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 13400

Article: 13400
Subject: Viewlogic setup with Xilinx and Altera Backend PAR
From: #YEO WEE KWONG# <P7102672H@ntu.edu.sg>
Date: Tue, 1 Dec 1998 17:53:08 +0800
Links: << >>  << T >>  << A >>
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE1D10.663C02E6
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

I had tried very hard to integrate the Viewlogic EDA tool (Workview
office) with the backend PAR for Xilinx alliance series and Altera
Maxplus 2 and encountered a lot of problems which I hope you could
render some assistance. 

The questions are as follow:
*	The steps or installation steps to integrate the above tools
*	How do we create versioning for different synthesis and device
fitting design
*	How to set up the environment variables for the above tools( you
may like to code me the details assuming that my xilinx installation is
at c:\xilinx and altera software at c:\maxplus2)
*	There seems to be 2 set of libraries available, one from
Viewlogic another from vendor, which should we use and how to set them?
*	How do we download the bit file (or configuration file) to
initialize the FPGA chip after it is generated at Intelliflow.

Your help in solving the above problems would grant me more faith in
Viewlogic tool. Hope you can help.

If possible, please direct your reply to yeosky@mbox3.singnet.com.sg
<mailto:yeosky@mbox3.singnet.com.sg>  Thank you

Yeo Wee Kwong

------_=_NextPart_001_01BE1D10.663C02E6
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>Viewlogic setup with Xilinx and Altera Backend PAR</TITLE>
</HEAD>
<BODY>

<P><FONT FACE=3D"Arial">Hi,</FONT>
</P>

<P><FONT FACE=3D"Arial">I had tried very hard to integrate the =
Viewlogic EDA tool (Workview office) with the backend PAR for Xilinx =
alliance series and Altera Maxplus 2 and encountered a lot of problems =
which I hope you could render some assistance. </FONT></P>

<P><FONT FACE=3D"Arial">The questions are as follow:</FONT>

<UL><LI><FONT FACE=3D"Arial">The steps or installation steps to =
integrate the above tools</FONT></LI>
<LI><FONT FACE=3D"Arial">How do we create versioning for different =
synthesis and device fitting design</FONT></LI>
<LI><FONT FACE=3D"Arial">How to set up the environment variables for =
the above tools( you may like to code me the details assuming that my =
xilinx installation is at c:\xilinx and altera software at =
c:\maxplus2)</FONT></LI>
<LI><FONT FACE=3D"Arial">There seems to be 2 set of libraries =
available, one from Viewlogic another from vendor, which should we use =
and how to set them?</FONT></LI>
<LI><FONT FACE=3D"Arial">How do we download the bit file (or =
configuration file) to initialize the FPGA chip after it is generated =
at Intelliflow.</FONT></LI>
<BR>
</UL>
<P><FONT FACE=3D"Arial">Your help in solving the above problems would =
grant me more faith in Viewlogic tool. Hope you can help.</FONT>
</P>

<P><FONT FACE=3D"Arial">If possible, please direct your reply to =
</FONT><A HREF=3D"mailto:yeosky@mbox3.singnet.com.sg"><U><FONT =
COLOR=3D"#0000FF" =
FACE=3D"Arial">yeosky@mbox3.singnet.com.sg</FONT></U></A><FONT =
FACE=3D"Arial"> Thank you</FONT>
</P>

<P><B><I><FONT COLOR=3D"#008080" FACE=3D"Bookman Old Style">Yeo Wee =
Kwong</FONT></I></B><B><I></I></B><B><I></I></B>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BE1D10.663C02E6--

Article: 13401
Subject: Re: Which parts are fastest for 3-state enables?
From: cartoffice@mail3.clio.it
Date: Tue, 01 Dec 1998 13:25:41 GMT
Links: << >>  << T >>  << A >>
dsaffcvncncncnOn Tue, 24 Nov 1998 23:37:40 GMT, "Steve"
<reply.through.newsgroup@paranoid.com> wrote:

>I'm interfacing to a Motorola MPC 860 @50Mhz.  I need to provide an
>acknowledge signal in <10ns after the Clock.  To do this my PLD/FPGA
>needs to have a (Clk-Q + comb delay + tristate enable) < 10ns.
>
>So far I haven't found a CPLD to do it.  What small FPGA's have the best
>crack at it???  ... or prove me wrong on the CPLD issue?
>
>So far my only practical solution is an XCS05XL-4.
>
>
>Comments?
>
>Steve
>
>

Article: 13402
Subject: Re: Will XILINX survive?
From: Tim Tyler <tim@BITS.bris.ac.uk>
Date: Tue, 1 Dec 1998 13:33:05 GMT
Links: << >>  << T >>  << A >>
Victor the Cleaner <jonathan@canuck.com> wrote:
: Tim Tyler (tim@BITS.bris.ac.uk) wrote:
: : Bob Sefton <nobody@home.com> wrote:
: : : What an absolute crock of shit. Save this crap for
: : : misc.invest.stocks where somebody might buy into it.

: : Unlikely - it's easy to ignore anonymous posters trying to diss companies
: : - they could equally well be from the backers of the competition.

: Or from a disenchanted employee wanting to get the "truth" out and fearing
: reprisals.

Few employees become so disenchanted with their company that they try to
back-stab it while still in a position to receive reprisals.  Ex-employees
would perhaps be more likely to behave in this manner - but they may well
have grudges to take into account.

I don't know what 'Anonymous' is trying to achieve with its post, but the
method used doesn't provide much in the way of credibility.
-- 
__________
 |im |yler  The Mandala Centre  http://www.mandala.co.uk/  tt@cryogen.com

...but officer, I was only going one way!

Article: 13403
Subject: HELP, Tool selection
From: jerry english <jenglish@planetc.com>
Date: Tue, 01 Dec 1998 09:47:01 -0500
Links: << >>  << T >>  << A >>
I am trying to make an informed decision on which FPGA HDL compiler
I should purchase. I have evaluated three, FPGA Express (Synopsys),
Synplify (Synplicity), and Leonardo Spectrum level 3 (Exemplar).
All three tools produce similar results, all have basically the same
features, all are priced about the same give or take a few thousand.
Any comments, positive or negative, about the tool you are using that
you would care to share would be greatly appreciated. If there is a
demand
I could tabulate and post the summary.

Regards
Jerry English



Article: 13404
Subject: Re: Logical Devices ALLPRO diagnostics
From: "Paul T. Shultz" <paul@csciences.com>
Date: Tue, 01 Dec 1998 12:24:10 -0500
Links: << >>  << T >>  << A >>
Ellis,

We have an AllPro 88, software version 2.90.  In our installation the
diagnostics are executed through the main program under the utilities menu
item.  The programs and directories that you referenced are not present in our
system.

Regards,

Paul T. Shultz
Chesapeake Sciences Corporation
<paul@csciences.com>

Ellis Easley wrote:

> This question is not exactly FPGA related, but this is the only newsgroup I
> can find that deals with programmable logic. I am looking for the diagnostic
> programs for the Logical Devices AllPro (40) or AllPro 88 programmer. The
> diagnostic programs provided with the AllPro 88 are called DIAG1, DIAG3, and
> DIAG4.EXE and are located in directories named DIAG1, 3, and 4, according to
> the service manual.
>
> I am trying to use an AllPro programmer that I found in an internet auction.
> Before I bid on it I contacted Logical Devices about support. They told me I
> could get a service manual from them that would provide all the information
> I would need to verify the machine was working before buying the programming
> software. I understood this to mean it would contain either program listings
> or programming information which I could use to exercise the hardware. I
> ordered the manual and find it does not contain the details I need. It
> covers a newer model (AllPro 88) and while the schematics it contains are
> similar to the hardware in my unit, there are no PAL equations and no
> schematics of the pindriver circuits. I have figured out the circuitry up to
> the pindrivers but don't want to go further till I know whether it is
> possible to turn on more than one source in a pindriver. I don't want to
> accidentally connect 25V at 1A to a TTL output or a FET to ground!
>
> Logical Devices says the AllPro 88 software recognizes the older AllPro
> model and will do all devices up to 40 pins (the 88 has 88 pindrivers, the
> old model has 40).
>
> Ellis Easley Jr.
> ehowell.easley@worldnet.att.net

Article: 13405
Subject: Re: Is it normal to have to edit the xnf file???
From: Todd Kline <todd@wgate.com>
Date: Tue, 01 Dec 1998 13:04:49 -0500
Links: << >>  << T >>  << A >>
Steve wrote:

> I just got off the phone with Xilinx support re a problem
> with FPGA Express creating an xnf which generated
> errors in the Xilinx tools.
>
> They said that editing the xnf to fix the error was the best
> solution.  Is this normal?  There's got to be a better way.
>
> BTW, I was going to try edif output as recommended in
> answer 2866 but neither I (nor support) know how to
> select edif output from Express!
>
> This is getting a little frustrating.
>
> Steve

1) What exactly is the error?

2) To the best of my knowledge, Virtex parts are the only
    ones which produce EDIF files.  You cannot get EDIF
    files for older families.

Todd

Article: 13406
Subject: Re: HELP, Tool selection
From: Tim Davis <timdavis@tdcon.com>
Date: Tue, 01 Dec 1998 11:39:55 -0700
Links: << >>  << T >>  << A >>
Jerry,

A client's engineer had a problem with FPGA express (in Xilinx Foundation). He
had a large state machine (> 32 states) that was one hot encoded. He decoded
states for some output signals and found that the synthesized logic was decoding
*all* state bits instead of just the signal for the particular state he wanted.
Customer service (at Xilinx I believe) told him to recode his state machine by
putting the next state and output decodes into one clocked process. This was
going to be a tremendous headache for him (it would change pipeline timing,
etc...). After trying the design in Leonardo the CLB count dropped 10-15% I
believe. IE a huge difference.

I use Leonardo primarily because it purports to work for both ASIC and FPGA
technologies. I only have experience using it for Xilinx designs where it works
really well. The TCL scripting language is very usefull for getting ones self
out of tricky implementation problems. I understand that Synplify recently added
TCL support but I don't know how deep it is.

IMHO if you are doing FPGA's only stick with Galeleo (Spectrum Level 2 I think)
or Synplify. If ASIC's are in your future pick Spectrum Level 3.

jerry english wrote:

> I am trying to make an informed decision on which FPGA HDL compiler
> I should purchase. I have evaluated three, FPGA Express (Synopsys),
> Synplify (Synplicity), and Leonardo Spectrum level 3 (Exemplar).
> All three tools produce similar results, all have basically the same
> features, all are priced about the same give or take a few thousand.
> Any comments, positive or negative, about the tool you are using that
> you would care to share would be greatly appreciated. If there is a
> demand
> I could tabulate and post the summary.
>
> Regards
> Jerry English

--

Tim Davis
Timothy Davis Consulting

TimDavis@TDCon.com - +1 (303) 426-0800 - Fax +1 (303) 426-1023

+-------------------------------------------------------------------+
: Engineering is like having an 8 a.m. class and a late afternoon   :
:          lab every day for the rest of your life.                 :
+-------------------------------------------------------------------+


Article: 13407
Subject: Re: Which parts are fastest for 3-state enables?
From: bob elkind <eteam@aracnet.com>
Date: Tue, 01 Dec 1998 10:42:58 -0800
Links: << >>  << T >>  << A >>
Alternatives:

1.  Don't use a tri-state output, use "open collector" instead (assumes
that
signal is active LOW, that de-assertion time isn't nearly as critical as
assertion time).

2.  Use combinatorial logic to assert signal, rather than register
output. If you need
output function to persist for a specific # of clock cycles, the "state
machine" output
term can be (simply) one of the combinatorial input terms.

You have two problems, it seems:

A.  Assert the uP bus interface quickly. This is done best with comb
logic,
without waiting for clock edge or skew or clk=>q delay.

B.  Keeping the signal asserted as long as necessary.  This is done best
with
clocked logic, which feeds into the combinatorial logic.

Muxing between the A. and B. logic is usually a straightforward problem.


Steve wrote:

> I'm interfacing to a Motorola MPC 860 @50Mhz.  I need to provide an
> acknowledge signal in <10ns after the Clock.  To do this my PLD/FPGA
> needs to have a (Clk-Q + comb delay + tristate enable) < 10ns.
>
> So far I haven't found a CPLD to do it.  What small FPGA's have the best
> crack at it???  ... or prove me wrong on the CPLD issue?
>
> So far my only practical solution is an XCS05XL-4.
>
> Comments?
>
> Steve

Bob Elkind  (fpga design consulting, etc. etc.)
eteam@aracnet.com

Article: 13408
Subject: Re: Will XILINX survive?
From: peter299@maroon.tc.umn.edu (Wade D. Peterson)
Date: Tue, 01 Dec 1998 19:03:15 GMT
Links: << >>  << T >>  << A >>
Froilan P Montenegro <froilan+@andrew.cmu.edu> wrote:

>  What about reconfigurable computing and systems-on-a-chip with
>programmable logic on the same die?  Something like a processor core and
>an FPGA on the same chip with some an interface for the two to
>communicate with each other.  That way if you need basic general
>processing abilities along with some custom logic, this might allow you
>to build a smaller, cheaper, faster implementation than current
>processor/FPGA combinations.  I'm no expert but these are two
>possibilities I've heard.
> 
> - Froilan

I think you're right...the reconfigurable logic field is going to be
quite exciting in the future.  From what I've seen, though, it's too
small right now to predict what's going to happen.

If I really polish my crystal ball (it's a bit rusty sometimes) I can
see entire systems made from reconfigurable logic.  The hardware would
be booted in much the same way as operating systems are now.  This is
a shear fantasy at this point, but it isn't beyond the realm of
possibilities.

Wade Peterson
Silicore Corporation



Article: 13409
Subject: Re: Is it normal to have to edit the xnf file???
From: "Steve" <reply.through.newsgroup@paranoid.com>
Date: Tue, 01 Dec 1998 19:10:03 GMT
Links: << >>  << T >>  << A >>
>
>1) What exactly is the error?
>

Support was able to solve it for me.  I had instantiated a BUFGTS
on a 9500XL pin and was also trying to use that pin as an output.

This seemed to confuse FPGA Express.  The solution was to also
instantiate an OBUF on the pin.  ie it looks like once you instantiate
anything on a pin you can't rely on the automatic buffer insertion
for that pin.   ... kind of makes sense...

BTW: the 9500XL parts have much faster tristate enable although
    the internal feedback is still slower than going out and coming
    back in the GTS pin.  (like I was trying to do)


>2) To the best of my knowledge, Virtex parts are the only
>    ones which produce EDIF files.  You cannot get EDIF
>    files for older families.
>


You're right.  The answer I read seemed to indicate you could
choose your output format but you can't.  The rule appears to be
    pre-Virtex parts use xnf
    Virtex use edf



Steve


Article: 13410
Subject: Re: Will XILINX survive?
From: peter299@maroon.tc.umn.edu (Wade D. Peterson)
Date: Tue, 01 Dec 1998 19:15:30 GMT
Links: << >>  << T >>  << A >>
"Mike & Jen" <rebane@enter.net> wrote:

>I agree with Wade's statement:
>>>I cannot see how the continued push towards ever larger FPGAs matches
>>>with what people actually design.

>What is the difference between a 500k/1M gate FPGA design and an ASIC in
>terms of design time and time to market?  Not much that I can see.  An FPGA
>will always be slower and consume more power than the ASIC.   Why re-invent
>the wheel for 75% of a 1M gate FPGA design that uses common
>cores/interfaces/functions in support of, on average 25% of the design for
>actual custom IP requiring re-programmable  logic?  Devices of that size do
>not show me any enhanced value from a designer's perspective.

I agree from a totally technical perspective.  However, the FPGA parts
are more viable in smaller quantities.  A lot of ASIC vendors won't
even talk with you unless you want to buy 10,000 - 20,000 parts.  The
majority of design wins out there are much smaller.  That's where the
FPGAs fit in.  

The FPGAs also cut time-to-market.  We've discovered that in the core
business, the time-to-market is much faster...usually 4-12 weeks (or
more) than an ASIC.


>>2) The system-on-a-chip market (which is in it's infancy now, at least
>>for FPGAs) will grow at a faster pace than the market at large.  These
>>parts require very large densities.

>I agree.  The densities will need to be very large to implement standard
>cores in FPGA logic and still have programmable logic left over to do
>something custom with.  That is why I believe that the only vendor who, to
>date, has actually displayed the ability to merge ASIC cores (dedicated
>silicon) with surrounding FPGA logic is Lucent Technologies (ORCA).  That is
>what I would define as true "systems on a chip"; not just a bigger array
>block.  They seem to have the designers in mind.

I kind of like those ORCA parts.  The Lucent place and route tools
arn't very glitzy, but they seem to be pretty rock solid (as compared
to other vendors).


>>4) FPGA and ASIC vendors will, at some point, become commodity
>>producers of gates.  These companies will complete on delivery, price
>>and (to a smaller extent) performance, but not so much on tools (which
>>will become the realm of third party EDA vendors).

>If just producing larger arrays of the same ( although renamed) architecture
>is the direction X and A are going to take, then FPGAs will become as you
>put it "... become commodity producers of gates."   However, I do not
>believe it will be reduced to quite that simplistic a level (e.g.. DRAMS).
>Just because a lot of gates fill a device does not qualify it as ASICish
>(forgive my new word).  I feel it would be a mistake for the marketing
>departments of the FPGA vendors to promote their larger devices as ASIC
>like.  To do so would insult the intelligence of most designers (I hope).
>FPGAs are used to implement specific system capabilities as do ASICs.
>I believe we are in store for the usual ASIC bidding wars with an FPGA
>twist.  The vendors that can successfully merge the technologies will offer
>true value and will set themselves apart from the traditional FPGA/ASIC
>vendors.
>Mike.

The true commodity FPGA (similar to DRAMs) probabably will never
entirely happen unless somebody standardizes on the guts.  DRAMs are
true commodity parts, but there are some pretty good JEDEC pinout and
timing standards.  FPGAs are becoming more similar, but there is no
true standard architecture yet.

My Christmas wish list this year includes an ANSI/IEEE standard for
standard FPGA arcitectures.  That way all of the application and tool
developers could get some of their sanity back.  That would also
really push FPGAs into the commodity realm.  Imagine 5-10 vendors of
FPGAs all with a similar architecture!

Wade Peterson
Silicore Corporation
http://www.silicore.net/






Article: 13411
Subject: Re: Will XILINX survive?
From: s_clubb@NOSPAMnetcomuk.co.uk (Stuart Clubb)
Date: Tue, 01 Dec 1998 19:38:31 GMT
Links: << >>  << T >>  << A >>
On 30 Nov 1998 21:57:44 GMT, "Austin Franklin"
<dark4room@ix.netcom.com> wrote:

>Valid point, but, er, cough, cough, isn't VHDL supposed to give you that

It does, within reason.

>ability today?  I believe you still needed to buy the vendors tool set for
>each architecture....

If you want to give the sales droid money for a back end P&R tool that
is useless unless you are buying his FPGAs, then that is your choice,
I suppose :-)

I'm sure the nice friendly FAE will be only too happy to sort you out
with P&R tools when you say magic phrases like "I have a Vendor
Independent toolset" and "Actually I could use brand Y for this
project, and they just gave me the P&R"

Stuart
For Email remove "NOSPAM" from the address

Article: 13412
Subject: Re: VHDL simulation of exported schematics..?
From: s_clubb@NOSPAMnetcomuk.co.uk (Stuart Clubb)
Date: Tue, 01 Dec 1998 19:38:35 GMT
Links: << >>  << T >>  << A >>
On Mon, 30 Nov 1998 15:25:12 GMT, thor@sm.luth.se.NoSpam (Jonas Thor)
wrote:

>It should be possible to drive a global signal in VHDL testbench...?
>Any ideas?

I think (but I haven't tested it)...

>From the M1.5 Design Manager menu:

Design -> Implement

Options

Now set the Simulation Template to "ModelSim VHDL" and Edit the
associated template.

Select the VHDL/Verilog tab and check the tick box to bring the GSR
out as a port.

I imagine that should do the trick.

Cheers
Stuart 
For Email remove "NOSPAM" from the address

Article: 13413
Subject: Re: Will XILINX survive?
From: Peter Alfke <peter@xilinx.com>
Date: Tue, 01 Dec 1998 13:16:36 -0800
Links: << >>  << T >>  << A >>
Wade D. Peterson wrote:

> My Christmas wish list this year includes an ANSI/IEEE standard for
> standard FPGA arcitectures.  That way all of the application and tool
> developers could get some of their sanity back.  That would also
> really push FPGAs into the commodity realm.  Imagine 5-10 vendors of
> FPGAs all with a similar architecture!
>  

Don't hold your breath. It worked for DRAMs because they really  have a
simple, narrowly defined function, ( at least until recently ). And
their business has turned into a disaster.
FPGAs must stay more diverse, and we need competitive innovation to
advance the state of the art. If that leads to some confusion and even
chaos, so be it.

Just look at the battle between Altera and Xilinx. It has been the
driving force that brought you better, bigger and faster devices at an
incredibly fast pace.

Americans like competition, and everybody wants to win.

Standardization would kill innovation and turn these FPGA houses into
drab commodity factories that compete on nothing but price.

Not my idea of fun.

Peter Alfke

Article: 13414
Subject: Re: HELP, Tool selection
From: "Ken Coffman" <kcoffman@intermec.com>
Date: Tue, 1 Dec 1998 13:48:05 -0800
Links: << >>  << T >>  << A >>
The last time I looked at FPGA Express it was not very good. I got faster
and better out of the box performance from Synplicity, but Exemplar is
better if you like tinkering under the hood. You can't go wrong with either
product. Exemplar got out first with their ASIC support, but its coming soon
from Synplicity.

jerry english wrote in message <36640165.A3CCB901@planetc.com>...
>I am trying to make an informed decision on which FPGA HDL compiler
>I should purchase. I have evaluated three, FPGA Express (Synopsys),
>Synplify (Synplicity), and Leonardo Spectrum level 3 (Exemplar).
>All three tools produce similar results, all have basically the same
>features, all are priced about the same give or take a few thousand.
>Any comments, positive or negative, about the tool you are using that
>you would care to share would be greatly appreciated. If there is a
>demand
>I could tabulate and post the summary.
>
>Regards
>Jerry English
>
>
>


Article: 13415
Subject: Re: Is it normal to have to edit the xnf file???
From: "Austin Franklin" <dark4room@ix.netcom.com>
Date: 1 Dec 1998 22:42:54 GMT
Links: << >>  << T >>  << A >>
> 2) To the best of my knowledge, Virtex parts are the only
>     ones which produce EDIF files.  You cannot get EDIF
>     files for older families.

I believe I'm using EDIF for my Spartan parts now.....

Austin

Article: 13416
Subject: XILINX FPGA reaches GHz speeds
From: "EKC" <alpha3.1@ix.netcom.com>
Date: Tue, 1 Dec 1998 18:37:54 -0500
Links: << >>  << T >>  << A >>
ELECTRONIC ENGINEERING TIMES via NewsEdge Corporation : San Jose, Calif. -
Xilinx Inc. claims to have reached GHz speeds with a prototype FPGA,
attributing the feat to the power of pure-play foundries.

http://www.eet.com/story/OEG19981130S0011


-EKC


Article: 13417
Subject: Re: XILINX FPGA reaches GHz speeds
From: jhallen@world.std.com (Joseph H Allen)
Date: Wed, 2 Dec 1998 02:18:41 GMT
Links: << >>  << T >>  << A >>
In article <741uk7$oia@dfw-ixnews9.ix.netcom.com>,
EKC <alpha3.1@ix.netcom.com> wrote:
>ELECTRONIC ENGINEERING TIMES via NewsEdge Corporation : San Jose, Calif. -
>Xilinx Inc. claims to have reached GHz speeds with a prototype FPGA,
>attributing the feat to the power of pure-play foundries.
>
>http://www.eet.com/story/OEG19981130S0011

And are they planning on releasing the GHz speed existing 4000-series FPGA? 
No!  Instead they are going to release a large, but much slower new FPGA
with this technology.  Oh well.

-- 
/*  jhallen@world.std.com (192.74.137.5) */               /* Joseph H. Allen */
int a[1817];main(z,p,q,r){for(p=80;q+p-80;p-=2*a[p])for(z=9;z--;)q=3&(r=time(0)
+r*57)/7,q=q?q-1?q-2?1-p%79?-1:0:p%79-77?1:0:p<1659?79:0:p>158?-79:0,q?!a[p+q*2
]?a[p+=a[p+=q]=q]=q:0:0;for(;q++-1817;)printf(q%79?"%c":"%c\n"," #"[!a[q-1]]);}

Article: 13418
Subject: Interfaces to an Asynchronous SRAM
From: lzhang@eecg.toronto.edu (Louis Zhang)
Date: 2 Dec 98 05:29:03 GMT
Links: << >>  << T >>  << A >>

Hi,

We are try to build an Xilinx FPGA to read and write to an asynchronous 
SRAM (due to the design requirement, we can only use asynchronous SRAM, 
not synchronous SRAM nor ZBT SRAM)

The problem of the asynchronous SRAM is with its write operation, where
the write enable (R/Wn) has to be asserted ('0') "shortly" after the 
address (and data) is stable, and deasserted ('1') "shortly" before the 
address (and data) changes.  In other words, R/Wn has to be deasserted
('1') when address or data change.


The desired waveform looks like this:

CLK= 30MHz

          _______	  _______	  ________        ________	
CLK      |       |       |       |	 |        |      |        |		
     ____|       |_______|       |_______|        |______|        |_____

R/Wn _______ 	       ______           _______________________________ 
	    |	       |     |	       |		     
	    |__________|     |_________|                      


Addr ------------------------------------------------------------------- 
      RA0 |     WA0        |    WA1        |	 RA1        |   RA2
	  |		   |		   |		    |
     -------------------------------------------------------------------	   

A programmable clock can be used to "mask" the original R/Wn (coming out
of a flip flop) to form the above desired waveform.  Unfortunately, 
programmable clock is also not an option for us right now, even though our 
CLK is only 30MHz .  

I was wondering if we can just use ONE clock (which is the system clock for 
Address, Data, etc.) to work around with the strict write requirement of the 
asynchronous SRAM.

Thanks for your help in advance.


-----------------------
Louis Zhang
lzhang@eecg.toronto.edu

Article: 13419
Subject: Re: VHDL simulation of exported schematics..?
From: thor@sm.luth.se.NoSpam (Jonas Thor)
Date: Wed, 02 Dec 1998 08:53:41 GMT
Links: << >>  << T >>  << A >>


On Tue, 01 Dec 1998 19:38:35 GMT, s_clubb@NOSPAMnetcomuk.co.uk (Stuart
Clubb) wrote:

>From the M1.5 Design Manager menu:
>
>Design -> Implement
>
>Options
>
>Now set the Simulation Template to "ModelSim VHDL" and Edit the
>associated template.
>
>Select the VHDL/Verilog tab and check the tick box to bring the GSR
>out as a port.
>
>I imagine that should do the trick.

Yep, for a back-end timing simulation it would. But I really don't
like timing simulation since I beleive the static timing analysis for
a synchronous design will cover the timing.

What I am doing here is a front-end simulation with the unisim
library, so I don't think this will help.

Thanks,

Jonas Thor

Article: 13420
Subject: Re: Is it normal to have to edit the xnf file???
From: "ez" <ez@ez>
Date: Wed, 2 Dec 1998 10:11:03 +0100
Links: << >>  << T >>  << A >>
>They said that editing the xnf to fix the error was the best
>solution.  Is this normal?  There's got to be a better way.
>
At least Xilinx is consistent! We have been told exactly the
same thing. The busses that FPGA Express outputs use
the A<i> syntax, now, if we want to instantiate this from
a Viewlogic schematics, this simply does not work
together! Their option to change the output format still
doesn't work in v2.0.3. It would take me 30 secs to fix
this in their code, I don't know what they're waiting for,
but we had to write our own script to patch their mess.
Well,we have been writing scripts for some time now
to patch whatever simple things Xilinx tools screw up.
On the same track, I am still wondering why I have to
click 9 things to create an .XNF file.

1) Click design source tree to expand tree
2) Click to select 'checked?' design
3) Click to create implementation
4) Click OK to accept settings
5) Click implemented design in chips tree to select
6) Click to optimise
7) Click optimised design in chips tree to select
8) Click to export netlist
9) Click to accept the default directory

Did the people that created FPGA Express ever used their
own tools ????? It seems they like to click their mouse all
the time, yes, when updating a design that contains errors,
you have to click another design first and then click the
updated design to see the errors. Oh, well, meanwhile you
can always entertain yourself watching their beautiful
'Internet Explorer copy' flashing lamp to keep yourself from
getting nervous.











Article: 13421
Subject: NEW ENGINEERING PAGE: Please Visit
From: metad@globalnet.co.uk (Scott Paul Johnston)
Date: Wed, 02 Dec 1998 13:58:59 GMT
Links: << >>  << T >>  << A >>
Please visit and comment on my Electronics and Electrical Engineering
pages located at:

http://www.users.globalnet.co.uk/~metad/eee.htm

Containing:
Introduction to EEE
Resources (over 100 web links)
Employment Statistics and newspaper excerpts
Engineering Poems, Quotations and Jokes
EEE at Glasgow University

In addition my homepage (http://www.users.globalnet.co.uk/~metad/)
contains:

A section about me
My CV
A James Bond Section
A guestbook
Humour
500+ cool links in the "new look" bookpage
Cool background MIDI and graphics
Literary quotations
Photo Album
Student Resources
Awards Page
Poems...

Basically, something for everyone!

PLEASE VISIT VIA MY MAIN HOMEPAGE ADDRESS!

Please send you comments via the guestbook or by Email (containing
your full name and Email and webpage addresses) and visit via
http://www.users.globalnet.co.uk/~metad/.

Thanks
Scott Johnston
metad@globalnet.co.uk   

Article: 13422
Subject: Re: XILINX FPGA reaches GHz speeds
From: "Steve" <reply.through.newsgroup@paranoid.com>
Date: Wed, 02 Dec 1998 14:42:20 GMT
Links: << >>  << T >>  << A >>
Let's see ... if that's an XC4000 ...
    I guess it would have to be a -001
    or would that be a -0001?

EKC wrote in message <741uk7$oia@dfw-ixnews9.ix.netcom.com>...
>ELECTRONIC ENGINEERING TIMES via NewsEdge Corporation : San Jose, Calif. -
>Xilinx Inc. claims to have reached GHz speeds with a prototype FPGA,
>attributing the feat to the power of pure-play foundries.
>
>http://www.eet.com/story/OEG19981130S0011
>
>
>-EKC
>
>


Article: 13423
Subject: Re: Is it normal to have to edit the xnf file???
From: "Steve" <reply.through.newsgroup@paranoid.com>
Date: Wed, 02 Dec 1998 14:52:02 GMT
Links: << >>  << T >>  << A >>

ez wrote in message <7430d0$8l3@miura.gent.bg.barco.com>...
>>They said that editing the xnf to fix the error was the best
>>solution.  Is this normal?  There's got to be a better way.
>>
>At least Xilinx is consistent!

A small point here.  I think you're complaining about the wrong
company (this time).  My understanding is that Synopsis still controls
the code and Xilinx only resells it.

Granted the FPGA Express user interface is kludgey ... especially
their editor, but for me I live with that because the quality of the
synthesis for the price (base VHDL) is unrivalled.

I do CPLD's and small FPGA's and I use the Altera and Xilinx
tools.  FPGA Express does a much better job than the Altera
VHDL,  and some day when they let Xilinx sell the newer version
we'll even get VHDL 93 compliance :)


Steve


Article: 13424
Subject: Re: Will XILINX survive?
From: Brian Boorman <XZY.bboorman@harris.com>
Date: Wed, 02 Dec 1998 10:20:09 -0500
Links: << >>  << T >>  << A >>


Austin Franklin wrote:

> (nipping the "problem"
> > company Neocad in the bud)
>
> I disagree that NeoCAD was a problem.  They (again) created a perception of
> making a better router.  Well, it really was only better under certain
> circumstances....but when used with a design that was floorplanned, it was
> NO better, in fact, sometimes worse.  They designed it to handle regular
> structures, and the Xilinx router did not handle them at all.
>
> All the rest of the NeoCAD tools were not really very good (say, the EPIC
> editor.....).  The NeoCAD 'vision' was to make a universal set of tools for
> all programmable logic, NOT just 'another' tool for the Xilinx parts.  I
> don't believe their 'vision' was ever going to materialize, and the
> acquisition of NeoCAD by Xilinx was most fortunate for NeoCAD.
>

Perhaps. But the NeoCAD tools also became the basis for the Lucent tools. I
have done several Lucent designs in the past (and present), and when I had to
do a full Xilinx 4028 design this summer in only 2 months, the commonality of
the Lucent tools I was used to and the Xilinx M1.4 tools was a big factor in
the successfull completion. Never having used to old Xilinx tools, I cannot
argue any points/opinions people have about new vs. old. However, the common
interface was a godsend for me. Also, it has been my case (having to complete
designs for people who left) that how good the tools operate for you is
directly related to how well you design the circuitry you want to fit. Poor
design practices makes for poor routing/timing in the final design.

Last, as buggy as Epic is at times (both the Lucent and Xilinx versions), the
capability to go in and make tweaks during debug/integration is very helpful.
The key is to save after every change. I worked with Altera MaxPlus+ for two
years and this was something of a sore point with their tools (at least up to
version 8.3, I don't know what they have now).

> I believe it was a most unfortunate move for the Xilinx users, since, now,
> we are stuck with this set of inferior tools, at least for the foreseeable
> future....
>

--
Brian C. Boorman
Harris RF Communications
Rochester, NY 14610
XYZ.bboorman@harris.com
<Remove the XYZ. for valid address>




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