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 8825

Article: 8825
Subject: 1998 DAC University Booth (First Call for Participation)
From: rauletta@erebor.cudenver.edu
Date: Thu, 29 Jan 1998 19:11:37 -0600
Links: << >>  << T >>  << A >>


                        ********************************
                        **** CALL FOR PARTICIPATION ****
                        ********************************

                           1998 DAC University Booth
                        http://erebor.cudenver.edu/sigda
                                     at the
                       35th Design Automation Conference
                                June 15-19, 1998
                                 Moscone Center
                               San Francisco, CA

          The  DAC  University  Booth  provides an opportunity for the
          university  community  to  demonstrate  EDA  tools,   design
          projects,   and  instructional  materials  at  the  ACM/IEEE
          sponsored Design Automation Conference (http://www.dac.com).

          You  do  not  have to have a paper at DAC to demonstrate EDA
          tools, design projects, and instructional materials  at  the
          DAC  University Booth, you simply need to attend DAC 1998 in
          San Francisco.

          The  University  Booth   also   provides   space   for   the
          presentation  of  EDA  vendor  literature  and  programs  of
          interest to the university community.

          The 1000 square foot ACM  SIGDA  University  Booth  provides
          booth  space,  poster  areas,  computers, printers, and a T1
          connection to the Internet for  participating  universities.
          (Please  visit  http://erebor.cudenver.edu/sigda for further
          information.)

          To sign up, simply send the  following  information  to  the
          email  address  below, or better yet sign up at our web site
          http://erebor.cudenver.edu/sigda.

                --------------------------------------------------

          1) Your Name (Primary Contact):

          2) E-Mail: (user@etc.etc.etc)

          3) University:

          4) Department:

          5) URL: (http://etc.etc.etc)

          6) Computer Needed:
          (Indicate if you will bring a machine.)

          7) Operating System:

          8) Brief Description of software, design project, or course





              --------------------------------------------------



-Rich Auletta
 1998 DAC University Booth Coordinator
 University of Colorado at Denver
 Electrical Engineering
 Email: rauletta@erebor.cudenver.edu
 URL: http://erebor.cudenver.edu/sigda
 Voice: 303-556-2357
 Fax: 303-556-2383

-------------------==== Posted via Deja News ====-----------------------
      http://www.dejanews.com/     Search, Read, Post to Usenet
Article: 8826
Subject: Introduction book on Verilog/VHDL
From: Xiao Wang <xiao@umr.edu>
Date: Fri, 30 Jan 1998 00:13:19 -0800
Links: << >>  << T >>  << A >>
I'm just beginning to study Verilog or VHDL.

Is there any difference between these two names? 

What will be the very suitable introductory book for this Verilog and/or
VHDL? Thanks.

Please email back: xiao@umr.edu
Article: 8827
Subject: Re: VHDL vs schematics
From: madarass@cats.ucsc.edu (Rita Madarassy)
Date: 30 Jan 1998 10:42:13 GMT
Links: << >>  << T >>  << A >>

Normally, you maybe right there is no difference between hdl and
schematic. But there is a huge difference when it really matters.
For example, you may need a very high performance 32 bit priority encoder.i

A 32  bit priority encoder will likely be a huge bottle neck in your
design if you this in vhdl. Synthesis will create this priority 
encoder using 4 input lookup tables and it will slow down your
design to a ridiculoys speed. 
On the other hand, in schematic I can customize the carry logic using
CY4 components in Xilinx and build a priority encoder that will perform
at about 10ns in a 4000XL-09. Not bad ahh? 
Doing this implementation in VHDL will take at least 20 ns if the routes
are decent and the synthesis tool is fairly good.

If you want to get your job done and meet deadlines vhdl is the
way to go. However, be prepared to replace some of your modules
by designs done in schematic when vhdl doesn't do it.

In article <6aqqra$3nl@borg.svpal.org>,
George Noten <garyk@svpal.svpal.org> wrote:
>Simon Ramirez (s_ramirez@email.msn.com) wrote:
> I am not sure there is a big difference.  You *can* do in VHDL absolutely
> the same design that you do in schematic (i.e. write a netlist).  The
> only difference is that VHDL will be more portable and easier to
> simulate.  OTOH, you will need a schematic to understand what you wrote
> in VHDL.
>
>
> So-called "behavioral" VHDL description comes down to 2-3 sequential
> statements that the synthesis tool is able to understand (and each
> vendor interprets them in a different way).  No big deal.  Sometimes
> I do use behavioral description, but usually I prefer netlist and para-
> metrized libraries (XBLOX or LPM).  Normally there is no difference in
> performance between this type of design and schematic.
>
> 	George.
>


Article: 8828
Subject: Re: ABEL to Altera-HDL? Group FAQ?
From: "D.H. Chung" <dhchung@soback.kornet.nm.kr>
Date: Fri, 30 Jan 1998 22:03:31 +0900
Links: << >>  << T >>  << A >>
In altera ftp suite(ftp.altera.com), you can find eau019.exe in /pub/utils
directory.
This files reads .tt2 file from Data I/O v4.0 or later and translate into
.tdf(AHDL) file.
Hope this helps.

Thanks
D.H.
Ying C. 이(가) <6am21r$hn2$1@agate.berkeley.edu> 메시지에서 작성하였습니다...
>
>I believe you can compile an ABEL file in Synario and get out an AHDL
>output file.
>
>Ying
>ying@csua.berkeley.edu
>
>
>
>In article <34CD9B42.52C0@altatech.com>,
>Douglas L Datwyler  <datwyler@altatech.com> wrote:
>>Does any one know of a RELIABLE ABEL to Altera-HDL conversion utility?
>>
>>Where is the FAQ for this news group?
>>
>>Thank you,
>>
>>Douglas L Datwyler
>>Alta Technology
>>datwyler@altatech.com
>
>
>--
>-----------------------------------
>http://www.csua.berkeley.edu/~ying


Article: 8829
Subject: How to design 3-staged pipelined multiplier in VHDL for Xilinx 4000XL
From: Johannes Slhusvik <jso@nocrc.abb.no>
Date: Fri, 30 Jan 1998 16:25:47 +0100
Links: << >>  << T >>  << A >>
Can anyone guide me with the design of a 3-staged pipelined multiplier
in VHDL that is optimised for Xilinx 4000XL series ?

-- 
Johannes Slhusvik, PhD, Electronic Systems
ABB Corporate Research Centre, Norway
Tel: +47 66 84 34 28, Email: jso@nocrc.abb.no
Article: 8830
Subject: Re: Comments about Xilinx Alliance m1.4 w/Novell and other problems
From: Todd KLine <tkline@wgate.com>
Date: Fri, 30 Jan 1998 11:07:41 -0500
Links: << >>  << T >>  << A >>
Marco Rivero wrote:
 
> 3. I cannot lock a pin onto a complemented NET on a .ucf file.  Tried a "-" as
> I used to do with XACT: didn't work.  Tried a "~" as done on Viewlogic netlists,
> and .ncd files!: Didn't work.  Called customer support and was told to remove

In a previous life I was the Viewlogic tools manager.  I insisted that
the "~" NOT be used for inverted nets because different tools had
different sets of acceptable characters.  While one should probably
expect trivial things like character sets to be common among tools by
now, such is not the case.  I would strongly reccomend using only
alpha-numeric characters and the "_" character in all net and module
names.  From my experiance, these are the only characters that you can
be sure will be accepted by other tools.

While it is somewhat cumbersome, I always use "_l" as a suffix for
active low nets and I would guess that other people have similar
constructs.

Todd
Article: 8831
Subject: Re: Comments about Xilinx Alliance m1.4 w/Novell and other problems
From: Todd KLine <tkline@wgate.com>
Date: Fri, 30 Jan 1998 11:08:45 -0500
Links: << >>  << T >>  << A >>
Marco Rivero wrote:
 
> 3. I cannot lock a pin onto a complemented NET on a .ucf file.  Tried a "-" as
> I used to do with XACT: didn't work.  Tried a "~" as done on Viewlogic netlists,
> and .ncd files!: Didn't work.  Called customer support and was told to remove

In a previous life I was the Viewlogic tools manager.  I insisted that
the "~" NOT be used for inverted nets because different tools had
different sets of acceptable characters.  While one should probably
expect trivial things like character sets to be common among tools by
now, such is not the case.  I would strongly reccomend using only
alpha-numeric characters and the "_" character in all net and module
names.  From my experiance, these are the only characters that you can
be sure will be accepted by other tools.

While it is somewhat cumbersome, I always use "_l" as a suffix for
active low nets and I would guess that other people have similar
constructs.

Todd
Article: 8832
Subject: Re: MAX+II Version 8.2
From: ecp@focus-systems.on.ca (Eric Pearson)
Date: Fri, 30 Jan 1998 17:41:51 GMT
Links: << >>  << T >>  << A >>
In article <34CF5524.255D@nortel.co.uk>,
Peter Rush  <prush@nortel.co.uk> wrote:
>Maybe it's just the way that I'm using it, but does anybody else out
>there think that 8.2 is a giant leap backwards from 8.1!
>
>For my project (60% full 10K50, 2 clocks 25 and 50 MHz) compile times
>have gone from about 0.5 hrs to about 3 hrs, cliques seem to be ignored,
>and I get 10 - 15 timing constraints not met (previously 5 not met).
>
>I'm just glad I can still use 8.1.

Wow, I thought I was the only one!

Our design in a 10K130 compiles under 8.1. Under 8.2 it does'nt even
get past the syntax checking stage, it blows up in one of the LPMs.
We would really like 8.2 as it has the improved timing models for the
10K130.

Just for interests sake we are at 92% of a 10K130 running at 40 Mhz,
compile times are chaotic. Minor changes in design give compile times
ranging from 1.2 hours to 44.5 hours ( at least thats the longest we've
waited before giving up, and making some minor change ). If I add timing
constraints the compiler gives up. We started out at 150% utilization,
the design only started to route when we reached 92% LE utilization.

Even with complaints though, we DO have a functioning design.

Eric Pearson 



-- 
Eric Pearson -- Focus Systems -- Waterloo, Ontario
     ecp@focus-systems.on.ca  (519) 746-4918
    "We Engineer Innovative Imaging Solutions"
Article: 8833
Subject: Re: VHDL vs schematics
From: Geir Harris Hedemark <geirhe@hridil.ifi.uio.no>
Date: 30 Jan 1998 21:16:16 +0100
Links: << >>  << T >>  << A >>
madarass@cats.ucsc.edu (Rita   Madarassy) writes:
> On the other hand, in schematic I can customize the carry logic using
> CY4 components in Xilinx and build a priority encoder that will perform
> at about 10ns in a 4000XL-09. Not bad ahh? 

So what you are saying is that a synthesis tool that does not know
these CY4 components will perform poorly. That does not surprise me.

What happens when you use a synthesis library that contains the CY4
component you mention? Will the synthesis tool still perform poorly,
and if so: Why won't it use the component?

Geir
Article: 8834
Subject: Re: VHDL vs schematics
From: "Adam J. Elbirt" <aelbirt@viewlogic.com>
Date: 30 Jan 1998 21:30:09 GMT
Links: << >>  << T >>  << A >>
Geir,

What really matters in this case is whether or not the synthesis tool is
smart enough to take advantage of the element for your given design.  Some
are, some aren't, and there are a lot of factors that come into play. 
Typically, each tool has a variety of settings that result in the usage of
different library elements.  A typical example is a resource sharing option
that would use an ADDSUB component when turned on and one each of ADD and
SUB components when turned off.  For situations like these, you need to
examine your synthesis tool outputs and typically tweak the appropriate
settings to try and get what you want.  Even then, you may have to resort
to schematics to implement EXACTLY what you want for optimal performance or
area utilization.

Adam

Geir Harris Hedemark <geirhe@hridil.ifi.uio.no> wrote in article
<xq3n2gdlnu7.fsf@hridil.ifi.uio.no>...
> madarass@cats.ucsc.edu (Rita   Madarassy) writes:
> > On the other hand, in schematic I can customize the carry logic using
> > CY4 components in Xilinx and build a priority encoder that will perform
> > at about 10ns in a 4000XL-09. Not bad ahh? 
> 
> So what you are saying is that a synthesis tool that does not know
> these CY4 components will perform poorly. That does not surprise me.
> 
> What happens when you use a synthesis library that contains the CY4
> component you mention? Will the synthesis tool still perform poorly,
> and if so: Why won't it use the component?
> 
> Geir
> 
Article: 8835
Subject: Re: VHDL vs schematics
From: madarass@cats.ucsc.edu (Rita Madarassy)
Date: 30 Jan 1998 23:49:44 GMT
Links: << >>  << T >>  << A >>
In article <xq3n2gdlnu7.fsf@hridil.ifi.uio.no>,
Geir Harris Hedemark  <geirhe@hridil.ifi.uio.no> wrote:
>madarass@cats.ucsc.edu (Rita   Madarassy) writes:
>> On the other hand, in schematic I can customize the carry logic using
>> CY4 components in Xilinx and build a priority encoder that will perform
>> at about 10ns in a 4000XL-09. Not bad ahh? 
>
>So what you are saying is that a synthesis tool that does not know
>these CY4 components will perform poorly. That does not surprise me.
>
>What happens when you use a synthesis library that contains the CY4
>component you mention? Will the synthesis tool still perform poorly,
>and if so: Why won't it use the component?

There are a few tradeoffs when using CY4 components. First of all, you
will inherently need to assign location constraints to the CY4s otherwise
the xilinx tools will complain with errors. Second, customizing the CY4 
components  for specific implementations is EXTREMELY HARD to
do. Only the latter generation synthesis tools like Synplicity will implement
CY4 for other non-arithmetic code.
Last but not least, when adding location constraints you may make
life very hard for the router since the xilinx tools will have to
deal with a large chunk of logic.   

 
>
>Geir


Article: 8836
Subject: Re: xilinx M1 protection
From: Jack Lai <jwlai@mmmpcc.org>
Date: Fri, 30 Jan 1998 20:24:55 -0600
Links: << >>  << T >>  << A >>
bootrecord wrote:
> 
> persons wishing to modify the Volume Serial Number may use the following
> process:
> 
> re-boot in MS-DOS mode.  (dos boxes can't write to the boot)
> DEBUG
> L 100 2 0 1    ; read a copy of the boot sector from #2 or C:
> D 100      ; dump 128 bytes, look for words FAT16 or FAT32
> ( you can view the present serial number at the offsets listed below)
> ( W 100 0 5 1    ; write a copy to sector 5 of #0 or A: scratch floppy
> if you wish )
> ( Later you can L 100 0 5 1 and W 100 2 0 1 to restore if you must )
> E127    ; edit starting at byte CS:127 for FAT16 or FAT12
> E143    ; edit starting at byte CS:143 for FAT32
> ( enter desired serial number, in reverse byte order, follow each with
> <SPACE>)
> ( serial number 1234-5678 is entered as 78 56 34 12 )
> W 100 2 0 1    ; write it back out
> 
> notes: in a previous post in this ng, S.E.E. indicated difficulty with
> the FAT32 version, and had to format.  S.E.E. placed his R/W buffer at
> 0x0 instead 0x100.  this may have been the source of the problem, it has
> been customary to use 0x100, because the PSP and other things (stack)
> are placed below 0x100.  it's not clear to me that these areas would be
> used by the L and W commands, though.  my tests were successful with
> FAT12, FAT16, and FAT32 volumes.  also, ? gives help in debug.
> 
> regards to all xilinx M1 users.
> bill


Great info Bill, what about NTFS?

Regards, Jack

Article: 8837
Subject: Re: VHDL book
From: "HDLsRFun" <brownsco@frii.com>
Date: Fri, 30 Jan 1998 22:56:21 -0700
Links: << >>  << T >>  << A >>
Absolutely agree. Whenver I have a question or issue with even the most
esoteric parts of VHDL, Ashenden has an example and a understandable
description. I began with VHDL in '88, and
looked at many books; Ashenden's is the best single resource
available.


Deependra Talla wrote in message <6am4ub$p0l@ftp.ee.vill.edu>...
>The Designer's Guide to VHDL by Peter Ashenden is the coolest book..
>
>--
>~~~~~~~~~~~~~~~~~~~~ deepu@ece.vill.edu ~~~~~~~~~~~~~~~~~~~~~~~~~~~
>| Deepu Talla              | If it is not necessary to change     |
>|                          |    it is necessary not to change ... |
>| Phone: (610)225-0243 (R) |                                      |
>|        (610)519-7371 (O) | whatever, whoever, wherever, whenever|
>~~~~~~~~~~~~~~~~ http://www.ece.vill.edu/~deepu ~~~~~~~~~~~~~~~~~~~


Article: 8838
Subject: Re: MAX+II Version 8.2
From: "Matthew Morris" <matthewm@teal.csn.net>
Date: Sat, 31 Jan 1998 16:27:33 -0700
Links: << >>  << T >>  << A >>
<snip snip>
>get past the syntax checking stage, it blows up in one of the LPMs.
>We would really like 8.2 as it has the improved timing models for the
>10K130.
>
You can always change the timing models, they are not linked to the
compiler. Just take the DMF file from 8.2 and put it in the MAXPLUS2
directory. I would back them up first of course. You can even tweak that
file if you want to do more than one corner on your simulation.

Matt


Article: 8839
Subject: Re: VHDL vs schematics
From: Ray Andraka <no_spam_randraka@ids.net>
Date: Sat, 31 Jan 1998 22:06:49 -0500
Links: << >>  << T >>  << A >>
Rita Madarassy wrote:
> 
> Normally, you maybe right there is no difference between hdl and
> schematic. But there is a huge difference when it really matters.
> For example, you may need a very high performance 32 bit priority encoder.i
> 
> A 32  bit priority encoder will likely be a huge bottle neck in your
> design if you this in vhdl. Synthesis will create this priority
> encoder using 4 input lookup tables and it will slow down your
> design to a ridiculoys speed.
> On the other hand, in schematic I can customize the carry logic using
> CY4 components in Xilinx and build a priority encoder that will perform
> at about 10ns in a 4000XL-09. Not bad ahh?
> Doing this implementation in VHDL will take at least 20 ns if the routes
> are decent and the synthesis tool is fairly good.
> 
> If you want to get your job done and meet deadlines vhdl is the
> way to go. However, be prepared to replace some of your modules
> by designs done in schematic when vhdl doesn't do it.
> 
> In article <6aqqra$3nl@borg.svpal.org>,
> George Noten <garyk@svpal.svpal.org> wrote:
> >Simon Ramirez (s_ramirez@email.msn.com) wrote:
> > I am not sure there is a big difference.  You *can* do in VHDL absolutely
> > the same design that you do in schematic (i.e. write a netlist).  The
> > only difference is that VHDL will be more portable and easier to
> > simulate.  OTOH, you will need a schematic to understand what you wrote
> > in VHDL.
> >
> >
> > So-called "behavioral" VHDL description comes down to 2-3 sequential
> > statements that the synthesis tool is able to understand (and each
> > vendor interprets them in a different way).  No big deal.  Sometimes
> > I do use behavioral description, but usually I prefer netlist and para-
> > metrized libraries (XBLOX or LPM).  Normally there is no difference in
> > performance between this type of design and schematic.
> >
> >       George.
> >

You can get the structure you need to get the desired performance and
density using an HDL, but you will likely have to resort to low level
instantiation and a host of extra attributes to control placement. 
Effectively, this is writing a textual description of the schematic. 
The current HDL synthesis tools really don't provide enough control over
the design implementation to quickly, reliably and successfully turn out
high performance designs.  Hopefully the CAE vendors will fix this soon,
but I am not holding my breath.

As I see it, the two real advantages an HDL has over well executed
schematics are:
1) the design concept can be simulated behaviorly before the detail
design is even started (increasingly important).  THis allows a higher
degree of HW/SW cosimulation, as well as early test bench development
and the ability to find systemic snafus before the design has gone on
for months.  
2)Archiving the design.  Schematics generally require the schematic tool
to access electronic archives of the design, and definitely need the
tool to make modifications.  The CAE vendors can and do make changes to
the database formats from rev to rev, which can make updating an old
design very painful (BTDT).  HDL's on the other hand only need a text
editor to access and modify the design's source.  You will still need a
compiler to convert the source to a bit stream of course.  

All of the other claims to an HDL's superiority are suspect at best, and
in most cases just pure marketing hogwash.  Here are a few of the more
common ones:

Portability between devices/technologies:  This is great as long as you
keep the design at an abstract enough level to make it work.  Doing this
avoids using all the device's architectural features so you wind up with
a hopelessly slow and large design.  Instantiation of the device
features can be easily done of course, but a well developed schematic
library does the same thing. I've translated designs between
technologies in schematics, and it is not very painful if the schematic
is hierarchical and the partitioning between hierarchical blocks is
logical.

Design readability: proponents of thsi argument all seem to imply that
schematics are flat and HDLs are hierarchical.  A schematic can be
easily made hierarchical so that no single schematic is more than a
single page (this is the way I do mine).  Proper partitioning makes a
very readable design.  On the otherhand, I have seen some absolutely
terrible VHDL code that while not flat, might just as well been.  I am
still a believer in the age-old adage that a picture is worth a thousand
words.  For those who argue that state machines are easier to read in
HDL, there are methods to make a schematic machine readable as well. 
For one-hot machines, the symbols for the components can be set up to
resemble a flowchart.  for encoded machines, the machine can be
represented as a mux and register so that the next state is directly
readable.

Comments are easy to use in HDLs:  and they should be or no one will
ever follow the code.  A well drawn schematic rarely needs comments. 
Refer to the picture=1000 words note above.

HDL and synthesis makes for a push button solution:  Using the push
button solution removes the designer from the loop.  The result is poor
implementation of the design due to the synthesizer's not so hot
abilities in partitioning the design and synthesizing the logic.  Many
of the synthesis engines have many controls over the style of
synthesis.  Unfortunately, these controls tend to have global influence
over the design, and the effects are not necessarily intuitive.  THe
synthesis results also vary from tool to tool.

As you can see, (and many of you already know this) I am not a big fan
of HDLs at the current time, mostly because the control over the
implementation in the current tools is not as complete or as intuitive
as those controls applied in schematic entry.  Give me that control, and
I think the two real advantages stated above would tip the balance in
favor of HDLs.  For now, however, I can get to the desired performance
and logic density faster using schematic entry, and since I get paid to
get a working design, that means I either earn more money by doing more
designs or have a little extra time to devote to my wife and kids or to
flying (which is where the extra money part comes in).

-Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email randraka@ids.net
http://users.ids.net/~randraka
Article: 8840
Subject: Re: How to design 3-staged pipelined multiplier in VHDL for Xilinx 4000XL
From: Ray Andraka <no_spam_randraka@ids.net>
Date: Sat, 31 Jan 1998 22:08:33 -0500
Links: << >>  << T >>  << A >>
Johannes Slhusvik wrote:
> 
> Can anyone guide me with the design of a 3-staged pipelined multiplier
> in VHDL that is optimised for Xilinx 4000XL series ?
> 
> --
> Johannes Slhusvik, PhD, Electronic Systems
> ABB Corporate Research Centre, Norway
> Tel: +47 66 84 34 28, Email: jso@nocrc.abb.no

How many bits and what's the clock rate?  Makes a big difference in the
best implementation, and the level of instantiation required.
Article: 8841
Subject: Re: How to design 3-staged pipelined multiplier in VHDL for Xilinx 4000XL
From: Addie Tang <kctang@ee.cityu.edu.hk>
Date: Sun, 01 Feb 1998 14:58:15 +0800
Links: << >>  << T >>  << A >>
Hi,

   If you are using Synopsys, you can instantiate a 3stage multiplier
from DesignWare. It is defined in Advance Math components
in library DW02. You can check the use of this component for component
instantiation or through "*" operator using --pragma map_to_operator in
the DesignWare manual.

Addie Tang
Department of Electronic Engineering
EDA Centre
City University of Hong Kong

Johannes Slhusvik wrote:

> Can anyone guide me with the design of a 3-staged pipelined multiplier
> in VHDL that is optimised for Xilinx 4000XL series ?
>
> --
> Johannes Slhusvik, PhD, Electronic Systems
> ABB Corporate Research Centre, Norway
> Tel: +47 66 84 34 28, Email: jso@nocrc.abb.no



Article: 8842
Subject: can u give me some advice?
From: engp7604@leonis.nus.sg (Gu Feng)
Date: 1 Feb 1998 07:21:22 GMT
Links: << >>  << T >>  << A >>
Hi, there.

I'd like to buy a basic development system for FPGA, can anybody give me
some suggestion which system I should buy?

Thanks in advance.

GuFeng

engp7604@leonis.nus.sg


--
-------------------------------------------------------------------------------
GuFeng                        
Electrical Engineering Dept.   
National University of Singapore
Tel:771-5252
email: engp7604@leonis.nus.sg   gufeng@ee.nus.sg
-------------------------------------------------------------------------------

Article: 8843
Subject: Re: VHDL vs schematics
From: z80@ds.com (Peter)
Date: Sun, 01 Feb 1998 13:08:46 GMT
Links: << >>  << T >>  << A >>
Good summary, IMHO.  

I would have said that an HDL results in a more readable source when
used for designs which are "random" in nature, e.g. decoder-heavy
stuff, and encoded state machines are certainly in that category.

One of the "problems" with schematics is that for some years now the
trend has been to generate a huge number of small sheets. 

I have seen very few designs that really have such a huge number of
functionally separate portions, so the sheet division tends to be
unnecessary, and people spend lots of time inventing names for
signals, just so they can connect them from one sheet to the next. The
result is a unreadable schematic.

I think one reason for this trend is the fact that many designers
don't have access to a decent A3/A2 hard copy device, e.g. a plotter.
I have seen ASIC design offices, with 10 people doing ASICs, and their
sole hard copy device is an A4 laser. This is stupid.

Also, not many people want to take care to draw a neat schematic these
days. They rather throw a few gates on a piece of paper, and a few
more gates on another piece of paper. Especially if they are full-time
employed. Drawing a schematic so it is readable takes a bit of
dedication, and most people can't do it well.

A while ago I did a design which filled a XC3090. This went nicely on
about five A3 sheets, each of which was entirely readable, although if
I was passing it to a customer I would have plotted it on A2 sheets.

Some time before that, I was on a project, of similar complexity,
which was designed on about 50 A4 sheets, filling a ring binder, and
most of it was incomprehensible.


Peter.

Return address is invalid to help stop junk mail.
E-mail replies to z80@digiXYZserve.com but
remove the XYZ.
Article: 8844
Subject: FPGA/ASIC - same difference?
From: Magnus Homann <d0asta@palver.dtek.chalmers.se>
Date: 01 Feb 1998 17:04:16 +0100
Links: << >>  << T >>  << A >>

Hi,

I'm just curious of how big a step there is from FPGA to ASIC? I
cnostatntly hear that ASIC is so much harder, but is that just to
protect the ASIC-designers postions?

I can accept that there is some added difficulty, but how hard can it
really be? After doing a few FPGA designs (1-3) and taking a week
course in ASIC design, would I be ready?

Homann
-- 
   Magnus Homann  Email: d0asta@dtek.chalmers.se
                  URL  : http://www.dtek.chalmers.se/DCIG/d0asta.html
  The Climbing Archive!: http://www.dtek.chalmers.se/Climbing/index.html
Article: 8845
Subject: Re: VHDL vs schematics
From: "rk" <stellare@erols.com.NOSPAM>
Date: 1 Feb 1998 16:18:11 GMT
Links: << >>  << T >>  << A >>
Peter from x said:
: Good summary, IMHO.  
: 
: I would have said that an HDL results in a more readable source when
: used for designs which are "random" in nature, e.g. decoder-heavy
: stuff, and encoded state machines are certainly in that category.

rk:
it's easy to see when a block is done in the "wrong" methodology: when
you're drawing a schematic to understand some vhdl or when you're writing
an equation or making a flow chart to understand a schematic.

peter:
: One of the "problems" with schematics is that for some years now the
: trend has been to generate a huge number of small sheets. 

rk:
fast, cheap laser "a-sized" laser printers on the desktop.  while b-sized
sheets fit relatively nicely onto this format and with some care are
readable, d-sized sheets are an eye-chart with reference designators often
unreadable.    while moderately priced d-sized plotters are available with
ink jet technology (i.e., designjet 200 series from hp) that are much
faster then their mechanical arm pen plotter predecessors, it still takes a
lot of time to plot out a drawing set.  depending on the design, i find i
do about half in all d-sized and half in all b-sized.

peter:
: I have seen very few designs that really have such a huge number of
: functionally separate portions, so the sheet division tends to be
: unnecessary, and people spend lots of time inventing names for
: signals, just so they can connect them from one sheet to the next. The
: result is a unreadable schematic.

rk:
seen too many designs done with small clusters of gates, 2 or 3 in a bunch,
no logical flow at all.  very unreadable.  on the other hand, in my
opinion, every macro instance and net should be visibly named.  then, when
using the drc, timing analysis, manual placement, and in-circuit debugging
tools, it's much easier to proceed faster, without having to search for the
invisible identifiers that the cae system does apply.  a bit of work and a
pain up front but for any significant design i find it worth the effort.

now when designing with an hdl or macro generator, frequently i'll use the
schematic generator to make a schematic out of it and depending on which
tool is used, it can be either tolerable with a bit of work or a total
mess.  with no net names and reference designators specified, the timing
path identified by the timing analyzer as too long is a pain to find
quickly, especially if the schematic generation is done in levels.  and if
using in circuit probing, it's much easier to proceed quickly from a
well-drawn, well-labelled schematic.

peter:
<snip>
: Also, not many people want to take care to draw a neat schematic these
: days. They rather throw a few gates on a piece of paper, and a few
: more gates on another piece of paper. Especially if they are full-time
: employed. Drawing a schematic so it is readable takes a bit of
: dedication, and most people can't do it well.

it takes a bunch of time to generate a schematic.  and the engineers
workload is increasing.  it wasn't too long ago when we could sketch a
schematic by hand on a d-sized piece of paper (much faster than *ANY* of
the cae tools) and the draftsman would either draw it or enter into the
system and maintain it when the engineer marked it up.  and now we have
secretaries that won't answer the phone, won't order supplies, won't type
papers, won't do the mail, won't photocopy, won't fax ...

-------------------------------------------------------------
rk

"there's nothing like real data to screw up a great theory" 
- me (modified from original, slightly more colorful version)
--------------------------------------------------------------
Article: 8846
Subject: Re: VHDL vs schematics
From: Steve Goodwin <steve@p2cl_DSPM.demon.co.uk>
Date: Sun, 1 Feb 1998 16:50:26 +0000
Links: << >>  << T >>  << A >>
In article <34d8731a.71748198@news.netcomuk.co.uk>, Peter <z80@ds.com>
writes

[good stuff snipped from Peter and Ray]

>I think one reason for this trend is the fact that many designers
>don't have access to a decent A3/A2 hard copy device, e.g. a plotter.
>I have seen ASIC design offices, with 10 people doing ASICs, and their
>sole hard copy device is an A4 laser. This is stupid.

Just as a aside, if you are using Windoze, look for the Cannon A3
inkjets. I have the BJ230 for exactly this reason, cheap and reasonably
fast. There may be one around in colour by now.

[and snipped again]

Also, on the HDL/schematic stuff, I wouldn't let myself do a real (i.e.
used by a customer) C++ program until I understood (at least reasonably
well) what the compilers were doing in the background (non-trivial). Now
I use it where its appropriate as I do with other languages.

My use of HDL will go the same way. Understand its no magic bullet but
will help in some cases and may be absolutely necessary in others. I
can't see myself leaving schematics completely though. Maybe I'm just
old <grin>. (No maybe about it really...).

By the way, as for commenting HDL, unless its the gotcha's or general
overview my belief is that the code should comment itself as much as
possible. Otherwise who's to say that the comments and code agree :). In
a way, schematics might benefit more from comments than HDL does.

Just my .002p worth
-- 
Steve Goodwin...  De-spamming active, remove any _DSPM from address
Article: 8847
Subject: Re: VHDL book
From: Richard Schwarz <aps@associatedpro.com>
Date: Sun, 01 Feb 1998 13:22:58 -0500
Links: << >>  << T >>  << A >>
Check at http://www.associatedpro.com/aps

bart plackle wrote:

> can anyone point out a very good advanced vhdl book
>
> Kind regards Bart Plackle



--
__/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/

Richard Schwarz, President              EDA & Engineering Tools
Associated Professional Systems (APS)   http://www.associatedpro.com
3003 Latrobe Court                      richard@associatedpro.com
Abingdon, Maryland 21009
Phone: 410.569.5897                     Fax:410.661.2760

__/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/


Article: 8848
Subject: Re: can u give me some advice?
From: "Steven K. Knapp" <sknapp@optimagic.com>
Date: Sun, 1 Feb 1998 10:24:19 -0800
Links: << >>  << T >>  << A >>
There are a few try-before-you-buy packages available for free (or a low
cost) via the web at http://www.optimagic.com/lowcost.html.

-----------------------------------------------------------
Steven K. Knapp
OptiMagic, Inc. -- "Great Designs Happen 'OptiMagic'-ally"
E-mail:  sknapp@optimagic.com
   Web:  http://www.optimagic.com
-----------------------------------------------------------

Gu Feng wrote in message <6b17pj$j52@nuscc.nus.sg>...
>Hi, there.
>
>I'd like to buy a basic development system for FPGA, can anybody give me
>some suggestion which system I should buy?
>
>Thanks in advance.
>
>GuFeng
>
>engp7604@leonis.nus.sg
>
>
>--
>---------------------------------------------------------------------------
----
>GuFeng
>Electrical Engineering Dept.
>National University of Singapore
>Tel:771-5252
>email: engp7604@leonis.nus.sg   gufeng@ee.nus.sg
>---------------------------------------------------------------------------
----
>


Article: 8849
Subject: Re: Introduction book on Verilog/VHDL
From: "Steven K. Knapp" <sknapp@optimagic.com>
Date: Sun, 1 Feb 1998 10:32:07 -0800
Links: << >>  << T >>  << A >>
I like one called "HDL Chip Design:  A Practical Guide for Designing,
Synthesizing, and Simulating ASICs and FPGAs using VHDL or Verilog" by
Douglas J. Smith.  ISBN 0-9651934-3-8.  Also see
http://www.doone.com/hdl_chip_des.html and
http://www.amazon.com/exec/obidos/ISBN=0965193438/9196-2112036-285223 to
order it via Amazon.com.

There are other books on VHDL, Verilog, and FPGA design listed at
http://www.optimagic.com/books.html.

-----------------------------------------------------------
Steven K. Knapp
OptiMagic, Inc. -- "Great Designs Happen 'OptiMagic'-ally"
E-mail:  sknapp@optimagic.com
   Web:  http://www.optimagic.com
-----------------------------------------------------------



Xiao Wang wrote in message <34D18B9F.44B1@umr.edu>...
>I'm just beginning to study Verilog or VHDL.
>
>Is there any difference between these two names?
>
>What will be the very suitable introductory book for this Verilog and/or
>VHDL? Thanks.
>
>Please email back: xiao@umr.edu




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