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 36500

Article: 36500
Subject: Re: ideas
From: Ray Andraka <ray@andraka.com>
Date: Fri, 09 Nov 2001 23:42:43 GMT
Links: << >>  << T >>  << A >>
Basically anything you can dream up a digital circuit for.  The FPGA is
just a sea of uncommitted logic that you can wire up for just about
anything your little heart desires.  We are typically pushing FPGAs for
performance and functionality.  Our recent stuff includes several
weather radar demodulator/processors, soft radio recievers, a spectrum
analyzer, HDTV resizing, reformatting and encryption, and other
applications.

Samuel Bogale wrote:

> Can someone suggest ideas what FPGA can be used for?
>
> -Sam

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 36501
Subject: Re: Implementation of filter with three set of coeffs
From: Ray Andraka <ray@andraka.com>
Date: Fri, 09 Nov 2001 23:44:24 GMT
Links: << >>  << T >>  << A >>
Depends on your circuit.  VHDL is just the language to describe your
circuit.

Banana wrote:

> Ok, the question in another form :
>
> Normally How you send the coefficient to a filter in VHDL ???
>
> Thanks
>
>     Banana

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 36502
Subject: Re: Implementation of filter with three set of coeffs
From: Ray Andraka <ray@andraka.com>
Date: Fri, 09 Nov 2001 23:48:54 GMT
Links: << >>  << T >>  << A >>
I don't know (nor do I really want to) your intended circuit topology.  If this is
just a behavioral thing, then you could use an array with all 3 sets of coefficients,
and index which part of the table you use by adding an offset to the table address.
The table can be passed as an array of integers, either through a generic, through a
signal, or through a package.  The type for the array has to be defined in a package
that is referenced by your entity so that the type definition is done before it needs
to be used in the entity block.

Banana wrote:

> I have a filter and want to use it with three different set of
> coefficients, I'm not sure
> about the vhdl implementation of this, I arrange the following :
>
> 1) I define a package SRRC_coeffs where I define the coeff type and
> the coeffs type, here it is :
>
> library ieee;
> use ieee.std_logic_1164.all;
> use ieee.std_logic_signed.all;
>
> package SRRC_coeffs is
>         subtype coeff is std_logic_vector(11 downto 0);
>         type coeffs is array(1 to 42) of coeff;
> end SRRC_coeffs ;
>
> 2) I define an entity named coeffs_selector in this way :
>
> library ieee;
> use ieee.std_logic_1164.all;
> use ieee.std_logic_signed.all;
> use work.SRRC_coeffs.all;
>
> entity coeffs_selector is
>                 port( coeffs_sel        :       in  std_logic_vector( 1 downto 0);
>                           coeffs_SRRC   :       out coeffs
>                  );
> end coeffs_selector;
>
> architecture coeffs_selector_arch of coeffs_selector is
>
>         Constant coeffs_SRRCx3 : coeffs := coeffs'(
>                                                 coeff'("111111100010") , --  1
>                                                 coeff'("111111111100") , --  2
>                                                 coeff'("000001000001") , --  3
>                                                 -- .....................
>                                                 coeff'("000000000000")); -- 42
>
>
>         Constant coeffs_SRRCx4 : coeffs := coeffs'(
>                                                 coeff'("111111100110") , --  1
>                                                 coeff'("111111110001") , --  2
>                                                 coeff'("000000011010") , --  3
>                                                 -- .....................
>                                                 coeff'("000000000000")); -- 42
>
>
>
>         Constant coeffs_SRRCx6 : coeffs := coeffs'(
>                                                 coeff'("000000000110") , --  1
>                                                 coeff'("111111110011") , --  2
>                                                 coeff'("111111100010") , --  3
>                                                 -- .....................
>                                                 coeff'("000000000110")); -- 42
>
> begin
>
>         with coeffs_sel select
>                 coeffs_SRRC <=  coeffs_SRRCx3 when "00",
>                                 coeffs_SRRCx4 when "01",
>                                 coeffs_SRRCx6 when others ;
> end coeffs_selector_arch;
>
> 3) On the filter where arrive the coefficient bus I've something like
> this :
>
> library IEEE;
> use IEEE.std_logic_1164.all;
> use work.SRRC_coeffs.all;
>
> entity SRRC_x_N is
>   port(
>        SRRCxN_coeffs : in coeffs;
>        clk_div_n : in STD_LOGIC;
>        in_fir_MSB : in STD_LOGIC;
>        reset : in STD_LOGIC;
>        count_n : in STD_LOGIC_VECTOR (2 downto 0);
>        polyphase_out : out STD_LOGIC_VECTOR (11 downto 0)
>   );
> end SRRC_x_N;
>
> architecture SRRC_X_N of SRRC_x_N is
>
> signal COEFF_1 : coeff ;
> signal COEFF_2 : coeff ;
> signal COEFF_3 : coeff ;
> -- ...............
> signal COEFF_42 : coeff ;
>
> component FIR_1
>   port (
>        clk_div_n : in STD_LOGIC;
>        coeff_a : in coeff;
>        coeff_b : in coeff;
>        coeff_c : in coeff;
>        coeff_d : in coeff;
>        coeff_e : in coeff;
>        coeff_f : in coeff;
>        coeff_g : in coeff;
>        in_fir_MSB : in STD_LOGIC;
>        reset : in STD_LOGIC;
>        out_fir : out STD_LOGIC_VECTOR (11 downto 0)
>   );
> end component ;
>
> begin
>         coeff_1   <= SRRCxN_coeffs(1);
>         coeff_2   <= SRRCxN_coeffs(2);
>         coeff_3   <= SRRCxN_coeffs(3);
>         -- ...............
>         coeff_42  <= SRRCxN_coeffs(42);
>
> U1 : FIR_1
>   port map(
>        clk_div_n => clk_div_n,
>        coeff_a => coeff_1,
>        coeff_b => coeff_2,
>        coeff_c => coeff_3,
>        coeff_d => coeff_4,
>        coeff_e => coeff_5,
>        coeff_f => coeff_6,
>        coeff_g => coeff_7,
>        in_fir_MSB => in_fir_MSB,
>        out_fir => to_in_0,
>        reset => reset
>   );
>
> -- ...............
>
> end SRRC_X_N;
>
> AND FINALLY HERE ARE MY QUESTION :
> A) Aldec don't permit me to do directly the assignment
>         U1 : FIR_1
>                 port map(
>                        coeff_a => SRRCxN_coeffs(1);
>                         .......
>                         );
>
>     I can't understand why and if there is another solution to avoid
> this double assignment
>
> B) there's a clever way to do this assignment of three different set
> of coefficients to the filter ??
>
> Thanks for your help

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 36503
Subject: Re: Decoupling capacitors on Virtex II
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sat, 10 Nov 2001 10:58:45 +1100
Links: << >>  << T >>  << A >>


pete dudley wrote:
> 
> Our bread and butter decoupling cap is .01uF (10nF) 0805 surface mount, like
> 1 per 4 VCC on fpga's, and we back them up with a few larger caps up to 10uF
> tantalums.
> 
> We use terminated differential signalling for the high speed stuff if
> possible to cancel the ground bounce and use slew rate control on the rest.
> 
> On the highest end of the switching spectrum the ground/power planes help
> you and my guess is that 100pF or smaller chip caps do nothing for you.
> 
> Once we built some multiprocessor boards that ran about 150MHz but the caps
> were bad so we removed all of them. The boards ran fine without any
> decoupling. I'd like to try that experiment again with high speed fpga's.
> 
> I like Austin's idea of tuning the caps to the operating frequency.

With *thin* layers (such as in 8 layer boards etc), all the caps
can be left off except one electro per board. The VCC and GND planes
must be adjacent to provide very low planar transmission line
impedance (lumped capacitance is meaningless at fast edges).

A bit of track to bypass caps and vias hardly matters relative
to the inductance in many internal bond wires.

Article: 36504
Subject: XIlinx SLOW configuration option
From: "Dave Brown" <dbrown@novatel.ca>
Date: Fri, 9 Nov 2001 17:14:12 -0700
Links: << >>  << T >>  << A >>
Is it possible to use the SLOW constraint on a SpartanXL output? The
Constaints guide says that SLOW is not applicable to SpartanXL's, yet the
SpartanXL datasheet talks about setting the slew rate to slow. I tried it
and on a scope, there was no difference in the edges of my outputs. Anyone
know if I can set a slow slew rate on my outputs?
Thanks,
Dave




Article: 36505
Subject: Re: Log2(x) for vhdl?
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sat, 10 Nov 2001 11:27:01 +1100
Links: << >>  << T >>  << A >>
Thanks RA, that works good. I thought of using a loop, but
thought a synthesis tool wouldn't like things like 'while'...


Ray Andraka wrote:
> 
> THis is one I wrote a few years ago.  I keep it in my common library so
> that it is always available.
> 
> function Log2( input:integer ) return integer is
>  variable temp,log:integer;
>  begin
>   temp:=input;
>   log:=0;
>   while (temp /= 0) loop
>    temp:=temp/2;
>    log:=log+1;
>    end loop;
>    return log;
>   end function log2;
> 
> Russell Shaw wrote:
> 
> > Hi all,
> >
> > There's an exponent operator (**), but no log-base-2 (from what
> > i could see). Such a function (with integer result) would be useful
> > would it not?:
> >
> > constant MAXADDR: natural:=1000;
> > .
> > .
> > .
> > signal addrcntr:unsigned(LOG2(MAXADDR) downto 0);
> >
> > LOG2 should round upwards.
> >
> > Could a function be written to do it?

Article: 36506
Subject: Re: XIlinx SLOW configuration option
From: "Dave Brown" <dbrown@novatel.ca>
Date: Fri, 9 Nov 2001 17:28:37 -0700
Links: << >>  << T >>  << A >>
I think I found the answer, I have to use SLEW = SLOW as the contstaint.
"Dave Brown" <dbrown@novatel.ca> wrote in message
news:9shrga$4rf$1@pallas.novatel.ca...
> Is it possible to use the SLOW constraint on a SpartanXL output? The
> Constaints guide says that SLOW is not applicable to SpartanXL's, yet the
> SpartanXL datasheet talks about setting the slew rate to slow. I tried it
> and on a scope, there was no difference in the edges of my outputs. Anyone
> know if I can set a slow slew rate on my outputs?
> Thanks,
> Dave
>
>
>



Article: 36507
Subject: Re: Log2(x) for vhdl?
From: Ray Andraka <ray@andraka.com>
Date: Sat, 10 Nov 2001 01:13:31 GMT
Links: << >>  << T >>  << A >>
This function is not intended to synthesize directly into hardware, rather it
is used to generate constants for synthesized hardware.  Typically for sizing
a bus to a maximum value or setting the number of layers in a tree
structure.  A hardware Log2 is a first '1' detect and encoder, or cna be done
with a normalizing shifter.

Russell Shaw wrote:

> Thanks RA, that works good. I thought of using a loop, but
> thought a synthesis tool wouldn't like things like 'while'...
>
> Ray Andraka wrote:
> >
> > THis is one I wrote a few years ago.  I keep it in my common library so
> > that it is always available.
> >
> > function Log2( input:integer ) return integer is
> >  variable temp,log:integer;
> >  begin
> >   temp:=input;
> >   log:=0;
> >   while (temp /= 0) loop
> >    temp:=temp/2;
> >    log:=log+1;
> >    end loop;
> >    return log;
> >   end function log2;
> >
> > Russell Shaw wrote:
> >
> > > Hi all,
> > >
> > > There's an exponent operator (**), but no log-base-2 (from what
> > > i could see). Such a function (with integer result) would be useful
> > > would it not?:
> > >
> > > constant MAXADDR: natural:=1000;
> > > .
> > > .
> > > .
> > > signal addrcntr:unsigned(LOG2(MAXADDR) downto 0);
> > >
> > > LOG2 should round upwards.
> > >
> > > Could a function be written to do it?

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 36508
Subject: Re: speed of HW JPEG implementations
From: kayrock66@yahoo.com (Jay)
Date: 9 Nov 2001 17:19:20 -0800
Links: << >>  << T >>  << A >>
We had one in .18u standard cell asic process at 24MHz (24M pixels
input rate).  I'm sure it would have gone faster, but that was all we
needed for our motion JPEG app (30fps).  FPGA version (HDL not
optimized for FPGA) ran at 6MHz.

Regards


"Seb" <someone@microsoft.com> wrote in message news:<E5VG7.47306$88.5467995@zwoll1.home.nl>...
> Hi group(s)
> 
> regarding any JPEG standard (JPEG, JPEG2000, JPEG-LS,...) what bitrates can
> be obtained with current implementation technologies (DSP, FPGA, ASIC)?
> 
> Indications and estimations are also welcome.
> 
> cheers
>     Seb

Article: 36509
Subject: Re: 64-bit PCI core for Lattice CPLD?
From: "clevin1234" <clevin1234@home.com>
Date: Sat, 10 Nov 2001 02:16:02 GMT
Links: << >>  << T >>  << A >>
You might want to check out the QL5064 from QuickLogic www.quicklogic.com .
They have integrated a Master/ Target 64bit 66 Mhz PCI core into an OTP
FPGA.   The device has a built in 4 channel (2 XMIT, 2 RCV) DMA Controller
along 74K Gates of programmable user logic. There is no NRE charge for using
the core and they have an RDK board available for evaluation.  You could
easily migrate the CPLD logic into the FPGA an also have the PCI
functionality in a single device.

Chuck

"Alex Rast" <arast@inficom.com> wrote in message
news:ANEE7.153$FU5.365042@news.uswest.net...
> Lattice has a core on their site for 32-bit PCI, but I'm wondering if
there is
> available from Lattice or third parties a 64-bit core. It would be ideal
if it
> can run at 66MHz to boot. I'm looking to target an ispLSI8600VE. Such a
core
> would be helpful because it would let me save an additional chip on our
> circuit board. Right now we're using a different (much smaller) CPLD on
our
> board for other, non-PCI functions. I'm getting ready to design the second
> revision. I'm leaning towards the Lattice chip because I've been
unsatisfied
> with the s/w tools for the chip I have now, and because I really could use
the
> internal tristate busses on the 8000 series. The 8600VE is way, way
overkill
> in terms of macrocell density as a direct replacement for our current
CPLD,
> but if I could integrate the PCI core onto the chip as well then I think
it's
> justifiable.
>
> BTW, anybody out there have any experience with Lattice's tools? What are
your
> thoughts on them?
>
> Alex Rast
> arast@qwest.net
> arast@inficom.com



Article: 36510
Subject: Re: 18V8Z and Philips SNAP compiler
From: Jim Granville <jim.granville@designtools.co.nz>
Date: Sat, 10 Nov 2001 15:34:51 +1300
Links: << >>  << T >>  << A >>
Tim Stewart wrote:
> 
> I have a file that was written for Philips SNAP compiler that I would
> like to modify.  The file was written for the 18V8Z35N, which
> apparently is now obsolete.  The ICT 18CV8 PEEL device is supposedly a
> direct replacement but seems I now have to learn a new compiler (i.e.,
> ICT's WinPLACE).  The quickest solution would be to obtain a copy of
> SNAP since I do have some 18V8Z35N's left.  Anyone know where I can
> get a copy of SNAP?
> 
> Thanks in advance,
> 
> Tim

 I have Slice on 5 1/4 floppies, but not sure I can read them anymore
:-)

 You could check AtmelCUPL V4.8, I think is still on their WEB, as that
used to support 18V8Z ( For a while Atmel cross licenced with Philips )
 The ICT ones have different fuse maps.

 -jg

Article: 36511
Subject: ZX81 production run, is there any interest?
From: McMeikan <mcmeikan@touch88.com.au>
Date: Sat, 10 Nov 2001 11:10:19 +0800
Links: << >>  << T >>  << A >>
I recently had the link
http://www.howell1964.freeserve.co.uk/ZX81/ZX81_FPGA/ZX81_FPGA.htm
pointed out to me, as I love the ZX81 and happen to have a bag of the
XC3042 used to replace the ULA in this design, I am thinking of getting
some boards made and putting together some machines.

Is there much interest? Obviously the more the cheaper.  I am in
Australia but postage of  the PCB should not be too expensive.

I have not started on this yet, but once I get some more info it should
not take long at all and having some feedback from others interested
will help me to scale the cost of making boards down.

    cya,    Andrew...


Article: 36512
Subject: Re: VHDL testbench question
From: "Clyde R. Shappee" <clydes.remove@world.std.com>
Date: Sat, 10 Nov 2001 03:34:34 GMT
Links: << >>  << T >>  << A >>
Yes.  It is possible and very useful. One can have a test bench that is
completely driven by vectors, easily edited by a text editor.

Many texts outline how to do this.

Clyde


Andrew Gray wrote:

> Hi
>
> Is it possible to read a data file into a test bench and use it as one of
> the data inputs to your design?
> Is it possible to write from one of the outputs of your design into a file?
>
> Thanks
>
> Andrew


Article: 36513
Subject: mixed language synthesis with Synplify
From: "I. Servan Uzun" <uzun@bornova.ege.edu.tr>
Date: Sat, 10 Nov 2001 09:27:45 +0000 (UTC)
Links: << >>  << T >>  << A >>
Hi,

How can I synthesize a project consisting of both vhdl and
verilog files in Synplify?

My top level file is a vhdl file and I want to instantiate
a module written in verilog file in my top-level vhdl file. 

Thanks in advance.

Servan



-- 
Posted from  [193.140.73.215] 
via Mailgate.ORG Server - http://www.Mailgate.ORG

Article: 36514
Subject: Re: XIlinx SLOW configuration option
From: "Falk Brunner" <Falk.Brunner@gmx.de>
Date: Sat, 10 Nov 2001 14:22:58 +0100
Links: << >>  << T >>  << A >>


"Dave Brown" <dbrown@novatel.ca> schrieb im Newsbeitrag
news:9shrga$4rf$1@pallas.novatel.ca...
> Is it possible to use the SLOW constraint on a SpartanXL output? The
> Constaints guide says that SLOW is not applicable to SpartanXL's, yet the

AFAIK SLOW is the default setting for all outputs.

> SpartanXL datasheet talks about setting the slew rate to slow. I tried it
> and on a scope, there was no difference in the edges of my outputs. Anyone

;-))    Did you really do a high-speed measurement? Or just hooked up a
probe to the output?
The difference between SLOW and FAST is not 20ns, its maybe somewhere 2 ns.
To see that, use a fast scope (>500MHz) and a fast probe. These 10:1 probes
wont work here.

--
MfG
Falk





Article: 36515
Subject: Funny voltage levels
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sun, 11 Nov 2001 00:41:46 +1100
Links: << >>  << T >>  << A >>
Hi all,

I compiled a vhdl design with leonardo/maxplus2 into an acex 1k30
device, and found that as well as the normal 3.3V outputs, there
were also some pulses at 1.5V and 0.3V. The pins are unloaded.
The signals come from a bit of code i've been trying to debug.
Can tools let these kinds of conflicting designs happen without
warnings?

Article: 36516
Subject: Re: Funny voltage levels
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sun, 11 Nov 2001 00:51:17 +1100
Links: << >>  << T >>  << A >>
Hi all,

I compiled a vhdl design with leonardo/maxplus2 into an acex 1k30
device, and found that as well as the normal 3.3V outputs, there
were also some pulses at 1.5V and 0.3V. The pins are unloaded.
The signals come from a bit of code i've been trying to debug.
Can tools let these kinds of conflicting designs happen without
warnings?

There's no tristate buses or gates in the design, and only one
clock.

Article: 36517
Subject: Re: ZX81 production run, is there any interest?
From: Andrew Owen <aowen@mail.ru>
Date: Sat, 10 Nov 2001 16:51:05 +0000
Links: << >>  << T >>  << A >>
in article 3BEC9A9B.2D0EA6EB@touch88.com.au, McMeikan at
mcmeikan@touch88.com.au wrote on 11/10/01 3:10 AM:

> I recently had the link
> http://www.howell1964.freeserve.co.uk/ZX81/ZX81_FPGA/ZX81_FPGA.htm
> pointed out to me, as I love the ZX81 and happen to have a bag of the
> XC3042 used to replace the ULA in this design, I am thinking of getting
> some boards made and putting together some machines.
> 
> Is there much interest? Obviously the more the cheaper.  I am in
> Australia but postage of  the PCB should not be too expensive.
> 
> I have not started on this yet, but once I get some more info it should
> not take long at all and having some feedback from others interested
> will help me to scale the cost of making boards down.

I know a few Speccy users in Oz who may be interested. If I could make a
suggestion, since you can build it any way you want, it might be worthwhile
having a switchable ROM with ZX80, ZX81, and TreeForth as options, and the
RAM can easily be extended to 64K. The German group ZX-Team probably has a
lot of helpful info on this kind of thing, and may even be interested in
buying a few boards but I can't remember their web or email address off had
- try google.

-(another) Andrew


Article: 36518
Subject: Re: ZX81 production run, is there any interest?
From: Peter Alfke <palfke@earthlink.net>
Date: Sat, 10 Nov 2001 17:06:32 GMT
Links: << >>  << T >>  << A >>


Andrew Owen wrote:

> iI know a few Speccy users in Oz who may be interested. If I could make a
> suggestion, since you can build it any way you want, it might be worthwhile
> having a switchable ROM with ZX80, ZX81, and TreeForth as options, and the
> RAM can easily be extended to 64K. The German group ZX-Team probably has a
> lot of helpful info on this kind of thing, and may even be interested in
> buying a few boards but I can't remember their web or email address off had
> - try google.

Putting ZX81 into google.com gave 12,000 hits ( in 0.1 sec ). Sorted by # of
references.
More than you ever want to know. Have fun !
Peter Alfke


Article: 36519
Subject: Re: Carry chain in Virtex II
From: Neil Franklin <neil@franklin.ch.remove>
Date: 10 Nov 2001 19:54:00 +0100
Links: << >>  << T >>  << A >>
ndeshmukh@yahoo.com (nitin) writes:

> Anybody have any ideas why Virtex II CLB has two carry chains passing
> through it linking two slices and each and propagating in columns from
> top to bottom...?
>
> Why didn't they link all fourslices in just one chain and joined CLBs
> in one column together and so on in different columns...

Well this is spaeculation (but well founded).

In the original Virtex there ae 2 slices and also 2 carry chains,
each going CLB to CLB taking only one slice per CLB.

Studying the info in XAPP151 is becomes obvious that the actial
transistors are layed out in an "butterfly" configuration, with the 2
slices left and right of the routing!

Sort of like this:


          .----|----|||||----|----.
          | .--|--. ||||| .--|--. |
          | |  C  | ||||| |  C  | |
          | `--|--' ||||| `--|--' |
          | Slice 1 ||||| Slice 0 |
          |    |    |||||    |    | CLB
          |    |    |||||    |    |
          -----|-----||------|-----
          -----|-------|-----|-----
          -----|------|||----|-----
          -----|-----||------|----- routing - and |
          -----|-------||----|-----
          -----|----||-------|-----
          |    |    |||||    |    |
          `----|----|||||----|----'
         carry1^             ^carry0


So having carry going through both slices would require it to meander
through the chip, massively increasing distance and so slowing it down.

I suspect Virtex-2 to have 2 slices each side of routing, so I assume
this to also apply there.


> It would be nice if some one with some insight into this matter can
> shed some light on it...

Hope it helped.


--
Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/
Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer
- Intellectual Property is Intellectual Robbery

Article: 36520
Subject: Re: Can Xilinx recognize the critical path in the design
From: Kuan Zhou <zhouk@rpi.edu>
Date: Sat, 10 Nov 2001 15:13:28 -0500
Links: << >>  << T >>  << A >>
Hi,
   Can you show me how to do it?Or what buttons to push in
the software?

   Thank you very much!

sincerely
-------------
Kuan Zhou
ECSE department


On Thu, 8 Nov 2001, Falk Brunner wrote:

> 
> "Kuan Zhou" <zhouk@rpi.edu> schrieb im Newsbeitrag
> news:Pine.SOL.3.96.1011108021708.6365A-100000@rcs-sun1.rcs.rpi.edu...
> > Hi,
> >    I did a fir design but don't know what's the critical
> > path in my design.Can Xilinx tool has someway to mesure and
> > recognize it?
> 
> Depends on your definition of "critical". The tools can "only" find the path
> with the longest delay in a synchronous design. What they cant do, is to
> look for bad design style (except detecting gated clocks or so).
> 
> --
> MfG
> Falk
> 
> 
> 
> 
> 
> 


Article: 36521
Subject: Re: Can Xilinx recognize the critical path in the design
From: "Jim" <johnsonw10@hotmail.com>
Date: Sat, 10 Nov 2001 23:17:54 GMT
Links: << >>  << T >>  << A >>
Use "Timing Analyzer" after "Implement". The tool will generate a report
showing the path delays.

Jim

"Kuan Zhou" <zhouk@rpi.edu> wrote in message
news:Pine.A41.3.96.1011110151214.22494A-100000@cortez.sss.rpi.edu...
> Hi,
>    Can you show me how to do it?Or what buttons to push in
> the software?
>
>    Thank you very much!
>
> sincerely
> -------------
> Kuan Zhou
> ECSE department
>
>
> On Thu, 8 Nov 2001, Falk Brunner wrote:
>
> >
> > "Kuan Zhou" <zhouk@rpi.edu> schrieb im Newsbeitrag
> > news:Pine.SOL.3.96.1011108021708.6365A-100000@rcs-sun1.rcs.rpi.edu...
> > > Hi,
> > >    I did a fir design but don't know what's the critical
> > > path in my design.Can Xilinx tool has someway to mesure and
> > > recognize it?
> >
> > Depends on your definition of "critical". The tools can "only" find the
path
> > with the longest delay in a synchronous design. What they cant do, is to
> > look for bad design style (except detecting gated clocks or so).
> >
> > --
> > MfG
> > Falk
> >
> >
> >
> >
> >
> >
>



Article: 36522
Subject: Reconfigrable Routers
From: ramnathgoenka@yahoo.com (Ramnath)
Date: 10 Nov 2001 20:11:09 -0800
Links: << >>  << T >>  << A >>
hi,

    I am in need of some ideas about reconfigrable routers( IP ) 
using FPGA's.

    Can some one point the door. Although i have got some papers from
google i could not digest it.

    Ramnath

Article: 36523
Subject: 8031 on board Xilinx XC4000XL
From: "Mohap" <NOSPAMamol_moh@hotmail.com>
Date: Sat, 10 Nov 2001 23:59:37 -0500
Links: << >>  << T >>  << A >>
I have a Xilinx XS40-05XL fpga. i see that it has an onboard 8031 which is
interesting because i needed an 8051 anyway. i need a 256 byte ram. i did a
search on google groups and found a discussion where there was a
disagreement on whether this 8031 has 128bytes RAM or 256.



Article: 36524
Subject: Re: Xilinx ISE false timing errors?
From: Philip Freidin <philip@fliptronics.com>
Date: Sun, 11 Nov 2001 07:56:52 GMT
Links: << >>  << T >>  << A >>
This is a bug that I have noticed too, but have not had the time to report it to
Xilinx yet.

Could you please report it !!!

Additional info: It seems to be triggered by turning on the "try to improve
timing after routing is completed" switch. I believe the "-d n" is such an
example, as is the new "-e n" switch. Leaving these out does not give
as good a result, but the incorrect error message goes away.

Philip

On Mon, 5 Nov 2001 10:19:29 -0700, "Dave Brown" <dbrown@novatel.ca> wrote:
>Hi,
>    When I run place and route using foundation ISE 4.1 (using FPGA Express
>as synthesis tool), I get the report that the submitted design did not meet
>timing requirements, those requirements with an asterix were not met. Then,
>in the list of timing requirements, none have an asterix, and at the
>completeion of place and route, it says all timing requirements met, 0
>errors. Same thing if I look at the static timing report, it says all
>constaints met. What gives? Is there a timing error in there? Seems like
>there isn't, anyone else noticed this?
>Thanks,
>Dave
>
>

Philip Freidin
Fliptronics



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