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 20950

Article: 20950
Subject: Recommended VHDL titles wanted ...
From: "Number Cruncher" <michel.heuts@pandora.be>
Date: Tue, 29 Feb 2000 21:14:53 GMT
Links: << >>  << T >>  << A >>
Workgroup MCU & UPL
Belgium


Hello guys,

                    We are doing some studywork and need titles about
recommended VHDL books.
So if you have read certain books and can recommend it give me a sign ...

regards,

Michel
WMU Belgium


Article: 20951
Subject: Re: Asynchronous flipflops in Cypress CPLDs with Warp VHDL
From: "Karl Olsen" <kro@post3.tele.dk>
Date: Tue, 29 Feb 2000 23:26:33 +0100
Links: << >>  << T >>  << A >>
KJ <Kevin.Jennings@Unisys.com> wrote in message
news:89gjav$sek$1@mail.pl.unisys.com...

> No you're not paranoid, you can't be sure that cpua_data(7) is properly
> latched.  Transparent latches always need the redundant cover term when
> implemented as combinatorial logic (i.e. without benefit of some
predefined
> latch in the hardware architecture).  I've seen this failure with several
> programmable devices.
>
> When I used Warp I couldn't find a way to have it keep the necessary
> redundant terms so unless the part you are targetting can guarantee that a
> latch will be implemeted properly I wouldn't suggest trying it.  If it
does,
> you might want to change the logic somewhat like...
>
> latch_enable <= not(cpua_wr_not) or not(cpub_wr_not);    -- Or together
the
> writes to create a latch enable
>
> Make your process sensitive to latch_enable instead of cpua_wr_not and
> cpub_wr_not and then use one of the write signals to select the data.
>
> process (reset, latch_enable, cpua_data, cpub_data)
> begin
>   if reset = '1' then
>     outport <= "00000000";
>   elsif latch_enable = '1' then
>     if (cpua_wr_not = '0') then
>         outport <= cpua_data;
>     else
>         outport <= cpub_data;
>     end if;
>   end if;
>
> [...]
>
> If that still gives a single combinatorial equation then either...
> - Pick another part that knows about latches.
> - Rethink the real reason for your async requirement.
> - Talk to the Cypress rep about how to keep redundant terms.  Like I said
> when I played with it, the VHDL signal attributes that they defined didn't
> include any sort of "don't touch this


Hi Kevin,

My part (CY37064) does support latches, and Warp uses them when I rewrite it
like in your example.  But won't the hold time be badly violated, since the
latch_enable signal is delayed by an extra pass through a macrocell?  E.g.
at the end of a CPU A write cycle, the "if (cpua_wr_not = '0')" disappears
"immediately" since it is implemented directly in product terms in front of
the latch, but the latch_enable edge comes one pass later.

I would prefer an asynchronous solution since I have some sensitive analog
stuff nearby, and I am worried by the noise from 28 flipflops running at
14 MHz.  (Should I be that?)  But since it is so non-trivial to get an
async solution right, I will go for a synchronous circuit like this, which
should be safe.  It delays the write strobes two clocks so that it can
latch the data at the clock edge that comes between one and two clock
periods after the falling edge of the real cpuX_wr_not.

Thanks a lot for your advice.

Regards,
Karl Olsen



library ieee;
use ieee.std_logic_1164.all;

entity sync is
  port (
    reset       : in std_logic;
    clock       : in std_logic;
    cpua_wr_not : in std_logic;
    cpua_data   : in std_logic_vector (7 downto 0);
    cpub_wr_not : in std_logic;
    cpub_data   : in std_logic_vector (7 downto 0);
    outport     : out std_logic_vector (7 downto 0)
  );
end sync;

architecture sync_arch of sync is

signal cpua_wr_not1 : std_logic;
signal cpua_wr_not2 : std_logic;
signal cpub_wr_not1 : std_logic;
signal cpub_wr_not2 : std_logic;

begin

process (reset, clock)
begin
  if reset = '1' then
    cpua_wr_not1 <= '1';
    cpua_wr_not2 <= '1';
    cpub_wr_not1 <= '1';
    cpub_wr_not2 <= '1';
  elsif rising_edge(clock) then
    cpua_wr_not1 <= cpua_wr_not;
    cpua_wr_not2 <= cpua_wr_not1;
    cpub_wr_not1 <= cpub_wr_not;
    cpub_wr_not2 <= cpub_wr_not1;
  end if;
end process;

process (reset, clock)
begin
  if reset = '1' then
    outport <= "00000000";
  elsif rising_edge(clock) then
    if cpua_wr_not1 = '0' and cpua_wr_not2 = '1' then
      outport <= cpua_data;
    elsif cpub_wr_not1 = '0' and cpub_wr_not2 = '1' then
      outport <= cpub_data;
    end if;
  end if;
end process;

end sync_arch;


Article: 20952
Subject: Re: Recommended VHDL titles wanted ...
From: jonathan@oxfordbromley.u-net.com (Jonathan Bromley)
Date: Tue, 29 Feb 2000 22:41:48 GMT
Links: << >>  << T >>  << A >>
On Tue, 29 Feb 2000 21:14:53 GMT, "Number Cruncher"
<michel.heuts@pandora.be> wrote:


>        We are doing some studywork and need titles about
>recommended VHDL books.
>So if you have read certain books and can recommend it give me a sign ...

I very much like "VHDL for Logic Synthesis" by Andrew RUSHTON (2nd ed)
which is reasonably cheap (by the standards of VHDL books) and is
not widely recommended.

"A designer's guide to VHDL" by Peter ASHENDEN gets the thumbs-up
from many people, but I haven't read it myself.

Jonathan Bromley

Article: 20953
Subject: Re: Extremely fault tolerant strategies
From: janm@penfold.transactionsite.com (Jan Mikkelsen)
Date: 29 Feb 2000 22:47:59 GMT
Links: << >>  << T >>  << A >>
A few books:

Transaction Processing: Concepts & Techniques
By Gray & Reuter
Morgan Kaufmann Publishing

Reliable Computer Systems: Design and Evaluation, Third Edition 
By Siewiorek & Swarz

Atomic Transactions: In Concurrent and Distributed Systems
by Lynch, Merritt & Weihl
Morgan Kaufmann Publishing

Fault Tolerance in Distributed Systems
By Pankaj Jalote

Tandem have (or had?) a thing called Tandem Information Manager (TIM) where you
could get the Tandem documentation on CD for around $150.  That might also be
useful.

Jan Mikkelsen
janm@transactionsite.com
Article: 20954
Subject: Re: Recommended VHDL titles wanted ...
From: rk <stellare@nospam.erols.com>
Date: Tue, 29 Feb 2000 18:16:59 -0500
Links: << >>  << T >>  << A >>
Jonathan Bromley wrote:

> On Tue, 29 Feb 2000 21:14:53 GMT, "Number Cruncher"
> <michel.heuts@pandora.be> wrote:
>
> >        We are doing some studywork and need titles about
> >recommended VHDL books.
> >So if you have read certain books and can recommend it give me a sign ...
>
> I very much like "VHDL for Logic Synthesis" by Andrew RUSHTON (2nd ed)
> which is reasonably cheap (by the standards of VHDL books) and is
> not widely recommended.
>
> "A designer's guide to VHDL" by Peter ASHENDEN gets the thumbs-up
> from many people, but I haven't read it myself.

_HDL Chip Design_ by Smith gets two thumbs up. :-)

rk

Article: 20955
Subject: Xilinx Tools Vs Altera tools
From: wamsi@my-deja.com
Date: Tue, 29 Feb 2000 23:37:43 GMT
Links: << >>  << T >>  << A >>
How does Quartus for APEX compare with the Xilinx tools version for Virtex.
Principally I am interested in post Synopsys features.
1. Timing analysis reports
2. Fitter/Router
3. any experience with compile time for 1mil gate designs
4. timing driven compilation
5. Incremental compile
6. Testability tools

Thanks
-Wamsi


Sent via Deja.com http://www.deja.com/
Before you buy.
Article: 20956
Subject: AMS board simple questions
From: javidiaz@my-deja.com
Date: Wed, 01 Mar 2000 02:08:23 GMT
Links: << >>  << T >>  << A >>
Hello guys,

I have a few questions to you.

First, I use the following components:
1. Active-HDL from Aldec
2. Xilinx Foundation Series 1.5
3. Wild-one AMS board
4. Visual C++ and Borland C++

I have been trying to do a synthesis for one of the example project
provided. BUT  I got the following message:
ELBREAD: Warning:
Component /GEN_BOARDS/WF1BOARD/U_PE0/U_IF/U_PADIntAck :
IPAD not bound.

I have noticed that it is the only Component (IPAD) without any
Architectural description on ams4ksim.vhd package file.What you should
I do? (do a dataflow 'bypass', i.e., just assign the output equal to
the input)

My other doubt is: can I use the DPMCs ports as direct communication
ports between PEs themselve and the host; given I don't have any
mezzanne memory card; Or those ports are wasted?

Thanks,
Javier


Sent via Deja.com http://www.deja.com/
Before you buy.
Article: 20957
Subject: Re: Delay Lines using FPGA ??
From: Phil Hays <Spampostmaster@sprynet.com>
Date: Tue, 29 Feb 2000 21:02:13 -0800
Links: << >>  << T >>  << A >>
henry wrote:

> i was wondering if a delay line could be made easily in an fpga and if so
> are there any
> limitations ie. max delay, max frequency.

The main limitation is accuracy.  A range exceeding three to one should be
expected over process, temperature and voltage.  Example: if the longest delay
is 30 ns, the shortest would be less than 10 ns.  

A better way to handle the same problem is a digital delay of some sort. 
Perhaps a high frequency clock runs a counter for the number of cycles that
makes up the delay.


-- 
Phil Hays

Article: 20958
Subject: Re: Book recommendations?
From: Edwin Naroska <edwin@ds.e-technik.uni-dortmund.de>
Date: Wed, 01 Mar 2000 10:58:26 +0100
Links: << >>  << T >>  << A >>
Hi,

Take a look at part 2 of the VHDL FAQ
(http://www.vhdl.org/comp.lang.vhdl/).
Section 2.6 lists some recommended books.

--
Edwin



Article: 20959
Subject: help me!...please...
From: "YEUNG-CHUEL" <stevens@ieee.org>
Date: Wed, 1 Mar 2000 19:47:41 +0900
Links: << >>  << T >>  << A >>
I am studying VHDL with Maxplus2 student 9.24 of Altera. but I can't design
RAM and

ROM. Although I don't have drive F:, Maxplus2 occurs error that is "Error:
Line 5: File

f:\maxw\code\exew\vhdl87\std\textio.vhd: Unsupported feature error: access

type is not supported"

Why? I don't understand.

Please correct my error.

below is my source.

---------------------------------------------------------------------------
library   IEEE;
use    IEEE.std_logic_1164.all;
use    IEEE.std_logic_unsigned.all;


entity  myROM  is
  port( ROM_ADDR : in ieee.std_logic_1164.std_logic_vector(4 downto 0);
    ROM_EN  : in ieee.std_logic_1164.std_logic;
    ROM_DATA : out ieee.std_logic_1164.std_logic_vector(7 downto 0) );
end myROM;

architecture behavior  of  myROM  is
COMPONENT asyn_rom_32x8
-- pragma translate_off
     GENERIC (LPM_FILE : string);
-- pragma translate_on
     PORT (Address : IN ieee.std_logic_1164.STD_LOGIC_VECTOR(4 DOWNTO 0);
           MemEnab : IN ieee.std_logic_1164.STD_LOGIC;
           Q       : OUT ieee.std_logic_1164.STD_LOGIC_VECTOR(7 DOWNTO 0)
     );
END COMPONENT;


BEGIN

   rom: asyn_rom_32x8
-- pragma translate_off
        GENERIC MAP (LPM_FILE => "myrom.mif")
-- pragma translate_on
   PORT MAP (Address => ROM_ADDR(4 downto 0), MemEnab => ROM_EN, Q =>
ROM_DATA(7 downto 0));

END behavior;


Article: 20960
Subject: Re: Xilinx Tools Vs Altera tools
From: Georg Berliner <gbe@pe.dk>
Date: Wed, 01 Mar 2000 11:52:23 +0100
Links: << >>  << T >>  << A >>
On Tue, 29 Feb 2000 23:37:43 GMT, wamsi@my-deja.com wrote:

>How does Quartus for APEX compare with the Xilinx tools version for Virtex.
>Principally I am interested in post Synopsys features.
>1. Timing analysis reports
>2. Fitter/Router
>3. any experience with compile time for 1mil gate designs
>4. timing driven compilation
>5. Incremental compile
>6. Testability tools
>
>Thanks
>-Wamsi
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.


The new version of of Quartus works OK now
There are some bugs with runtime errors  But there is a workarround 

>Principally I am interested in post Synopsys features.
I am using  Mentor Leonardo as front end. it works fine . There are
however some I/F mismatches between Leonardo and Apex resulting in
protection errors , which can be avoided by using some work arround


The Xilinx 2.1i  has a simple interface : 16-bit look alike but it is
apparently working fine and it does the job.

>1. Timing analysis reports
Is fine in both Xilinx and  Altera

>2. Fitter/Router
works Fine  on both of them

>3. any experience with compile time for 1mil gate designs
I only managed to fill 50% of Apex  400 K gates Chip

As far as know Altera and Xilinx gate counting method  can not be
directly compared. 
Roughly one must divide the Xilinx gate count by factor 2 to get the
same  gate count.

>5. Incremental compile

It will come in the future

>6. Testability tools
Altera has a IP for logic analyzer working via JTAG

Georg Berliner






Article: 20961
Subject: Re: Xilinx Tools Vs Altera tools
From: "David Hawke" <dhawke@skynow.net>
Date: Wed, 1 Mar 2000 12:29:44 -0000
Links: << >>  << T >>  << A >>
See my comments below

Georg Berliner wrote in message ...
>On Tue, 29 Feb 2000 23:37:43 GMT, wamsi@my-deja.com wrote:
>
>>How does Quartus for APEX compare with the Xilinx tools version for
Virtex.
>>Principally I am interested in post Synopsys features.
>>1. Timing analysis reports
>>2. Fitter/Router
>>3. any experience with compile time for 1mil gate designs
>>4. timing driven compilation
>>5. Incremental compile
>>6. Testability tools
>>
>>Thanks
>>-Wamsi
>>
>>
>>Sent via Deja.com http://www.deja.com/
>>Before you buy.
>
>
>The new version of of Quartus works OK now
>There are some bugs with runtime errors  But there is a workarround
>
>>Principally I am interested in post Synopsys features.
>I am using  Mentor Leonardo as front end. it works fine . There are
>however some I/F mismatches between Leonardo and Apex resulting in
>protection errors , which can be avoided by using some work arround
>
>
>The Xilinx 2.1i  has a simple interface : 16-bit look alike but it is
>apparently working fine and it does the job.
>
>>1. Timing analysis reports
>Is fine in both Xilinx and  Altera

Xilinx offers very good timing driven compile syntax, and accepts timing
information from
all 3 main Synthesis Vendors, namely Synopsys, Leonardo and Synplify
>
>>2. Fitter/Router
>works Fine  on both of them
>
>>3. any experience with compile time for 1mil gate designs
>I only managed to fill 50% of Apex  400 K gates Chip
I have designs that are using upto 99% of V1000's and routing and running at
over 100MHz
>
>As far as know Altera and Xilinx gate counting method  can not be
>directly compared.
>Roughly one must divide the Xilinx gate count by factor 2 to get the
>same  gate count.
>
Not quite, an Apex 20K400E is equivalent to a V600(E). For rough guidlines
use the Logic Cell count,
but this still doesn't take into account things like the Mult-And, and
SRL16's which are VERY useful, and are
again automatically inferred by the systhesis tools.

>>5. Incremental compile
>
>It will come in the future

Xilinx 2.1i already has the cabability and has always had it. In fact
Exemplar have an app note that details
how you can use it with Leonardo.

>
>>6. Testability tools
>Altera has a IP for logic analyzer working via JTAG
Xilinx has the Probe functionality in FPGA Editor, and has JTAG test tools.
Probe is far more usable, as
it doesn't have to be inserted into the design, and therfore altering the
desing timing, and can be manipulated
and re-downloaded in seconds....


My 2 cents...



Article: 20962
Subject: Re: AMS board simple questions
From: "Simon" <simonb@tile.demon.co.cuthis.uk>
Date: Wed, 1 Mar 2000 12:33:31 -0000
Links: << >>  << T >>  << A >>
javidiaz@my-deja.com wrote in message <89hu2l$om6$1@nnrp1.deja.com>...
>
>I have been trying to do a synthesis for one of the example project
>provided. BUT  I got the following message:
>ELBREAD: Warning:
>Component /GEN_BOARDS/WF1BOARD/U_PE0/U_IF/U_PADIntAck :
>IPAD not bound.
>
>I have noticed that it is the only Component (IPAD) without any
>Architectural description on ams4ksim.vhd package file.What you should
>I do? (do a dataflow 'bypass', i.e., just assign the output equal to
>the input)

Try defining IPAD for synthesis by adding this to your code:

-- synthesis translate_off
library unisim;
use unisim.vcomponents.all;
-- synthesis translate_on



Article: 20963
Subject: Re: Xilinx Tools Vs Altera tools
From: wamsi@my-deja.com
Date: Wed, 01 Mar 2000 13:54:42 GMT
Links: << >>  << T >>  << A >>


> >>2. Fitter/Router
> >works Fine  on both of them
> >

So you mean Quartus fixed the bugs which plagued MAXPLUS-2 for timing
driven compilation? I could never get even a 10K40 design to route using
timing driven compilation...it would just go on for hours at no end!

> >>3. any experience with compile time for 1mil gate designs
> >I only managed to fill 50% of Apex  400 K gates Chip
> I have designs that are using upto 99% of V1000's and routing and
running at
> over 100MHz
> >

Any experiance with APEX at over 130MHz speeds? How well do the Altera
-3 speed grade compare with the Xilinx -6? I am told the Xilinx -6 can
do a 16x16 multiply at 175MHz ... any such benchmarks for altera?


> >As far as know Altera and Xilinx gate counting method  can not be
> >directly compared.
> >Roughly one must divide the Xilinx gate count by factor 2 to get the
> >same  gate count.
> >
> Not quite, an Apex 20K400E is equivalent to a V600(E). For rough
guidlines
> use the Logic Cell count,
> but this still doesn't take into account things like the Mult-And, and
> SRL16's which are VERY useful, and are
> again automatically inferred by the systhesis tools.
>

But does the 20K not have the ESB feature in it which can be used either
as RAM or as combinatorial logic? By using it as combinatorial logic you
can get a lot more gates than the equivalent Xilinx part.

> >>5. Incremental compile
> >
> >It will come in the future
>
> Xilinx 2.1i already has the cabability and has always had it. In fact
> Exemplar have an app note that details
> how you can use it with Leonardo.
>
> >

Are there ways to use the incremental compile in the
verilog->synopsys->fpga tools design flow?


> >>6. Testability tools
> >Altera has a IP for logic analyzer working via JTAG
> Xilinx has the Probe functionality in FPGA Editor, and has JTAG test
tools.
> Probe is far more usable, as
> it doesn't have to be inserted into the design, and therfore altering
the
> desing timing, and can be manipulated
> and re-downloaded in seconds....
>

Do you mean that Xilinx actually has the testability function embedded
in the chip and it just needs to be configured?

-Wamsi


Sent via Deja.com http://www.deja.com/
Before you buy.
Article: 20964
Subject: Re: Materials on PCI
From: Mark <M.Korsloot@computer.removethis.org>
Date: Wed, 01 Mar 2000 15:41:19 +0100
Links: << >>  << T >>  << A >>
The best book (as far as I know) is:

PCI System Architecture (fourth edition !!)
Mindshare Inc (Tom Shanley, Don Anderson)
ISBN 0-201-30974-2
Available through Amazon (usually stock).

Mark

keshav wrote:
> 
> Hi all,
>     I would like to know which book is the Best for PCI.
> 
> Thanks in advance,
> Ken
> 
> * Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
> The fastest and easiest way to search and participate in Usenet - Free!

Article: 20965
Subject: Re: AMS board simple questions
From: Ray Andraka <randraka@ids.net>
Date: Wed, 01 Mar 2000 14:44:11 GMT
Links: << >>  << T >>  << A >>


javidiaz@my-deja.com wrote:

> Hello guys,
>
> I have a few questions to you.
>
> First, I use the following components:
> 1. Active-HDL from Aldec
> 2. Xilinx Foundation Series 1.5
> 3. Wild-one AMS board
> 4. Visual C++ and Borland C++
>

I haven't used the wild-one, but I can make these observations based on
the wildforce and wildstar boards:

Simulation is a whole lot easier using Modelsim because all the AMS macros
have been set up for modelsim.  For use with Aldec, you pretty much have
to re-write all the macros.  After spending too much time trying to get
the board level sim running on Aldec I gave up and went to modelsim, where
it pretty much ran the first time.

Annapolis obscures the VHDL code with all kinds of cute things.  That
makes the translation to aldec a pain.  One area Aldec seemed to have
particular problems was with the entity and architecture definitions being
in separate files in separate directories (Annapolis does that too much).
I suspect that is the root of the warning you are getting.

On the wildstar board, the DPMCs ports can be used as connections between
the PEs if the mezzanine memories are not installed (or installed but not
needed).  To do this you use the SHUNT definitions of the port.

>
> I have been trying to do a synthesis for one of the example project
> provided. BUT  I got the following message:
> ELBREAD: Warning:
> Component /GEN_BOARDS/WF1BOARD/U_PE0/U_IF/U_PADIntAck :
> IPAD not bound.
>
> I have noticed that it is the only Component (IPAD) without any
> Architectural description on ams4ksim.vhd package file.What you should
> I do? (do a dataflow 'bypass', i.e., just assign the output equal to
> the input)
>
> My other doubt is: can I use the DPMCs ports as direct communication
> ports between PEs themselve and the host; given I don't have any
> mezzanne memory card; Or those ports are wasted?
>
> Thanks,
> Javier
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.

--
-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: 20966
Subject: Re: Materials on PCI
From: Mark <M.Korsloot@computer.removethis.org>
Date: Wed, 01 Mar 2000 15:44:27 +0100
Links: << >>  << T >>  << A >>
The best book (as far as I know) is:

PCI System Architecture (fourth edition !!)
Mindshare Inc (Tom Shanley, Don Anderson)
ISBN 0-201-30974-2
Available through Amazon (usually stock).

Mark

keshav wrote:
> 
> Hi all,
>     I would like to know which book is the Best for PCI.
> 
> Thanks in advance,
> Ken
> 
> * Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
> The fastest and easiest way to search and participate in Usenet - Free!
Article: 20967
Subject: Error in Xilinx application note XAPP131?
From: Steve Charlwood <s.m.charlwood@bham.ac.uk>
Date: Wed, 01 Mar 2000 16:14:11 +0000
Links: << >>  << T >>  << A >>
Hi all,

I'm using a FIFO design based on Xilinx's synchronous FIFO with
independent clocks, described in XAPP131. In this, it states that in the
worst-case scenario (due to the asynchronous nature of the READ and
WRITE clocks), the FULL and EMPTY flags would simply stay active one
cycle longer and that this would _not_ generate an error.

I have a simulation in which EMPTY goes low, indicating that there is
data to be read from the FIFO, and then, due to the clock edges from the
two different domains coinciding, the signal is forced into an unknown
state, which could be high or low. In the case where the signal is
actually interpreted as low, the FIFO would indicate that it was not
empty, when in fact it might be. This causes a problem when trying to
read data from the FIFO using the EMPTY signal as the read enable: for
each occurence, and additional word may be inserted into the data
stream.

Does anyone know of a solution to this problem (apart from avoiding the
use of multiple clock domains!).

Cheers,

Steve

 
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ 

  Heterogeneous, Parallel & Reconfigurable Architectures Group 
  School of Electronic & Electrical Engineering 
  University of Birmingham, Edgbaston, Birmingham, B15 2TT 
  e-mail: s.m.charlwood@bham.ac.uk 
  tel: +44 (0)121-414-4340 (shared)/fax: +44 (0)121-414-4291 

  Signal Processing & Imagery Department 
  Defence Evaluation & Research Agency (DERA) 
  DERA Malvern, St. Andrews Road, Malvern, Worcs., WR14 3PS 
  e-mail: charlwood@signal.dera.gov.uk
  tel: +44 (0)1684-895452 (shared)/fax: +44 (0)1684-894389 

_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Article: 20968
Subject: Re: Xilinx PCI pinout ?
From: "Holger Kleinert" <Holger@hollgi.de.nuspam>
Date: Wed, 1 Mar 2000 17:18:52 +0100
Links: << >>  << T >>  << A >>
HI !

You can download this information direct from Xilinx Web Site.
I do not know the exact name, but the document is >1MByte...
I got it there ...

Ken Schmidt <kenschmidtNOSPAM@earthlink.net> schrieb in im Newsbeitrag:
BNmt4.2460$pP5.45867@newsread1.prod.itd.earthlink.net...
> Do you mind my asking you a couple of questions? First, what is your cost
> target for your PCI Xilinx? Second,
> are you planning on turning the Xilinx into a hard ASIC?
>
> Thanks,
> --Ken Schmidt
>     Peerless Systems Corp.
>
> "Nicolas Matringe" <nicolas@dotcom.fr> wrote in message
> news:38B56158.3C805F18@dotcom.fr...
> > Hi
> > I plan to buy a Xilinx PCI Core for a SpartanII device but I can't find
> > any information about the core pinout. I'd like to start working on the
> > PCB layout as soon as possible.
> > The planned device is an XC2S50-FG256.
> >
> > Any help, link... is greatly appreciated
> >
> > Thanks in advance,
> > Nicolas MATRINGE           DotCom S.A.
> > Conception electronique    16 rue du Moulin des Bruyeres
> > Tel 00 33 1 46 67 51 11    92400 COURBEVOIE
> > Fax 00 33 1 46 67 51 01    FRANCE
>
>


Article: 20969
Subject: Re: AMS board simple questions
From: "Andy Peters" <apeters.Nospam@nospam.noao.edu.nospam>
Date: Wed, 1 Mar 2000 09:38:28 -0700
Links: << >>  << T >>  << A >>
Simon wrote in message
<951925119.333.0.nnrp-11.9e9832fa@news.demon.co.uk>...
>javidiaz@my-deja.com wrote in message <89hu2l$om6$1@nnrp1.deja.com>...
>>
>>I have been trying to do a synthesis for one of the example project
>>provided. BUT  I got the following message:
>>ELBREAD: Warning:
>>Component /GEN_BOARDS/WF1BOARD/U_PE0/U_IF/U_PADIntAck :
>>IPAD not bound.
>>
>>I have noticed that it is the only Component (IPAD) without any
>>Architectural description on ams4ksim.vhd package file.What you should
>>I do? (do a dataflow 'bypass', i.e., just assign the output equal to
>>the input)
>
>Try defining IPAD for synthesis by adding this to your code:
>
>-- synthesis translate_off
>library unisim;
>use unisim.vcomponents.all;
>-- synthesis translate_on

Simon,

that would make the library visible for simulation.  it would be ignored for
synthesis.

a


--
-----------------------------------------
Andy Peters
Sr Electrical Engineer
National Optical Astronomy Observatories
950 N Cherry Ave
Tucson, AZ 85719
apeters (at) noao \dot\ edu

"Money is property; it is not speech."
            -- Justice John Paul Stevens



Article: 20970
Subject: Re: Error in Xilinx application note XAPP131?
From: bobperl@best_no_spam_thanks.com (Bob Perlman)
Date: Wed, 01 Mar 2000 17:24:17 GMT
Links: << >>  << T >>  << A >>
Hi, Steve - 

I've looked at XAPP131, but not a closely as you have, so I can't
comment on the problem you're seeing.  However, I have taken a
(fairly) close look at a design done at QuickLogic by John Birkner and
his associates, and it looks solid to me.

Take a look at:
http://www.quicklogic.com/support/anqn/qan25.pdf

There's also a set of design files at:
http://www.quicklogic.com/support/anqn/async.exe

As I recall, I had to download their design software so I could read
some of the schematics.

As always, my comments come with the famous 30-second Usenet
warranty...

Take care,
Bob Perlman

-----------------------------------------------------
Bob Perlman
Cambrian Design Works
Digital Design, Signal Integrity
http://www.best.com/~bobperl/cdw.htm
Send e-mail replies to best<dot>com, username bobperl
-----------------------------------------------------
Article: 20971
Subject: Re: Bit Serial Arithmetic De-mystified
From: "Andy Peters" <apeters.Nospam@nospam.noao.edu.nospam>
Date: Wed, 1 Mar 2000 10:26:58 -0700
Links: << >>  << T >>  << A >>
Peter Alfke wrote in message <38B361E8.7378AF8A@earthlink.net>...
>It, the heck, is when you perform arithmetic serially, one bit at a
>time, LSB first, instead of on parallel words.


Peter,

that made my day!


-- a
-----------------------------------------
Andy Peters
Sr Electrical Engineer
National Optical Astronomy Observatories
950 N Cherry Ave
Tucson, AZ 85719
apeters (at) noao \dot\ edu

"Money is property; it is not speech."
            -- Justice John Paul Stevens



Article: 20972
Subject: Re: AMS board simple questions
From: "Simon" <simonb@tile.demon.co.cuthis.uk>
Date: Wed, 1 Mar 2000 17:40:53 -0000
Links: << >>  << T >>  << A >>
Yep.  Engage brain...  I tripped over the combination of
the ams4ksim (simulation?) support stuff and what appeared
to be a message from the Aldec simulator.  As you point out,
this won't help synthesis!


Andy Peters wrote in message <89jh86$2jm$1@noao.edu>...
>Simon wrote in message
><951925119.333.0.nnrp-11.9e9832fa@news.demon.co.uk>...
>>javidiaz@my-deja.com wrote in message <89hu2l$om6$1@nnrp1.deja.com>...
>>>
>>>I have been trying to do a synthesis for one of the example project
>>>provided. BUT  I got the following message:
>>>ELBREAD: Warning:
>>>Component /GEN_BOARDS/WF1BOARD/U_PE0/U_IF/U_PADIntAck :
>>>IPAD not bound.
>>>
>>>I have noticed that it is the only Component (IPAD) without any
>>>Architectural description on ams4ksim.vhd package file.What you should
>>>I do? (do a dataflow 'bypass', i.e., just assign the output equal to
>>>the input)
>>
>>Try defining IPAD for synthesis by adding this to your code:
>>
>>-- synthesis translate_off
>>library unisim;
>>use unisim.vcomponents.all;
>>-- synthesis translate_on
>
>Simon,
>
>that would make the library visible for simulation.  it would be ignored
for
>synthesis.
>



Article: 20973
Subject: Re: Recommended VHDL titles wanted ...
From: "Number Cruncher" <michel.heuts@pandora.be>
Date: Wed, 01 Mar 2000 19:41:39 GMT
Links: << >>  << T >>  << A >>
Hi guys,

                thanks a lot!

regards,

Michel


Article: 20974
Subject: Re: Xilinx Tools Vs Altera tools
From: Jerry English <jenglish@planetc.com>
Date: Wed, 01 Mar 2000 14:45:29 -0500
Links: << >>  << T >>  << A >>
Glad this thread started. I too am looking at Altera/Xilinx for a large design,
about half million ASIC
gates. What that equates to in fpga gates is up for debate. What I want to know
is what kind of
time am I looking at as far as place and routes go,. When I make small changes
in the design do I
have to start from "bare die" or can the existing layout be modified? Since it
appears that Synopsys's
FPGA express is the front end tool for HDL what kind of times should one expect
for compiling?
If you do respond please indicate the platform.

Thanks
Jerry

wamsi@my-deja.com wrote:

> How does Quartus for APEX compare with the Xilinx tools version for Virtex.
> Principally I am interested in post Synopsys features.
> 1. Timing analysis reports
> 2. Fitter/Router
> 3. any experience with compile time for 1mil gate designs
> 4. timing driven compilation
> 5. Incremental compile
> 6. Testability tools
>
> Thanks
> -Wamsi
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



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