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 36775

Article: 36775
Subject: Re: Modelsim
From: "Andrew Gray" <andrewgray@iafrica.com>
Date: Tue, 20 Nov 2001 01:09:55 +0200
Links: << >>  << T >>  << A >>
I have their evaluation version but my design exceeds the memory limitation.

Arthur Sharp <arthur@nospam.com> wrote in message
news:3bf98132$0$21711$afc38c87@news.optusnet.com.au...
>
> "Andrew Gray" <andrewgray@iafrica.com> wrote in message
> news:3bf946c4.0@news1.mweb.co.za...
> > Hi
> >
> > Does anyone know where I can get hold of a full-version licence for
> > Modelsim? I only need it for 2 or 3 days.
>
> You should be able to get an evaluation license for Modelsim for 20-30
days
> from their site.
>
> >
> > Is there any alternative VHDL simulator like modelsim that is freely
> > available?
>
> As good as Modelsim, for free, not really.
>
> >
> > Thanks
> >
> > Andrew
>
> A.S.
>
>



Article: 36776
Subject: Re: Incrementing counter from state-machine
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Tue, 20 Nov 2001 10:54:12 +1100
Links: << >>  << T >>  << A >>


Andy Peters wrote:
> 
> Russell Shaw wrote:
> 
> > I traced from the clock input pin to the clock divider, and looked at
> > the fanout from the last flip-flop. I found there were a few counters
> > implemented as scattered logic, not being recognized as counter
> > templates. I found just one or two extra lines in a process can prevent
> > recognition by the compiler as a counter.
> 
> I noticed that about Leonardo -- it's VERY picky about what it considers
> a counter.
> 
> What's bizarre is that I ran Leonardo on two different machines (one a
> Sparc running Solaris 7, the other running Solaris 2.5) and I got
> different results (the 2.5 machine did the right thing for a counter;
> the S7 machine didn't!) for identical code with identical scripts.  I
> haven't gotten to the bottom of that one yet.
> 
> Moral: pay attention to the report and log files.

Maybe there's an option for time-limiting the optimizations.
Did the faster machine optimize better?

Article: 36777
Subject: Re: Xilinx and Multirate clock ??
From: Peter Alfke <peter.alfke@xilinx.com>
Date: Mon, 19 Nov 2001 16:00:00 -0800
Links: << >>  << T >>  << A >>

--------------C3B9749BDAE72B58F3470A77
Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; x-mac-creator="4D4F5353"
Content-Transfer-Encoding: 7bit



rickman wrote:

> Are the DCM on the Virtex II chips fully functional? I have heard
> through the grapevine that there are problems. I have not heard anything
> specific, just that I should check with Xilinx before planning to use
> the DCM.

Virtex-II production devices ( the ones not marked ES ) do everything the data
sheet promises ( see http://www.xilinx.com/partinfo/ds031.htm ).

Engineering samples ( marked ES ) have the following restrictions:

The frequency synthesis output FX can have values of M and D up to 32,
but the M/D quotient may not exceed 4. (This limits the min input frequency to
about 6 MHz min)

In ES devices, the vatiable phase shift function ( increment or decrement by
clock period divided by 256 ) does not work. This is fixed in 3000ES, and of
course in all production ( non-ES ) devices.

Peter Alfke, Xilinx Applications

--------------C3B9749BDAE72B58F3470A77
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
&nbsp;
<p>rickman wrote:
<blockquote TYPE=CITE>Are the DCM on the Virtex II chips fully functional?
I have heard
<br>through the grapevine that there are problems. I have not heard anything
<br>specific, just that I should check with Xilinx before planning to use
<br>the DCM.</blockquote>
Virtex-II production devices ( the ones not marked ES ) do everything the
data sheet promises ( see <u><A HREF="http://www.xilinx.com/partinfo/ds031.htm">http://www.xilinx.com/partinfo/ds031.htm</A></u>
).
<p>Engineering samples ( marked ES ) have the following restrictions:
<p>The frequency synthesis output FX can have values of M and D up to 32,
<br>but the M/D quotient may not exceed 4. (This limits the min input frequency
to about 6 MHz min)
<p>In ES devices, the vatiable phase shift function ( increment or decrement
by clock period divided by 256 ) does not work. This is fixed in 3000ES,
and of course in all production ( non-ES ) devices.
<p>Peter Alfke, Xilinx Applications</html>

--------------C3B9749BDAE72B58F3470A77--


Article: 36778
Subject: ISA interface
From: "Jason Berringer" <jberringer@trace-logic.com>
Date: Mon, 19 Nov 2001 21:16:05 -0500
Links: << >>  << T >>  << A >>
Hello all,

I'm new to the newsgroup and am curious if anyone has VHDL code for a 16 bit
full ISA interface. Everything seems to PCI from what I can see, and I don't
require that level of complexity. If you could help me out or point me to a
source I would be most appreciative.

Thanks



Article: 36779
Subject: Re: Modelsim
From: Ray Andraka <ray@andraka.com>
Date: Tue, 20 Nov 2001 03:24:07 GMT
Links: << >>  << T >>  << A >>
Talk to the sales droid. If you can convince him that you are a serious
prospect, you can get him to turn off the memory limitation for the 20-30 day
eval.

Andrew Gray wrote:

> I have their evaluation version but my design exceeds the memory limitation.
>
> Arthur Sharp <arthur@nospam.com> wrote in message
> news:3bf98132$0$21711$afc38c87@news.optusnet.com.au...
> >
> > "Andrew Gray" <andrewgray@iafrica.com> wrote in message
> > news:3bf946c4.0@news1.mweb.co.za...
> > > Hi
> > >
> > > Does anyone know where I can get hold of a full-version licence for
> > > Modelsim? I only need it for 2 or 3 days.
> >
> > You should be able to get an evaluation license for Modelsim for 20-30
> days
> > from their site.
> >
> > >
> > > Is there any alternative VHDL simulator like modelsim that is freely
> > > available?
> >
> > As good as Modelsim, for free, not really.
> >
> > >
> > > Thanks
> > >
> > > Andrew
> >
> > A.S.
> >
> >

--
--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: 36780
Subject: Re: Synopsys+Xilinx vs Synplicity
From: shengyu_shen@hotmail.com (ssy)
Date: 19 Nov 2001 19:29:18 -0800
Links: << >>  << T >>  << A >>
Hi

all tools come from FPGA vendor are very slow, include ISE and
quartus, I synthesis my design(nnARM at about 150,000 gate) in ISE and
Quartus, both took me about 4 hours(in Athlon 1G), but with synplicity
pro, synthesis take only 44 minites, and another 2 hour to fit in
Quartus II

But Synplicity's synthesis result sometimes is very bad, it generate a
very long cascade chain in my APEX20K400E, so I must divide the chain
into tree like structure in RTL code.

I think it is better to use Synplicity Pro to syn and then use Quartus
or ISE to place and route, but you must write your RTL code more
carefully

khtsoi@cse.cuhk.edu.hk wrote in message news:<9sqnh1$j6u$1@eng-ser1.erg.cuhk.edu.hk>...
> Hi,
> 
> Someone told me that tools from Synplicity is better than the Synopsys
> Xilinx combination in FPGA place and route process. I am sick with the
> bad routing design in Alliance3.1i. Also, it takes me more than 1 day
> to par a design using only 45% slices of a XCV1000E (on a Sun E4500).
> 
> I really hope someone can give me some advices about the performance
> of the Synplicity tools. I will use it on either or both Sun E4500 with
> SunOS and P4 1.4GHz PC with Win/Linux. Does anyone has experience on
> implementing a design in similar size under these environment? Also,
> most of my current design is developed under Synopsys Design Compiler
> which cannot be synthsised directly on FPGA Express. Can the Synplicity
> tools under the synopsys coding style? Last, is there any performance
> differences between the commercial version and evaluation version?
> 
> ---- Brittle
> 
> PS I am now downloading the evaluation version of Synplify.

Article: 36781
Subject: Re: modelsim: free, evaluation or full !?
From: shengyu_shen@hotmail.com (ssy)
Date: 19 Nov 2001 19:33:43 -0800
Links: << >>  << T >>  << A >>
the free or evalueation version have limit on the size of the design,
my design(nnARM,150,000 gate after syn) can not sim in Modelsim eva
version, it kill modelsim immediately, and the ActiveHDL still live,
but become a dumb stupid


"Seb" <someone@microsoft.com> wrote in message news:<P9bK7.8929$98.1458446@zwoll1.home.nl>...
> Hello all.
> 
> With our ISE tools, we got Modelsim Xilinx edition, but no license.
> We installed the free version, it runs fine, also with the Xilinx
> components.
> 
> My question: what is the difference (what are the nags) between the three
> Modelsim versions on the cd:
> - full version
> - evaluation version
> - free version
> 
> thanx
> cheers,
>     Seb

Article: 36782
Subject: how to imply CAM in APEX20K
From: shengyu_shen@hotmail.com (ssy)
Date: 19 Nov 2001 19:45:34 -0800
Links: << >>  << T >>  << A >>
I have APEX20k400E prototype board, but the doc say only SPEX20KE/KC
can imply CAM, fow can I deal with this

Article: 36783
Subject: Re: Does anybody knows where have a free(open hardware) FPGA PCI Development board whith PCB data.....?
From: shengyu_shen@hotmail.com (ssy)
Date: 19 Nov 2001 19:49:31 -0800
Links: << >>  << T >>  << A >>
Hi 
www.opencores.org have prototype board that hold Virtex, and maybe
they also have spartan version, and I think you can search some
research group of univ on web

and Xess(www.xess.com) sell fpga board at very low price for education
purpose

tt889@163.net (apple88888) wrote in message news:<c99f3d6.0111171400.20ad425e@posting.google.com>...
> It may like Spartan-II PCI develop board(DS-PCI32S-KIT2) or something like this.
> 
> remail me :  tt889@163.net

Article: 36784
Subject: what is the price of XC2V2000?
From: shengyu_shen@hotmail.com (ssy)
Date: 19 Nov 2001 20:04:48 -0800
Links: << >>  << T >>  << A >>
I need such a device to mply my design, but I must decide on its
price, but I can not find its price on xilinx site.and further about
XC2V10000.

Article: 36785
Subject: Re: what is the price of XC2V2000?
From: "Tim" <tim@rockylogic.com.nospam.com>
Date: Tue, 20 Nov 2001 04:13:19 -0000
Links: << >>  << T >>  << A >>
6K is biggest part in production.

Prices on www.findchips.com

"ssy" <shengyu_shen@hotmail.com> wrote in message
news:f4a5f64f.0111192004.7adf8716@posting.google.com...
> I need such a device to mply my design, but I must decide on its
> price, but I can not find its price on xilinx site.and further about
> XC2V10000.



Article: 36786
Subject: Synthesis in Active-VHDL
From: bhars18@rediffmail.com (Bharathi)
Date: 19 Nov 2001 21:43:42 -0800
Links: << >>  << T >>  << A >>
Can anyone help in getting a VHDL Design (done in Active-VHDL) synthesised?
Is any extra software necessary to do that?
Thanks in advance.

Article: 36787
Subject: Re: Modelsim
From: "Leon Heller" <leon_heller@hotmail.com>
Date: Tue, 20 Nov 2001 06:54:30 -0000
Links: << >>  << T >>  << A >>

"Andrew Gray" <andrewgray@iafrica.com> wrote in message
news:3bf9932d.0@news1.mweb.co.za...
> I have their evaluation version but my design exceeds the memory
limitation.

You could try downloading the Xilinx version from the Xilinx web site. I've
a feeling it might be the full version, but only for Xilinx chips. I haven't
noticed  any mention of restrictions.


Leon
--
Leon Heller, G1HSM leon_heller@hotmail.con
http://www.geocities.com/leon_heller
Low-cost Altera Flex design kit: http://www.leonheller.com





Article: 36788
Subject: Re: Modelsim
From: "Leon Heller" <leon_heller@hotmail.com>
Date: Tue, 20 Nov 2001 06:56:54 -0000
Links: << >>  << T >>  << A >>
Sorry, I've just checked the Xilinx version. It is only for small designs.

--
Leon Heller, G1HSM leon_heller@hotmail.con
http://www.geocities.com/leon_heller
Low-cost Altera Flex design kit: http://www.leonheller.com



Article: 36789
(removed)


Article: 36790
(removed)


Article: 36791
Subject: Synplify use question
From: dottavio@ised.it (Antonio)
Date: 19 Nov 2001 23:33:55 -0800
Links: << >>  << T >>  << A >>
Good Morning, 
I've the following counter divider 3 , my problem is that it works
only at a maximum of 150MHz while I need to keep it working to 165 MHz
at least on a Xilinx XCV1000 BG560 -4 . I try to use Synplify Pro that
tells that to speed up I've to "changing pad type from OBUF to
OBUF_F_24 for pad clk_div_3_obuf to improve timing" and the same for
"count_3_obuf" but how I could do this ??
Following is the counter, by the way do you have in mind another way
to speed it up (other than buy another device !!!). Ciao



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

entity counter_divider_3 is
	port (
		clk			: in  STD_LOGIC;
		reset		: in  STD_LOGIC;
		count_3		: out STD_LOGIC_VECTOR (2 downto 0);
		clk_div_3 	: out STD_LOGIC 
		);
end counter_divider_3;

architecture counter_divider_3_arch of counter_divider_3 is
	signal int_count_3      : STD_LOGIC_VECTOR (1 downto 0) ;
	signal reset_clk_a_b    : STD_LOGIC_VECTOR (3 downto 0) ;
	signal count_0_delayed  : STD_LOGIC;
begin		

	process (clk, reset)
	begin 
		if reset='1' then  
			int_count_3 <= "01"; 
		elsif rising_edge(clk) then	   
			-- & funziona come aggregatore di bit non  un and logico !!
			int_count_3 <= int_count_3(0) & not(int_count_3(0) or
int_count_3(1));
		end if;
	end process;

	process(clk)
	begin
   		if falling_edge(clk) then
       		count_0_delayed <= int_count_3(0);
   		end if;
	end process;
	
	clk_div_3 <= int_count_3(0) nor count_0_delayed;
	
	with int_count_3 select
		count_3 <= 	"010" when "00" ,
				  	"001" when "10" ,
			 		"000" when "01" ,
			 		"XXX" when others;
end counter_divider_3_arch;

Article: 36792
Subject: Re: Virtex2 gate-level simulation: SDF and timing errors
From: assaf_sarfati@yahoo.com (Assaf Sarfati)
Date: 19 Nov 2001 23:45:46 -0800
Links: << >>  << T >>  << A >>
Andy Peters <andy@exponentmedia.deletethis.com> wrote in message news:<3BF96DA4.B8172C4C@exponentmedia.deletethis.com>...
> Assaf Sarfati wrote:
> > 
> > Hi everyone,
> > 
> > I am trying to simulate the gate-level VHDL file generated by Xilinx
> > P&R tools. My test design is a bunch of counters connected to an
> > inferred distributed-RAM. The target device is a Virtex-2 chip.
> > 
> > When I simulate the gate-level VHDL by itself, I get timing violation
> > warnings (sometimes) when writing to the distributed-RAM; watching the
> > simulator waveforms, it appears that the clock to the RAM has a 100-pS
> > phase difference to the counters' clock (the clock is routed as a
> > global clock net).
> 
> That sounds like one of Xilinx' bad models.  I've looked through the
> Xilinx functional models, and there's all sorts of things like:
> 
> 	foo_i <= foo after 1 ps;
> 
> and such.  I've ranted before: there should be NO timing information in
> a FUNCTIONAL model.  Check the archives of comp.lang.vhdl for more on
> that subject.  Summary: the Xilinx people oughta learn how to write
> proper models.
>  
> > When I add the gate-level SDF file to the simulation, all the timing
> > violation warnings disappear (for all cases: min, max and typ).
> 
> Right, because the post-route delay information is "real."
>  
> > Trying to trace the generated VHDL code, I see that signals are routed
> > through buffer entities, with built-in delays; apparently the VHDL
> > design itself contains all required delays.
> 
> Again: why does a functional model have timing info?
>  
> --a

What I'd really like is the ability to generate two gate-level models:
one with full timing info (which would be accurate and slowwww) and
one which is good for cycle-based simulations; I would use the first
model for checking timing problems which hadn't been caught in static
timing analysis; I'd use the second for functional verification.

I think that Xilinx (and others) don't like to generate a
functional-only model because it may be synthesizable; it may then
allow me to re-target and synthesize IPs to which I don't have the
source (e.g. Xilinx PCI core).

  Regards
  Assaf Sarfati

Article: 36793
Subject: Synplicity and BlockRAM?
From: VR <thisisntvalid@invaldireturn.co>
Date: Tue, 20 Nov 2001 09:57:02 +0000 (UTC)
Links: << >>  << T >>  << A >>
Hey all.

For a project in VHDL, it was suggested by a colleague of mine for doing
RAM intensive applications, that it is easier to instantiate a small RAM
as opposed to creating an array of std_logic_vector.

I was passed on this code, however when synthesizing for an XCV-800
(Virtex I) using Synplicity Synplify Pro 7.0, it for some odd reason uses
1024 tri-buffs! (BUFFT).

The same code works fine in Spectrum (I checked) and I told works with XSV
synthesis. Neither use BUFFTs or an inordinate amount.

I can't get this simple code to even P&R in Foundation (after being
synthesized by Synplicity), I am told that I cannot have more than 86
BUFFTs on a net and that I actually have 4 sets of 256 BUFFTs.

Any ideas on how to correct? Code is as follows:

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.STD_LOGIC_ARITH.ALL;
use IEEE.STD_LOGIC_UNSIGNED.ALL;

entity ram is 
 port (clk : in std_logic; 
 	we  : in std_logic; 
 	a   : in std_logic_vector(11 downto 0); 
 	di  : in std_logic_vector(3 downto 0); 
 	do  : out std_logic_vector(3 downto 0)); 
 end ram; 
 
 architecture syn of ram is 
 
 type ram_type is array (4095 downto 0) of std_logic_vector (3 downto 0); 
 signal RAM : ram_type; 
 signal read_a : std_logic_vector(11 downto 0); 
 
 begin 
 process (clk) 
 begin 
 	if (clk'event and clk = '1') then  
 		if (we = '1') then 
 			RAM(conv_integer(a)) <= di; 
 		end if; 
 		read_a <= a; 
 	end if; 
 end process; 
 
 do <= RAM(conv_integer(read_a));
 
 end syn;
 
Thanks!
VR.

Article: 36794
Subject: Re: Xilinx Fpga Editor support for Virtex 2...does it exist in 3.x? How about 4.x?
From: hamish@cloud.net.au
Date: 20 Nov 2001 10:27:09 GMT
Links: << >>  << T >>  << A >>
Austin Franklin <austin@da22rkroom.com> wrote:
> I thought I had a full installation of the latest 3.x tools...and brought up
> the FPGA Editor...but it doesn't seem to allow me to select any of the
> Virtex 2 parts...  Does anyone know if this is supposed to be available in
> FPGA Editor with the 3.x tools?  If not, then is it available with the 4.x
> tools?

You need 3.1i, plus the Virtex-II device update, plus service pack 8,
installed in that order. The device update includes service pack 6, but
if you install it after SP8, you need to reinstall SP8.

I don't know if you can download the device update. I have it on CD-ROM.
I'd stick with 4.1i SP2 anyway -- much better for new designs IMHO.

Hamish
-- 
Hamish Moffatt VK3SB <hamish@debian.org> <hamish@cloud.net.au>

Article: 36795
Subject: Re: how to imply CAM in APEX20K
From: "Wolfgang Loewer" <wolfgang.loewer@elca.de>
Date: Tue, 20 Nov 2001 11:30:11 +0100
Links: << >>  << T >>  << A >>
The EP20K400E device is part of the APEX 20KE family and so it supports the
implementation of CAM in it's ESBs (Embedded System Block).

- Wolfgang
http://www.elca.de

"ssy" <shengyu_shen@hotmail.com> schrieb im Newsbeitrag
news:f4a5f64f.0111191945.bb9c30d@posting.google.com...
> I have APEX20k400E prototype board, but the doc say only SPEX20KE/KC
> can imply CAM, fow can I deal with this



Article: 36796
Subject: Re: Log2(x) for vhdl?
From: Jos De Laender <Jos.De_Laender@Alcatel.Be>
Date: Tue, 20 Nov 2001 11:32:40 +0100
Links: << >>  << T >>  << A >>

A recursive approach is preferred in a synthesis environment here.
As suggested, check google for details that I gave already.

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: 36797
Subject: Re: Synthesis in Active-VHDL
From: Edwin Naroska <edwin@ds.e-technik.uni-dortmund.de>
Date: Tue, 20 Nov 2001 12:17:56 +0100
Links: << >>  << T >>  << A >>
Hi,

Bharathi wrote:

> Can anyone help in getting a VHDL Design (done in Active-VHDL) synthesised?
> Is any extra software necessary to do that?
> Thanks in advance.
> 


You need a synthesis tool. You may check out section
4.3 (part 1) of the VHDL FAQ:

	http://www.vhdl.org/comp.lang.vhdl/FAQ1.html#4.3

Xilinx provides a free tool chain (WebPack) including a
synthesis tool. See

	http://www.xilinx.com	

--
Edwin




Article: 36798
Subject: Altera: diff betw. MAX3000 and MAX7000?
From: Utku Ozcan <ozcan@netas.com.tr>
Date: Tue, 20 Nov 2001 15:21:29 +0200
Links: << >>  << T >>  << A >>

Subject explains: What is the most obvious difference
between MAX3000A and MAX7000A (3.3 V) devices? 

Everything seems to be same: speed grade, I/O support,
JTAG.

The only difference I can find so far is that MAX7000A
is a family of wider gate-count spectrum.

Utku

Article: 36799
Subject: Re: Synplicity and BlockRAM?
From: Tom Dillon <tdillon@dilloneng.com>
Date: Tue, 20 Nov 2001 13:28:14 GMT
Links: << >>  << T >>  << A >>
That is the exact code that Exemplar Leonardo will turn into a=20
"write_first" block RAM.

Tom Dillon
Dillon Engineering, Inc.
http://www.dilloneng.com



>>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<

On 11/20/01, 3:57:02 AM, VR <thisisntvalid@invaldireturn.co> wrote=20
regarding Synplicity and BlockRAM?:


> Hey all.

> For a project in VHDL, it was suggested by a colleague of mine for doi=
ng
> RAM intensive applications, that it is easier to instantiate a small R=
AM
> as opposed to creating an array of std_logic_vector.

> I was passed on this code, however when synthesizing for an XCV-800
> (Virtex I) using Synplicity Synplify Pro 7.0, it for some odd reason u=
ses
> 1024 tri-buffs! (BUFFT).

> The same code works fine in Spectrum (I checked) and I told works with=
=20
XSV
> synthesis. Neither use BUFFTs or an inordinate amount.

> I can't get this simple code to even P&R in Foundation (after being
> synthesized by Synplicity), I am told that I cannot have more than 86=

> BUFFTs on a net and that I actually have 4 sets of 256 BUFFTs.

> Any ideas on how to correct? Code is as follows:

> library IEEE;
> use IEEE.STD_LOGIC_1164.ALL;
> use IEEE.STD_LOGIC_ARITH.ALL;
> use IEEE.STD_LOGIC_UNSIGNED.ALL;

> entity ram is
>  port (clk : in std_logic;
>       we  : in std_logic;
>       a   : in std_logic_vector(11 downto 0);
>       di  : in std_logic_vector(3 downto 0);
>       do  : out std_logic_vector(3 downto 0));
>  end ram;

>  architecture syn of ram is

>  type ram_type is array (4095 downto 0) of std_logic_vector (3 downto =
0);
>  signal RAM : ram_type;
>  signal read_a : std_logic_vector(11 downto 0);

>  begin
>  process (clk)
>  begin
>       if (clk'event and clk =3D '1') then
>               if (we =3D '1') then
>                       RAM(conv_integer(a)) <=3D di;
>               end if;
>               read_a <=3D a;
>       end if;
>  end process;

>  do <=3D RAM(conv_integer(read_a));

>  end syn;

> Thanks!
> VR.



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