        "Fold me up into 4D space."
         \___\/




muzok@msn.com (muzo) wrote:

>hi,
>I am looking for suggestions on good interview questions for a digital designer.
>Interesting verilog coding issues, synthesis questions, issues which shows the
>understanding of basic design fundemantals (meta-stability etc) are welcome.

>thanks

>muzo

>standard disclaimer




Article: 2657
Subject: Re: PLD JDEC Files
From: "Paul E. Bennett" <peb@transcontech.co.uk>
Date: Sat, 20 Jan 96 00:47:46 GMT
Links: << >>  << T >>  << A >>
In article <4do6ve$t7f@portal.gmu.edu>
           swood@melpar.esys.com "S.G. Wood, Jr." writes:

> I'm searching for a tool (program) that will take a PLD JDEC file and 
> reproduce the logic equations for the PLD.  We are currently trying to debug 
> circuit cards with PLDs on them, however the documentation for the PLDs are 
> out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that 
> we are using are 22V10.  I would appreciate any pointers that people may have.
> 
I believe (not certain though) that e-mailing mpe@mpe.demon.co.uk will get you 
some information on some such software. I think it was a part of their X-Shell 
package which is a Forth based Cross-Compiler System.

Their Tel: is +44 (0) 1703-780084.

-- 
Paul E. Bennett
peb@transcontech.co.uk
Going Forth Safely


Article: 2658
Subject: Re: GRRR!!! Xilinx Makebits defaults changing
From: Scott Kroeger <Scott.Kroeger@mei.com>
Date: Sat, 20 Jan 1996 08:44:00 -0600
Links: << >>  << T >>  << A >>
Matthew Marks wrote:
> 
> My gripe concerns Xilinx changing a default in makebits
> between versions 4.5 (?) and 5.1 (?).  
--SNIP--
> I had one design where the configuration ROM held 8 configurations and the
> Xilinx chip itself was able to initate re-configuration if it detected that
> this was necessary.  This stopped working.  (Master Parallel Mode)
--SNIP--
> Although it was easy to explain why the first design ceased to function, I am
> somewhat perplexed about the other.  It appears that the Xilinx chip is unable
> to produce the last (or it seems like the last two, when I hook a counter on
> it) RDY pulses to haul in all of the configuration.  Surely if the IOBs
> take over too early, this is a problem with the chip?
> 
> Incidentally, I managed to get the chip to configure without the -xa option
> by altering the length count - but I'm glad I've found out how to do it
> properly!
> 

I've run into this problem too and lost a day or two of work.
The length count wierdness is documented in the makebits chapter
as well as in an app note or two.  I was peeved at the makebits
command line change as well as the easter egg hunt I had to
undertake to discover a problem that a good portion of the Xilinx
community has probably suffered through.

I guess we've all earned our Xilinx purple hearts over this!

Cheers,
Scott


Article: 2659
Subject: Re: PLD JDEC Files
From: mzenier@netcom.com (Mark Zenier)
Date: Sat, 20 Jan 1996 21:45:50 GMT
Links: << >>  << T >>  << A >>
in <4do6ve$t7f@portal.gmu.edu>, S.G. Wood, Jr. wrote:
: I'm searching for a tool (program) that will take a PLD JDEC file and
: reproduce the logic equations for the PLD.  We are currently trying to debug
: circuit cards with PLDs on them, however the documentation for the PLDs are
: out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that
: we are using are 22V10.  I would appreciate any pointers that people may have.

Opal Jr, National Semi's free or cheap Pal assembler package included
the JED2EQN program that would revert a 22V10.  (If you were lucky
enough to get a chip not copy protected.)  Should be a common part of
any recent PAL assembler by the competing manufacturers.

I'm not sure of any current net location, and National has dropped their
GAL/PAL line, so their sales rep may no longer carry it.  You can buy
the software disks for a GAL programmer from the hobby magazine Elektor
Electronics that includes OPAL Jr. 

Here are a few URL's that may or may not still be valid for PLD software.
ftp://ftp.funet.fi/pub/cae.  
ftp://ftp.cs.indiana.edu/pub/goo/PLD/
http://www.cs.ucla.edu/csd-grads-gs3/loving/www/m2l.html

Mark Zenier  mzenier@eskimo.com  mzenier@netcom.com





Article: 2660
Subject: Re: PLD JDEC Files
From: Larry Martin <larrym@wco.com>
Date: 21 Jan 1996 01:24:22 GMT
Links: << >>  << T >>  << A >>
swood@melpar.esys.com (S.G. Wood, Jr.) wrote:
>I'm searching for a tool (program) that will take a PLD JDEC file and 
>reproduce the logic equations for the PLD.  We are currently trying to debug 
>circuit cards with PLDs on them, however the documentation for the PLDs are 
>out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that 
>we are using are 22V10.  I would appreciate any pointers that people may have.
>
>-Simon

I believe that ABEL by Data-IO can do this.  Also the PLDshell that Intel 
used to sell (now part of Altera) could do it (I think).  You could also 
check with Logical Devices, who makes CUPL to see if they can do this.

Good luck-
Larry

-- 
 _         __      __      _____
| |       |  \    /  |    / ___ \        Larry Martin Consulting
| |       | \ \  / / |   / /   \_\       Analog - RF - Microwave
| |       | |\ \/ /| |   | |              
| |       | | \  / | |   | |    __       Frequency Synthesizers
| |____   | |  \/  | |   \ \___/ /
|______|  |_|      |_|    \_____/        Voice:  707-829-0633




Article: 2661
Subject: how to write place and route software
From: edwint@bnr.ca (Edwin Tsang)
Date: 21 Jan 1996 16:33:48 GMT
Links: << >>  << T >>  << A >>
As a user of FPGA and EPLD devices. Very often I bumped into routing 
problem e.g. 
 1) not able to fit my design into my target chip even it is in the 80% 
   range. (vendor A)
 2) routed my design but 100s of ns late (vendor B)
 3) routed my design, but if I spend a day in routing it manually, I could
    get decrease the area used for 50% + the improvement in timming. 
    (vendor C)
 4) routed my design and then I fixed the pin they give me and they failed
    to route it again with the pin fixed. (vendor D + A)
 5) It crash during place and route.
Yes I know it is very HARD to design the place and route software. If you 
asked me to design one, I have no idea where to start. After optimized the
logic equation, how would one start place the logic?

Is there any introduction article on place and route software on the NET. 
So that FPGA/EPLD users like me could understand a bit more and 
appreciate the hard work our P&R designer do. 

May be it could help us to design a more routable design.  
Yes, it may affect the protability and ability to reuse the design. 
But more than often, for some parts of my designs I would use schematic
instead of VHDL to take advantage of the target device's architecture, 
though I wish I don't need to do that if the software could take care
of it.

--
Edwin Tsang, Email:edwint@bnr.ca , NORTEL
Opinion is mine only and I reserved the right to change it 


Article: 2662
Subject: Re: Virtual Computer Corp. still in business?
From: cburns@crl.com (Charlie Burns)
Date: 21 Jan 1996 12:35:24 -0800
Links: << >>  << T >>  << A >>
Yes, I attended a presentation by Steven about two weeks ago in
Mountain View. They have this SBUS card with a single Xilinx
chip that was pretty interesting. Here's the address:

Virtual Computer Corp
6925 Canby Ave #103
Reseda, Ca 91335
818 342-8294
info@vcc.com

Arrigo Benedetti (benedett@dsi.unimo.it) wrote:
: Does anyone know if Steven Casselman's Virtual Computer Corporation
: is still alive?


Article: 2663
Subject: Re: Emulation for a wireless chip
From: devb@lys.vnet.net (David Van den Bout)
Date: 21 Jan 1996 19:46:20 -0500
Links: << >>  << T >>  << A >>
In article <4ddfu1$4kd@news.csie.nctu.edu.tw>,
Chuang Hsien-Ho <eea80593@orioles.EE.NCTU.edu.tw> wrote:
>Chuang Hsien-Ho (eea80593@maddux.EE.NCTU.edu.tw) ´£¨ì:
>: We have a project, "A Baseband Chip Set for Digital Cellular Phone", in
>: progress. We will soon finish the verilog coding. Now we want to prepare
>: for the emulation step. We have about tens of thousands of gates, targeting
>: on at least 6MHz.
>: Being a academic project, we might not able to afford a complete commercial
>: system. We plan to buy some RAM-based FPGAs(like XC4000 or ALTERA FLEX series),
>: maybe some FPICs(like Aptix).
>: Could you experienced persons give me some comments or suggestions? Or where
>: can I get the detail information of these vendors?  Thanks a lot.
>
>Many thanks to those who gave me suggestion. It sounds that ALTERA will be
>better in rounting, but Xilinx will be easy for first user. I am really 
>interested in the average utilization of XC4000 and FLEX8000(FLEX10K).
>Does anyone has experience?
>
>However, no one told me about the FPICs. The only vendor I know is Aptix,
>but I don't know if they have agent in Taiwan. Could anyone tell me?
>I also heard that the FPIC is expensive. If we don't use FPIC, is there
>any alternatives for the emulation board?
>
Try the chips from I-Cube.  They are smaller than the Aptix chips
(160 I/O although I think they are coming out with a 320 I/O version)
but easier to work with.  They have a cross-bar architecture so they don't
need a complicated router.  I-Cube used to supply a simple C library you
could use to make your own downloading software.  I can personally vouch
that I-Cube's stuff worked very well (at least back in 1993).

-- 
|| Dave Van den Bout  --  XESS Corp. ||
|| 2608 Sweetgum Dr., Apex, NC 27502 ||
|| (919) 387-0076 FAX:(919) 387-1302 ||
|| devb@xess.com       devb@vnet.net ||


Article: 2664
Subject: Re: PLD JDEC Files
From: devb@lys.vnet.net (David Van den Bout)
Date: 21 Jan 1996 19:51:43 -0500
Links: << >>  << T >>  << A >>
>> I'm searching for a tool (program) that will take a PLD JDEC file and 
>> reproduce the logic equations for the PLD.  We are currently trying to debug 
>> circuit cards with PLDs on them, however the documentation for the PLDs are 
>> out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that 
>> we are using are 22V10.  I would appreciate any pointers that people may have.
>> 

ALTERA's free PLDSHELL tools include a disassembler that will take a JEDEC
file and return a PLDasm file for a particular part.  I believe one of the
parts it works for is the 22V10.  You can call ALTERA and get them to send
you the tools.  You can also download them from 
ftp://ftp.vnet.net/pub/users/xess/PLDSHELL/pldsh.zip (I think -- I can 
never get the path right).


-- 
|| Dave Van den Bout  --  XESS Corp. ||
|| 2608 Sweetgum Dr., Apex, NC 27502 ||
|| (919) 387-0076 FAX:(919) 387-1302 ||
|| devb@xess.com       devb@vnet.net ||


Article: 2665
Subject: Re: PLD JDEC Files
From: trev@ss11.wg.icl.co.uk (Trevor Hall)
Date: 22 Jan 1996 06:40:28 GMT
Links: << >>  << T >>  << A >>
In article <4do6ve$t7f@portal.gmu.edu>
           swood@melpar.esys.com "S.G. Wood, Jr." writes:

> I'm searching for a tool (program) that will take a PLD JDEC file and 
> reproduce the logic equations for the PLD.  We are currently trying to debug 
> circuit cards with PLDs on them, however the documentation for the PLDs are 
> out dated and incorrect (as can be seen with a logic analyzer).  The PLDs that 
> we are using are 22V10.  I would appreciate any pointers that people may have.
> 

Old versions of AMD's PALASM have a JEDEC "disassembler". The new (MINC based)
MACH XL for AMD devices does not have such a feature.
I believe ABEL from Data I/O has this feature.

T.H.




Article: 2666
Subject: How Big Chips Will Be Designed In The Not Too Distant Future
From: jcooley@world.std.com (John Cooley)
Date: Mon, 22 Jan 1996 07:16:21 GMT
Links: << >>  << T >>  << A >>
cnuddep@sh.bel.alcatel.be (peter cnudde sh146 8218) wrote:
> We all agree that the level of abstraction will increase, how else could
> we design 10M gate ASICs? 

Simon J Davidmann <simond@vchips.com> wrote:
>I am sure we won't be writing it in VHDL, what we shall be doing is doing
>a lot of design re-use and purchasing complete packaged blocks from 3rd
>parties - like our PCI, PCMCIA, USB, ATM cores - chip designers will be
>putting blocks together - with small bits of new design - not designing
>10M gates from scratch...
>
>Simon Davidmann
>Virtual Chips, Inc.

Although Simon works at a company that sells 3rd party designs, I think his
view on this is right on.  That is, the big designs of the future will
contain a hell of a lot of pre-packaged subdesigns that will be provided
by people/companies other than the 10M chip designer with the value add
of the 10M chip designer being in the significantly smaller custom glue
logic & subdesigns he/she creates.

                           - John Cooley
                             Part Time EDA Consumer Advocate
                             Full Time ASIC, FPGA & EDA Design Consultant

===========================================================================
 Trapped trying to figure out a Synopsys bug?  Want to hear how 3713 other
 users dealt with it ?  Then join the E-Mail Synopsys Users Group (ESNUG)!
 
      !!!     "It's not a BUG,               jcooley@world.std.com
     /o o\  /  it's a FEATURE!"                 (508) 429-4357
    (  >  )
     \ - /     - John Cooley, EDA & ASIC Design Consultant in Synopsys,
     _] [_         Verilog, VHDL and numerous Design Methodologies.

     Holliston Poor Farm, P.O. Box 6222, Holliston, MA  01746-6222
   Legal Disclaimer: "As always, anything said here is only opinion."


Article: 2667
Subject: Chosing VHDL or Verilog Does Have An Impact For U.S. Engineers
From: jcooley@world.std.com (John Cooley)
Date: Mon, 22 Jan 1996 07:31:56 GMT
Links: << >>  << T >>  << A >>
Tony Goodloe <tgoodloe@adtran.com> wrote:
>Verilog vs. VHDL - IT DOESN'T MATTER! People have shipped products 
>and made money using each. Learn one. Don't fret about the decision.
>The hard part is understanding what HDLs are all about in general. 
>Once you learn one, the other will come.

Sorry, Tony, but I have to disagree with you on this one here.  Unlike Europe
or in Japan, U.S. engineers are basically hired by the buzzwords they have
on their resume *and* how well the engineers know them.  Most U.S. companies
don't like to train if they can get a person with the right buzzwords in the
first place.  In light of the last 10 years, most U.S. workers know they're
disposable at any moment -- so it pays to keep your career such that the 
right buzzwords are on your resume at all times.

                           - John Cooley
                             Part Time EDA Consumer Advocate
                             Full Time ASIC, FPGA & EDA Design Consultant

===========================================================================
 Trapped trying to figure out a Synopsys bug?  Want to hear how 3713 other
 users dealt with it ?  Then join the E-Mail Synopsys Users Group (ESNUG)!
 
      !!!     "It's not a BUG,               jcooley@world.std.com
     /o o\  /  it's a FEATURE!"                 (508) 429-4357
    (  >  )
     \ - /     - John Cooley, EDA & ASIC Design Consultant in Synopsys,
     _] [_         Verilog, VHDL and numerous Design Methodologies.

     Holliston Poor Farm, P.O. Box 6222, Holliston, MA  01746-6222
   Legal Disclaimer: "As always, anything said here is only opinion."


Article: 2668
Subject: Re: PLD JDEC Files
From: David Pashley <david@fpga.demon.co.uk>
Date: Mon, 22 Jan 96 09:51:48 GMT
Links: << >>  << T >>  << A >>
As well as giving you the PALASM equations, PLDLab95 from iNt (in 
Munich) also generates state-machine descriptions and schematics 
from JEDEC files.

-- 
David Pashley                 <
 ------------------------  <  <  <  ---------- Email: david@fpga.demon.co.uk
| Direct Insight Ltd    <  <  <  <  >            Tel: +44 1280 700262      |
| *The EDA Source*         <  <  <               Fax: +44 1280 700577      |
 ---------------------------  <  ------------------------------------------


Article: 2669
Subject: Re: How Big Chips Will Be Designed In The Not Too Distant Future
From: Brian Childs <brian@vizef.demon.co.uk>
Date: Mon, 22 Jan 1996 11:54:44 +0000
Links: << >>  << T >>  << A >>
In article <DLKMv9.Is5@world.std.com>, John Cooley
<jcooley@world.std.com> writes
>cnuddep@sh.bel.alcatel.be (peter cnudde sh146 8218) wrote:
>> We all agree that the level of abstraction will increase, how else could
>> we design 10M gate ASICs? 
>
>Simon J Davidmann <simond@vchips.com> wrote:
>>I am sure we won't be writing it in VHDL, what we shall be doing is doing
>>a lot of design re-use and purchasing complete packaged blocks from 3rd
>>parties - like our PCI, PCMCIA, USB, ATM cores - chip designers will be
>>putting blocks together - with small bits of new design - not designing
>>10M gates from scratch...
>>
>>Simon Davidmann
>>Virtual Chips, Inc.
>
>Although Simon works at a company that sells 3rd party designs, I think his
>view on this is right on.  That is, the big designs of the future will
>contain a hell of a lot of pre-packaged subdesigns that will be provided
>by people/companies other than the 10M chip designer with the value add
>of the 10M chip designer being in the significantly smaller custom glue
>logic & subdesigns he/she creates.
>
>                           - John Cooley
>                             Part Time EDA Consumer Advocate
>                             Full Time ASIC, FPGA & EDA Design Consultant
>

I also agree with this trend. In fact it is already here... You only
have to look at the companies who used to (and probably still do)
provide custom design services, who now publish/market an extensive
range of 'off the shelf cores'. What appears to be missing is any good
design tools to assist designers (or do you know different). 

I see the challenge for the EDA tool companies is to provide tools which
allow designers to 'manage' this process.

Some parallels have been drawn, in the past, with PCB design techniques
i.e. the value added is how you put these blocks together. Whilst some
aspects of PCB design are relevant, I think there is one fundimental
difference which will have a major impact on the style of design tools
required. In the PCB world when you use, say a pentium chip, you can
only use that chip they way it was designed to work i.e. you can't
change the gate implementation. With embeded cores you can and typically
get synthisable HDL which is also user configurable. This means that
each gate level implementation is going to be different. It will have
different timing characteristics, implementation etc. The factors which
will influence the implementation are not only the configuration
parameters you set as a designer, but the synthsis tool used, its
version (different versions of the same tool typically build gates
differently) but also the user specified logic surrounding the core. 

This leads me to the question of how do you evaluate one companies core
against anothers, without doing a full design? At the time you would
evaluate a core, typically very early in the design cycle, the factors
which would influence the suitablity of one core over another are not
known. Is this a silicon definition of 'Catch 22'?

What I think is needed is effectively on line data books (similar to
what PCB designers have) which provide you with not only the functional
data, but 'generic' timing data and simulatable code for functional
suitabliity verification. The generic timing data would be used with a
black boxed level of hierarchy, so that the synthesis tool has something
reasonable to work on and then depending on wether you can meet these
timing constraints would dictate your next step. 

Whilst I don't see any alternative to using cores, if you want chips >
10m gates, new EDA tools are going to be required to help manage this
subtly different design approach. I have not seen any tools out there
today that come close to providing this sort of functionality. If you
know different please do let me know.


PS. If this mail appears to come from Brian Childs, sorry, this is not
correct. It should come from Steve Bird steve@vizef.demon.co.uk. We have
installed some new mailer s/w and it is proving to be difficult to
configure correctly....

-- 
Steve Bird steve@vizef.demon.co.uk

Opinions are mine, mine, mine...


Article: 2670
Subject: Re: Chosing VHDL or Verilog Does Have An Impact For U.S. Engineers
From: edwint@bnr.ca (Edwin Tsang)
Date: 22 Jan 1996 14:39:39 GMT
Links: << >>  << T >>  << A >>
John Cooley (jcooley@world.std.com) wrote:

: Sorry, Tony, but I have to disagree with you on this one here.  Unlike Europe
: or in Japan, U.S. engineers are basically hired by the buzzwords they have
: on their resume *and* how well the engineers know them.  Most U.S. companies
: don't like to train if they can get a person with the right buzzwords in the
: first place.  In light of the last 10 years, most U.S. workers know they're
: disposable at any moment -- so it pays to keep your career such that the 
: right buzzwords are on your resume at all times.

John, you are right on. What are those buzzwords (for FPGA/ASIC designer):
 VHDL, Verilog, ASIC, FPGA, Synopsys, Viewlogic, PAL/GAL, EPLD, CPLD,
 simulation, synthesis, Vendors(Xilinix, Lattice, Altera, ATEL, ATMEL ....)
 HDL, ABEL
Anybody want to add to this list?

--
Edwin Tsang, Email:edwint@bnr.ca , NORTEL
Opinion is mine only and I reserved the right to change it 


Article: 2671
Subject: Re: How Big Chips Will Be Designed In The Not Too Distant Future
From: spp@bob.EECS.Berkeley.EDU (Steve Pope)
Date: 22 Jan 1996 16:50:16 GMT
Links: << >>  << T >>  << A >>
Brian Childs <brian@vizef.demon.co.uk> writes:

> I also agree with this trend. In fact it is already here... You only
> have to look at the companies who used to (and probably still do)
> provide custom design services, who now publish/market an extensive
> range of 'off the shelf cores'. 

This is because whenever a design shop designs a chip for one
customer, its elements then become potential "off the shelf cores" for
the next customer.    Problem is, the actual level of circuit
re-use is always lower than one hopes.   Rule of thumb -- if you
can get by with less than half of a design being newly-designed
circuitry, you're doing really well.

If the design shop is accumulating large numbers of "off the shelf
cores" it's a good bet they are desinging these things anew for each
customer and that is why they have so many of them.

> What appears to be missing is any good
> design tools to assist designers (or do you know different). 

We've been saying this for the last 15 years. :)

Steve


Article: 2672
Subject: Re: Virtual Computer Corp. still in business?
From: mark.stephens@gsfc.nasa.gov (Mark Stephens)
Date: Mon, 22 Jan 1996 16:11:16 -0500
Links: << >>  << T >>  << A >>
In article <BENEDETT.96Jan19105901@iago.dsi.unimo.it>,
benedett@dsi.unimo.it (Arrigo Benedetti) wrote:

>Does anyone know if Steven Casselman's Virtual Computer Corporation
>is still alive?
>
>Thanks in advance,
>
>-Arrigo Benedetti

Try 

Virtual Computer Corporation
6925 Canby Ave #103
Reseda, CA
teL: 818-342-8294 fax: 818-342-0240
John Schewel, VP Marketing & Sales
jas@vcc.com


mark

-- 
mark stephens                                    "In constraint,
NASA GSFC Code 521                                is freedom"
Greenbelt, MD
(301) 286-4269         mark.stephens@gsfc.nasa.gov


Article: 2673
Subject: Re: good interview questions ?
From: mbutts@netcom.com (Mike Butts)
Date: Mon, 22 Jan 1996 22:29:00 GMT
Links: << >>  << T >>  << A >>
Very interesting.  I had to change 'input' and 'output' because
Cadence Verilog-XL considers them reserved keywords.  And I added
a $monitor statement to see what went on.  So I got:

> Highest level modules:
> test
> 
>                    0 inputa = 0, outputa = 0
>                   10 inputa = 1, outputa = 0
>                   11 inputa = 0, outputa = 0
>                   20 inputa = 0, outputa = 1
> 19 simulation events
> CPU time: 0.3 secs to compile + 0.2 secs to link + 0.0 secs in simulation
> End of VERILOG-XL 2.2.1   Jan 22, 1996  14:11:05

Is this result because once the output = 1 is scheduled
at time 20, the output = 0 which would occur at time 16 isn't
scheduled because there's already a later event scheduled?

In fact gate-level event-driven simulators I've used in the
past used an "inertial delay" model which would basically
swallow input pulses that were shorter than the propagation
delay of the gate.  Since that's roughly what the real
gate would really do.  Such a simulator would never pass a 1
to the output with that stimulus.

           --Mike


-- 
Mike Butts, Portland, Oregon   mbutts@netcom.com



Article: 2674
Subject: XILINX XACT 6.0.0 Tools flaky
From: gavin@cypher.co.nz (Gavin Melville)
Date: Tue, 23 Jan 1996 04:40:11 GMT
Links: << >>  << T >>  << A >>
Hi All,

I am using the windows based XACT (step) tools, and have found them to
be a little "flaky", with occasional lockups and crashes -- in
particular the Design Editor (V5.2.0).   This version has run fine
from DOS, but the need for simulation has forced me to use the Windows
tools.   One particular problem -- XDE doesn't always seem to generate
the same bitstream each time DebugLoad is run.

The machine has been OK in the past, and I have tried both QEMM and
HIMEM, with no observable difference.   The problem occurs 1-5 times
per day.

Has anyone else, in particular anyone who depends on these tools found
Ver 6.0.0 to be flaky ?


--
Gavin Melville,
gavin@cypher.co.nz





Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search