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
2017JanFebMarApr2017

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 450

Article: 450
Subject: Re: Which PC HDL synthesis tool is best?
From: "Charles Michael Heard heard@btr.com" <heard@btr.btr.com>
Date: 20 Nov 1994 03:31:10 GMT
Links: << >>  << T >>  << A >>
In article <1994Nov18.031530.22623@afterlife.ncsc.mil>,
Chris Tscharner <cgtscha@afterlife.ncsc.mil> wrote:
>>
>>   2) Which HDL to use: Verilog HDL or VHDL?
>>
> I would go with Verilog.  You can then stay with one language from RTL 
> all the way to post-route structural code containing delay information.
> I believe with VHDL you are probably stuck with going to a proprietary 
> simulator for the latter at the present time (ex: Viewsim in the case of
> Viewlogic).  In addition, a unified language enables true mixed level
> simulation (some parts of the design hierarchy described as RTL code, 
> some as schematics, possibly some as post-routed structural code).
> Both the Xilinx toolset and the Altera toolset have the capability of
> generating post-route Verilog.  Neocad now has the same capability for
> the FPGAs they support with their toolset.

As I understand it VHDL _is_ capable of mixed-level description (e.g., some
behavioural, some RTL, some purely structural).  In other posts to this
newsgroup I have seen several claims that Verilog-to-VHDL translators exist
-- which would seem to imply that VHDL is at least as powerful as Verilog.
Translating in the other direction is said to be difficult owing to VHDL
constructs which have no analog in Verilog.

I will readily grant that I've seen no tool set which generates post-route
VHDL for Xilinx or any other FPGA.  What I'd really like is a synthesis
tool which could show what kind of hardware the high-level constructs were
actually mapping to -- similar to a compiler giving a mixed source/assembly
output.  Chris, is this what you are referring to by post-route Verilog?
Or is it something more limited -- such as translation of an LCA file into
a netlist description without correlating it with your original source code?

Mike
-- 
C. M. Heard
VVNET, Inc.                           phone:  +1 408 247 9376
4040 Moorpark Ave. Suite 206          fax:    +1 408 244 3651
San Jose, CA 95117 USA                e-mail: heard@btr.com


Article: 451
Subject: Re: any XC4000 Horror Stories?
From: "Charles Michael Heard heard@btr.com" <heard@btr.btr.com>
Date: 21 Nov 1994 02:27:52 GMT
Links: << >>  << T >>  << A >>
In article <3aimv2$845@mailman.etecnw.com>,
Richard George <rkg@etecnw.com> wrote:
>I'm evaluating some FPGAs for use in a project, and I'm looking for
>"horror stories" having to do with the xilinx XC4000 series parts.

Actually this is not really a horror story so much as a success
story.  The project in question (at a place where I'm no longer
employed) used XC4005-6 devices as full-speed peripherals on a
25 Mhz 80386-DX local bus.  The one XC4005 problem we had on this
project was owing to a bad lot of parts:  devices with one specific
date code (but no others) would 'glitch' the 80386 data bus just
as they transitioned from configuration download mode to user mode.
This generally caused the 80386 processor to GP fault.  Once we swapped
out the offending lot of parts we had no further problems.  In all other
respects we were VERY happy with the 4005s.  Probably the conservative
design practices of the lead engineer had something to do with this.

> I've heard some nasty rumors of difficulties with ground bounce, timing,
> etc., [...]

I've not heard of this ... do you know where these rumours originated?

G'Day,

Mike
--
C. M. Heard
VVNET, Inc.                           phone:  +1 408 247 9376
4040 Moorpark Ave. Suite 206          fax:    +1 408 244 3651
San Jose, CA 95117 USA                e-mail: heard@btr.com


Article: 452
Subject: RE: Looking for VHDL & VIEWLOGIC FPGA Experts/Consult
From: randraka@ids.net
Date: Mon, 21 Nov 94 09:30:27 GMT
Links: << >>  << T >>  << A >>
In Article <3afvn5$969@steel.interlog.com>
"David I.Roach" <daver@interlog.com> writes:
>Looking for experts/consultants for VHDL & VIEWLOGIC FPGA
>
The Andraka Consulting Group is a digital hardware design firm 
specializing in getting maximum performance from FPGAs.   Our 
services include complete design, development, simulation and
integration of these devices.  We also evaluate, troubleshoot and
improve existing designs.   Please call or send E-mail for a free 
brochure.
-Ray Andraka, Group Chairman
Andraka Consulting Group
Tel  401/884-7930,  Fax 401/884-7950
Email randraka@ids.net
>


Article: 453
Subject: Re: any XC4000 Horror Stories?
From: dh@fncrd7.fnal.gov (don husby)
Date: 21 Nov 1994 14:47:54 GMT
Links: << >>  << T >>  << A >>
rkg@etecnw.com writes:
> Hi,
>    I'm evaluating some FPGAs for use in a project, and I'm looking for
> "horror stories" having to do with the xilinx XC4000 series parts. 
> I've heard some nasty rumors of difficulties with ground bounce,
> timing, etc., and I was looking for some firsthand accounts of
> problems with these beasties.

  I had heard that the new Xilinx chips were more powerful than ever 
and I decided to put that rumor to the test.  It was a cold and rainy
night and I was working late down in the pits of the colliding
detector.  I could almost feel the anti-proton flux in the air (but
of course I knew I was safe since the place is always swarming with
safety geeks.)  I set my new X4010 chips into the circuit board, 
attached the X-Cehcker a started a download.  About 10 seconds into
the download there was a searing lightning strike.
  There was a crash of thunder followed by the whoosh of helium as the
superconducting magnets lost ther containment.  I had been lucky: the 
anti proton beam just missed me.  "Whew, I could have been killed", I 
thought. 
   "Better luck next time" said high squeaky voice that seemed to come
from my test equipment.  I looked for the source of the voice but 
couldn't find it.  I moved stuff around and when I reached to unplug
the smoking mess that was my Xilinx test board, I was knocked on my 
ass by a bolt of electricity.  "Strike two", the little voice said...

--
Ground bounce?


Article: 454
Subject: Re: Help: Seeking Your Opinion of EDN Article
From: jww@cadence.com (John Willoughby)
Date: Mon, 21 Nov 1994 21:50:33 GMT
Links: << >>  << T >>  << A >>
John, I changed addresses recently and missed that issue (alas!). I'll try
and find it somewhere around here.

I think benchmarks are good things, and why wouldn't I? Cadence does very
well in benchmarks (like Cadence's Leapfrog being the fastest VHDL simulator
on the market). As a recently former user and now a tool vendor employee,
however, I tend to take a wary approach. It's too easy to focus on
specific tasks in a design cycle and lose sight of the overall design
process. As a user, I want to know what it's going to take to get my design
done, not just how a tool functions on one specific task. The use of more
complex tools, such as synthesis tools, means that there is a very real
learning curve for the tool, and many user-interface issues that can
have a dramatic impact on total usability. Cadence's Synergy synthesis tool,
for example, meets or beats the competition in many raw benchmarks. It's real
appeal to many of it's users, however, is it's ease of use, integration and
pricing.

A Learjet is a great way to get to work, but how long does it take to learn
how to drive it? How much does it cost? Not to mention the problem of parking!

So- I think benchmarks are good, in general, but authors need to make sure
not to obscure the impact on the total design methodology by too narrow
a focus on individual tasks.


standrd disclaimer:
my opinions are my own,etc., etc., etc.


John Willoughby


Article: 455
Subject: Re: any XC4000 Horror Stories?
From: nelson@eng.auburn.edu (Victor P. Nelson)
Date: Tue, 22 Nov 1994 23:17:18 GMT
Links: << >>  << T >>  << A >>
> Hi,
> 
> I'm evaluating some FPGAs for use in a project, and I'm looking for "horror stories"
> having to do with the xilinx XC4000 series parts.  I've heard some nasty rumors of
> difficulties with ground bounce, timing, etc., and I was looking for some firsthand
> accounts of problems with these beasties.
> 
> Thanks in advance,
> 
> rkg
> 
> (Richard George)
> 
> 


One of my senior design student groups managed to blow a 1 cm diameter
chunk out of the ceramic package above the die in an XC4003 - does that
qualify as a "horror story"?  Other than that, the XC4003 and 4005 devices
have done quite well in student projects.

Vic Nelson
Auburn University



Article: 456
Subject: Converts Fortran or C to VHDL
From: chu@bnlku4.phy.bnl.gov (Chu Ng)
Date: Wed, 23 Nov 1994 14:46:22 GMT
Links: << >>  << T >>  << A >>
Does anybody know any software package that converts
Fortran (or C) to VHDL? 

If this is not the right user group to post this question,
can you sugguest me which user group I should post this?

Chu



Article: 457
Subject: Re: any XC4000 Horror Stories?
From: cshelor@cpdsc.com (Charles Shelor)
Date: 23 Nov 1994 16:54:16 GMT
Links: << >>  << T >>  << A >>

>Hi,
>
>I'm evaluating some FPGAs for use in a project, and I'm looking for "horror stories"
>having to do with the xilinx XC4000 series parts.  I've heard some nasty rumors of
>difficulties with ground bounce, timing, etc., and I was looking for some firsthand
>accounts of problems with these beasties.
>
>Thanks in advance,
>
>rkg
>
>(Richard George)
>
>

HORROR 1: _THE_TOOLS_
My project started with 4008 using PC tools.  Went to 4010, had to
upgrade to 16 MBytes.  Went to 4013, had to upgrade to 32 MBytes.  PPR
ran all weekend without a solution.  Eventually got NeoCad on the SUN
which generated 65% utilization of the 4013 in a couple of hours.

HORROR 2: _XILINX_NON_SUPPORT_
Xilinx won't tell you about known problems until you rediscover them
yourself.  Example 1:  The transparent/latch sense for the gates of the
I/O latches are reversed from the XC3000s and incorrect in the 1st AND
2nd editions of their data books.  Example 2:  First release of PPR for
4013 would not 'honor' pin locations and moved pins freely trying to
improve routing.  (Only a problem if you are updating a design that is
already in a PCB!)

Corner your local Xilinx AE and threaten severe mental and physical
anguish if he doesn't tell everything he knows about the part!

HORROR 3:  Pullup resistors sometimes 'wear out'.  We had several parts
where the pullup resistors got weaker and weaker until we had to add
external pullups.

HORROR 4:  Optimistic timing:  Preliminary reviews by Xilinx and company
personnel using Xilinx databooks indicated typical operation at 30 MHz
with full temp operation at 24 MHz.  We were able to get as high as
20 MHz in the lab with 16 MHz being normal.  We never achieved greater
than 12 MHz through the temperature range.

We did not experience ground bounce problems in a 32-bit bidirectional
microprocessor bus interface.

Good Luck,

Charles F. Shelor

SHELOR ENGINEERING              VHDL Training, Consulting, and models
3308 Hollow Creek Rd            (817) 467-9367
Arlington,  TX  76017-5346      cfshelor@acm.org

Rule 1:  Never sacrifice global efficiency during local optimization.
Rule 2:  Every interaction between two organizations is a
         marketing opportunity; make it a positive one!







Article: 458
Subject: Horror Story
From: sc@vcc.com (Steve Casselman)
Date: Wed, 23 Nov 1994 17:29:13 GMT
Links: << >>  << T >>  << A >>
> Hi,
> 
> I'm evaluating some FPGAs for use in a project, and I'm looking for "horror stories"
> having to do with the xilinx XC4000 series parts.  I've heard some nasty rumors of
> difficulties with ground bounce, timing, etc., and I was looking for some firsthand
> accounts of problems with these beasties.
> 
> Thanks in advance,
> 
> rkg
> 
> (Richard George)
> 

All products horror stories let me know when you find the
*perfect* product. One problem I had with the XC4010 was
with the prototype of our EVC. I had hooked a pin to a open-collector
buffer that drives a *level* sensitive interupt on the SUN. Every
time I loaded the device I would get an interupt. I fixed this 
by routing a pin the the o-c bufs enable now everything works
fine. The interupt line is the only asyncronous line on the
SBus and very touchy. So the bottom line is I should have
been more careful.

Steve Casselman


Article: 459
Subject: Re: any XC4000 Horror Stories?
From: knickels@bird.cs.uiuc.edu (Kevin Nickels)
Date: 23 Nov 1994 17:29:27 GMT
Links: << >>  << T >>  << A >>
In article <3avs3oINNoi0@sun004.cpdsc.COM>, cshelor@cpdsc.com (Charles Shelor) writes:

# 2nd editions of their data books.  Example 2:  First release of PPR for
# 4013 would not 'honor' pin locations and moved pins freely trying to
# improve routing.  (Only a problem if you are updating a design that is
# already in a PCB!)
The current version (5.0) of the XACT/Mentor tools have this error unless the 
 LOC property is assigned on the top level schematic.  They had it fixed, then
 broke it again in this version.  Makes your top-level schematic design
 really ugly.  Aargh!

I've also had problems with the tools removing dead circuitry
 that...wasn't.  I think this is due to the same problem as the above
 comment, though (certain properties don't propagate from level to level,
 LOC and INIT=gnd being two examples.)

# HORROR 4:  Optimistic timing:  Preliminary reviews by Xilinx and company
# personnel using Xilinx databooks indicated typical operation at 30 MHz
# with full temp operation at 24 MHz.  We were able to get as high as
# 20 MHz in the lab with 16 MHz being normal.  We never achieved greater
# than 12 MHz through the temperature range.
Hmmm.  I run state machines off a 32MHz clock in a 4005 at room
 temperature.  I'd have to look to see what the data book claims for this
 chip, though.

I also have intermittent problems with the xchecker software not
 recognizing the xchecker cable unless I remove the cable header from the 
 board and replace it WITH THE BOARD POWERED UP.  It just tells me that the
 xchecker cable is not correctly connected to the system.  I break/replace
 the connection between the xchecker header and the board and it works
 fine.  This happens both on the xilinx demo boards and a wire-wrapped
 microprocessor interface board we built.

--kevin nickels
-- 
| Kevin Nickels - knickels@uiuc.edu                                     |
| Home:                 210 Kenwood #337        Champaign, IL  61821    |
| Office: 217-244-1372  2414 Beckman Institute  Urbana,    IL  61801    |
| Lab:    217-333-9591  268 Everitt Lab         Urbana,    IL  61801    |

The scientist is elated by being confused. 
   -Dudley Hershbach, Harvard University (Nobel Prize, Chemistry, 1986) 


Article: 460
Subject: Re: Help: Seeking Your Opinion of EDN Article
From: onmate@iohk.com (onmate)
Date: 24 Nov 1994 11:38:27 GMT
Links: << >>  << T >>  << A >>
In article <Czn0oB.E5t@Cadence.COM>, jww@cadence.com (John Willoughby) says:
>
>So- I think benchmarks are good, in general, but authors need to make sure
>not to obscure the impact on the total design methodology by too narrow
>a focus on individual tasks.
>

I can't agree more. The point is quite simple, the raw performance can be boost
in some way (Better Hardware, More RAM, etc) However, time cost by bad/stupid
user interface can never be reduced.

Best Regards.

Frankie Chung


Article: 461
Subject: QuickLogic Design V5.0 (QuickWorks)
From: onmate@iohk.com (onmate)
Date: 24 Nov 1994 11:42:37 GMT
Links: << >>  << T >>  << A >>
Have anyone used the new QuickLogic Design Tools ? It claim to have very
good features: VHDL/Verilog Simulation & Synthesis. Does it really work ?

Best Regards

Frankie Chung 
Onmate Technology




Article: 462
Subject: Should I jump to Actel when using Synopsys/Altera?
From: toledo@gtex02 (Juan Toledo Cota)
Date: 24 Nov 1994 16:16:59 GMT
Links: << >>  << T >>  << A >>
        We've been working with a project on Altera-PC for a long
time. When compiling, it comes out that only with soft-buffering option On
it will finish the job, and it uses up too much logic. We've been thinking
about translating the .tdf to .vhd for use with Synonsys in a Sun, to see
wether the logic optimization is better, and not have to use soft
buffering.

	Is it worth the trouble or should we just jump directly to Actel?

-------------------------------------------------------------------------
Juan Toledo Cota                      ESCUELA SUPERIOR DE INGENIEROS
c/ Santa Maria de los Reyes, 8.       DE SEVILLA.
41008 Sevilla, SPAIN.                 toledo@gtex02.us.es
-- 
-------------------------------------------------------------------------
Juan Toledo Cota                      ESCUELA SUPERIOR DE INGENIEROS
c/ Santa Maria de los Reyes, 8.       DE SEVILLA.
41008 Sevilla, SPAIN.                 toledo@gtex02.us.es


Article: 463
Subject: WWW page: Asynchronous/Self-timed FPGA systems
From: Rob Payne <rep@dcs.ed.ac.uk>
Date: Thu, 24 Nov 1994 16:44:52 GMT
Links: << >>  << T >>  << A >>

As part of my thesis research, I have been constructing a WWW page on
Asynchronous/Self-timed FPGA systems research. The page contains links
to most of the Web sources on the topic, a self-timed FPGA
bibliography, and some background to my own research on a dedicated
self-timed bundled-data FPGA architecture.

The page's URL is:

http://www.dcs.ed.ac.uk/students/pg/rep/selfTimedFPGA/selfTimedFPGA.html

Any suggestions for improvements or additions to the page are welcome.

Rob Payne.

 -------------------------------------------------------
|Rob Payne                              rep@dcs.ed.ac.uk|
|Dept. of Computer Science, University of Edinburgh, UK.|
|tel: (+44) 131 650 5169                                |
 -------------------------------------------------------






Article: 464
Subject: Re: Should I jump to Actel when using Synopsys/Altera?
From: bjb@bei.pic.net (Billy J. Beckworth)
Date: Fri, 25 Nov 1994 12:55:25
Links: << >>  << T >>  << A >>
In article <3b2e9r$rri@obelix.cica.es> toledo@gtex02 (Juan Toledo Cota) writes:
>From: toledo@gtex02 (Juan Toledo Cota)
>Subject: Should I jump to Actel when using Synopsys/Altera?
>Date: 24 Nov 1994 16:16:59 GMT

>        We've been working with a project on Altera-PC for a long
>time. When compiling, it comes out that only with soft-buffering option On
>it will finish the job, and it uses up too much logic. We've been thinking
>about translating the .tdf to .vhd for use with Synonsys in a Sun, to see
>wether the logic optimization is better, and not have to use soft
>buffering.

>        Is it worth the trouble or should we just jump directly to Actel?

>-------------------------------------------------------------------------
>Juan Toledo Cota                      ESCUELA SUPERIOR DE INGENIEROS
>c/ Santa Maria de los Reyes, 8.       DE SEVILLA.
>41008 Sevilla, SPAIN.                 toledo@gtex02.us.es

Not having used anything but Actel, I can't tell you what would be best.  I 
can say that I have used Actel for years on several designs with very good 
results.  Design turn-around time has been very good and I have always 
achieved full routing with only minor tweaking.

regards, BJB

===================================================================
Billy J. Beckworth                                  bjb@bei.pic.net


Article: 465
Subject: FPGAs not CPLDs was Re: any XC4000 Horror Stories?
From: cgtscha@afterlife.ncsc.mil (Christopher G. Tscharner)
Date: Fri, 25 Nov 1994 23:23:35 GMT
Links: << >>  << T >>  << A >>
References: <3aimv2$845@mailman.etecnw.com> <3avs3oINNoi0@sun004.cpdsc.COM>
Sender: 
Followup-To: 
Distribution: world
Organization: US Dept of Defense
Keywords: 

In article <3avs3oINNoi0@sun004.cpdsc.COM> cshelor@cpdsc.com writes:
SNIP
>
>HORROR 4:  Optimistic timing:  Preliminary reviews by Xilinx and company
>personnel using Xilinx databooks indicated typical operation at 30 MHz
>with full temp operation at 24 MHz.  We were able to get as high as
>20 MHz in the lab with 16 MHz being normal.  We never achieved greater
>than 12 MHz through the temperature range.
>
Trying to predict timing strictly through use of databooks is dangerous
for any FPGAs IMHO.  The routing determines timing to a great extent,
and it is nondeterministic.  A great contributor is device utilization
percentage.  In addition, efficiency of the designer or
designer/follow-on-optimizer (schematics) or synthesizer/optimizer (HDLs)
can effect both levels of "atomic logic blocks" (ex CLBs in Xilinx) as
well as routability (ex: use of hard macros for floorplanning of complex
often used functions like adders).  

Chris Tscharner


Article: 466
Subject: Viewdraw's Change Label Sense
From: H.A.B.Tiggeler@ee.surrey.ac.uk (Hans Tiggeler)
Date: Sat, 26 Nov 1994 11:31:11
Links: << >>  << T >>  << A >>

Does anybody know how to preserve active low labels in Viewdraw-XACT5.0? 
As far as I understand the Change Label Sense command places a tilde in front 
of an active low label name. This tilde is then changed by WIR2XNF into a 
hyphen which has a special meaning in Viewsim. I have tried several ways of 
preserving the label sense by creating a CRS file during WIR2XNF execution. I 
tehn tried to restore the label sense during XNF2WIR ( -s -c file.crs). 
However, this was unsuccessful.

I find it hard to believe that you can not use active low label in your 
design. I have now changed all my active low label to "LABEL_L", this does 
however not improve readabily.

Regards,

Hans Tiggeler
University of Surrey


  


Article: 467
Subject: Re: Should I jump to Actel when using Synopsys/Altera?
From: kerry@altera.com (Kerry Veenstra) <72712.1243@CompuServe.COM>
Date: 26 Nov 1994 23:13:06 GMT
Links: << >>  << T >>  << A >>
Juan,

> >        We've been working with a project on Altera-PC for a long
> >time. When compiling, it comes out that only with soft-buffering option On
> >it will finish the job, and it uses up too much logic. We've been thinking
> >about translating the .tdf to .vhd for use with Synonsys in a Sun, to see
> >wether the logic optimization is better, and not have to use soft
> >buffering.
> 

I've forwarded your post to an Altera Applications Engineer.
She should be able to help you.  You probably will be asked
to e-mail a copy of your TDF to sos@altera.com once we are
back in the office on Monday.

Kerry Veenstra
kerry@altera.com


Article: 468
Subject: Re: Viewdraw's Change Label Sense
From: fliptron@netcom.com (Philip Freidin)
Date: Sun, 27 Nov 1994 09:37:21 GMT
Links: << >>  << T >>  << A >>
In article <H.A.B.Tiggeler.1.000B856B@ee.surrey.ac.uk> H.A.B.Tiggeler@ee.surrey.ac.uk (Hans Tiggeler) writes:
>
>Does anybody know how to preserve active low labels in Viewdraw-XACT5.0? 
>As far as I understand the Change Label Sense command places a tilde in front 
>of an active low label name. This tilde is then changed by WIR2XNF into a 
>hyphen which has a special meaning in Viewsim. I have tried several ways of 
>preserving the label sense by creating a CRS file during WIR2XNF execution. I 
>tehn tried to restore the label sense during XNF2WIR ( -s -c file.crs). 
>However, this was unsuccessful.
>
>I find it hard to believe that you can not use active low label in your 
>design. I have now changed all my active low label to "LABEL_L", this does 
>however not improve readabily.
>
>Regards,
>
>Hans Tiggeler
>University of Surrey
>

Unfortunately, I believe that the solution you are using (using label 
names that include sense) is the only thing that works reliably. Both the 
change_label_sense, and the symbol editor's change_pin_sense seem to 
cause problems either with wir2xnf or later on when backanotating timing. 
The problem seems to be related to resrictions in the xnf language. 
The tilde, minus, slash, and backslash all have special meaning somewhere 
in the whole tool chain, and the conversions that wir2xnf makes seem to 
just push the problems from one place to another.

I went through trying to make sense of the mess about 2 years ago, and 
was unable to find a reliable way of using the change sense commands. I 
now either remember which signals are active low, or call them things 
like write_enable_bar. Wordy, but it does work. I dont believe this has 
changed in any version of wir2xnf since I last wasted time trying this.

There are also problems with the vielogic parameter substitution logic in 
wir2xnf. Fortunately the only fancy use I had for this has since been 
solved much more nicely with the new RLOC system. Simple parameter
substitutions in viewlogic/wir2xnf still work fine. Basically, 
substitution works if the first character of the assignment is an '@' 
character. I tend to create macros with multiple FFs in them, and on each 
one I will have an attribute like 'LOC=@Q0LOC' and 'LOC=@Q1LOC'. On the
instatiated symbol, I will have attributes like '@Q0LOC=CLB_R2C3', and
'@Q1LOC=CLB_R4C5'. This works fine. Attributes like 'LOC=CLB_@Q0LOC', 
with the instantiated symbol having '@Q0LOC=R2C3' does not work because of 
the position of the '@'. The new RLOC is easier to use and is more 
sophisticated than the simple substitutions I used to use.

All The Best,
	Philip Freidin
	


Article: 469
Subject: --- Libraries free of charge ---
From: leh@itiv.etec.uni-karlsruhe.de (Gunther Lehmann)
Date: 28 Nov 1994 09:44:46 GMT
Links: << >>  << T >>  << A >>
Hi,

we're developping a students course which provides
an insight in VHDL synthesis and FPGA programming.
We're using the Synopsys FPGA compiler and have 
ordered the xilinx tools. Since the delivery of these
tools takes a long time (no, this is not Xilinx's fault)
and we like to start synthesizing now, I like to know,
if there is any possibility to get the Synopsys
libaries:

    xfpga_xxx.db
    xprim_xxx.db

fast (ftp, ... ?) and free of charge?

Thanks
Gunther Lehmann

e-mail: leh@itiv.etec.uni-karlsruhe.de

 



Article: 470
Subject: Re: Any one with i_Logix experience?
From: Terry Mayhugh <mayhugh@austin.asc.slb.com>
Date: 28 Nov 1994 21:28:26 GMT
Links: << >>  << T >>  << A >>
We are evaluating the i-Logix Statemate and ExpressV-HDL packages for
use in a large scale simulation project. Are there any current users
of either of these tools who would be willing to share their 
experiences before we make a purchasing decision. You may E-mail
me directly or phone (512) 331-3118  8:30-5:30 CST if you wish. Thanks.


Article: 471
Subject: Re: Should I jump to Actel when using Synopsys/Altera?
From: tjl@research.canon.oz.au (Tim Lindquist)
Date: Tue, 29 Nov 1994 06:40:12 GMT
Links: << >>  << T >>  << A >>
I have been using MAX7000 parts in my design. Although there have been
many problems with the tools, the situation has been improving. All my
design is in VHDL. We are running Mentor 8.2 and use AutoLogic for
synthesis. Most problems on MAX parts seem to come down to the fact
that muxes on the interconnect are partially populated, so if your
pins are locked (which is normally the real world scenario anyway)
then small design changes can mean your project doesn't fit. This
situation has improved in version 5 of MaxPlus2. If your project no
longer fits due to pin allocations, it now can find the minimal set of
changes to fix the problem.

I DONT THINK going to Actel will help your problem, unless the
synthesis tool is absolutely top-notch. We have another design here
which is presently using 4 x 1460s. Autologic synthesis result was
pretty depressing in terms of both utilisation and speed. The designer
ended up doing hand synthesis to get what he wanted. Again, the
initial design was in VHDL.

My impression from running through my own VHDL is that the Altera VHDL
option is reasonably good (better utilisation and much faster than
AutoLogic) and that place and route for FLEX devices leaves Xilinx and
Actel in the dust.

Regards

Tim Lindquist


Article: 472
Subject: Does a digital comparator use 9 Mcells in Altera's 5000?
From: altuser1@us.es (Altera User #1)
Date: 29 Nov 1994 09:35:08 GMT
Links: << >>  << T >>  << A >>
  
        We are having a little trouble here. Is it right that a digital
	comparator (74684 macro-function) uses up 9 macro-cells in
        a 5000 series device? We've tried every possible compiler option.
        Is it because of the optimiser or the device will just not get anymore?
 
        Thanks.
        
	




Article: 473
Subject: XC3090 PERFROMANCE...
From: SANJAYB@ECLX.PSU.EDU (Sanjay Balasubramanian)
Date: 29 Nov 1994 17:45:34 GMT
Links: << >>  << T >>  << A >>
Hi,
	We are currently using the XC3090PG175 to implement a circit which
demands a speed of 25MHz (40ns). However any attempt to reach a speed higher
than 12ns has proved futile with the XC3090. (We are only using 38 CLBs and
abot 70 I/O pads, 150 speed grade). Are there existing designs which run at 
higher speeds on the XC3090 and if so any pointers to increase the performance
will be greatly appreciated.

Please e-mail us at:
s1b@eclu.psu.edu
Center for Electronic Design, Communication and Computing
University Park.

Thanks in advance,
Sanjay B.

	


Article: 474
Subject: PLD's, FPGA's & VITAL
From: jcooley@world.std.com (John Cooley)
Date: Tue, 29 Nov 1994 18:37:39 GMT
Links: << >>  << T >>  << A >>

From: "Lenihan, William E III" <wlenihan@msmail4.hac.com>
To: John Cooley (ESNUG)

I have heard a little about "VITAL", the VHDL Initiative Toward ASIC
Libraries. My limited understanding of it is that it allows a designs'
gate-level implementation, with real timing, to be 'linked-back' to the
source VHDL code and its' simulator, to enhance verification integrity by
using one common stim/response methodology (i.e. - test bench). Do I have
that right?

(1) Do you know if VITAL's reference to "ASIC Libraries" covers PLD's as well
as the obvious Gate Array/Standard Cell technology?

(2) If not, do you know of any work toward a PLD-targeted effort comparable
to VITAL?

Bill Lenihan
Hughes Aircraft


----

 I'm posting this for Bill because he doesn't have access to the vhdl & fpga
newsgroups.  Please, feel free to post replies publicly (to benefit the user
community) and to "cc" him via e-mail (to benefit him personally.)

                              - John Cooley
                                the ESNUG guy
                                (and EDA & ASIC Design Consultant, too!)

===========================================================================
 Trapped trying to figure out a Synopsys bug?  Want to hear how 3114 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."




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
2017JanFebMarApr2017

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