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
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

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 103825

Article: 103825
Subject: Re: How do I use the DDS core in a verilog flow?
From: langwadt@ieee.org
Date: 12 Jun 2006 13:37:22 -0700
Links: << >>  << T >>  << A >>

xilinx_user skrev:

> Aurelian Lazarut wrote:
> > xilinx_user wrote:
> > > I am new to using Xilinx cores. My intention is to use the DDS core,
> > > for which I was able to generate a number of files using coregen. It
> > > appears that only the "padded" EDIF can be run through the backend
> > > stages. My question - as naive as it is - is what do you instantiate in
> > > a higher level verilog module in order to implement the design?
> > >
> > take a look at core_name.veo (core_name.vho for vhdl) vit a text editor,
> > these files have instantiation template for your core (all the
> > param/generics and ports)
> >
> > (replace core_name with your actual core name)
> > have fun,
> > Aurash
>
> I tried that. The problem is that when you instantiate a verilog module
> it will implicitly refer to a ".V" file, which is the simulation model.
> It appears that only the EDIF file can be used for
> synthesis/place&route.
>
> What am I misinterpreting?

if I use coregen to generate a, for example, a memory block called
"ram2kx8"  I get
from coregen:

ram2kx8.v: that is the wrapper you need to include in you synthesis and
simulation.
ram2kx8.veo: that is a file with an example of how to instatiate the
module (cut-n-paste from this file as needed)
ram2kx8.edn: the edif file need to be in a place where the place and
route toold can find it

-Lasse


Article: 103826
Subject: Virtex4 DCM in DRP mode
From: "Guru" <ales.gorkic@email.si>
Date: 12 Jun 2006 14:15:05 -0700
Links: << >>  << T >>  << A >>
Hi all!

I made a clocking module for Virtex4FX12 that accepts 66MHz from DCM0
(FX output at 6/9 ratio from 100MHz). Output is giving frequencies
ranging from 35 to 66 MHz in 2MHz increments by changing Multiplier
value (while D is fixed at 32) using DCM_ADV's Dynamic Reconfiguration
(DRP). The output drives CMOS imaging chip MT9V403. The structure of
this module is very simple and works perfectly in simulation and on
oscilloscope, but I am not satisfied with the performance in real
usage. When I change the clock settings the imaging chips fails to
respond. The only setting that works perfectly is 66MHz (M=D=32). I
also added BUFGCE at the output to enable clock only when DCM lock and
syncronised to full clock pulse. That doesn't help either. Does anyone
has the idea what can be wrong?   

Thanks in advance

Guru


Article: 103827
Subject: Looking for patent attorney specialized in programmable logic
From: "HT Chang" <htchang@comcast.net>
Date: Mon, 12 Jun 2006 15:30:02 -0700
Links: << >>  << T >>  << A >>
Looking for a qualified patent attorney/agent specialized in programmable 
logic technology to file FPGA architecture and software algorithm related 
patents.
Worked with a patent agent advertized in IEEE magazine but this guy don't 
get it.
I'm in Bay area but the attorney/agent doesn't have to be in Bay area. Any 
recommendations?
Thanks in advance,
HT 



Article: 103828
Subject: Re: PCI Express - Root Complex ?
From: "Aashish Malhotra" <amalhotr@altera.com>
Date: 12 Jun 2006 15:32:04 -0700
Links: << >>  << T >>  << A >>
Hi Jerome, Sylvan,

Looking at your architecture requirements, I would say that your need
is to design an endpoint solution. Here is a link that would perhaps
help you see where an endpoint solution sits in the model PCI Express
system architecture.
http://www/products/ip/images/pci_express_fig1_typ_app.pdf

Going a little bit further, Altera devices (and I think Xilinx is the
same) are capable of doing both root complex as well as end point
applications. Altera offers a complete portfolio of solutions for PCI
Express applications, some highlights as it pertains to Stratix II GX
are given below:
End point IP Core (x1, x4, x8) with proven interoperability at the PCI
SIG (e.g. x8 IP core:
http://www/products/ip/iup/pci-express/m-alt-pcie8.html)
FPGA (Stratix II GX) with embedded transceivers that are shipping today
Stratix II GX based PCI Express development kit orderable today
(http://www/products/devkits/altera/kit-pciexpress_s2gx.html)

Besides Stratix II GX, Altera supports PCI Express on Cyclone II,
Stratix II as well as Stratix GX. Here is a link that would provide you
with an overall idea of Altera PCI Express solutions:
http://www.altera.com/technology/high_speed/protocols/pci_exp/pro-pci_exp.html

Please feel free to contact me if you have any questions
(amalhotr@altera.com)

Regards,
Aashish

Sylvain Munaut <SomeOne@SomeDomain.com> wrote:
> > Looking to use some Xilinx V4FX or Altera Stratix GXparts for designing
> > several endpoints  using PCI Express for the 1st time. I dont have the
> > PCI-Express spec  yet but am wondering whether I need a root complex?
> >
> > My understanding is that root complex differs from endpoint in that it
> > is used
> > for interfacing to main CPU memory.
> > However if I only want to communicate  between various FPGA endpoints
> > (using a switch) do I really need a root complex IP bridge? Seems like
> > all IP cores
> > are endpoints only?
> >
> > Any Xilinx or Altera PCI Express IP core recommendations based on
> > experience?
>
> I would guess it's more or less like PCI and you need someone to
> enumerate the busses and assign address.
> 
> 
> Sylvain


Article: 103829
Subject: Re: Looking for patent attorney specialized in programmable logic
From: Austin Lesea <austin@xilinx.com>
Date: Mon, 12 Jun 2006 15:35:27 -0700
Links: << >>  << T >>  << A >>
HT,

Sure.  Email me at xilinx.

Austin

HT Chang wrote:
> Looking for a qualified patent attorney/agent specialized in programmable 
> logic technology to file FPGA architecture and software algorithm related 
> patents.
> Worked with a patent agent advertized in IEEE magazine but this guy don't 
> get it.
> I'm in Bay area but the attorney/agent doesn't have to be in Bay area. Any 
> recommendations?
> Thanks in advance,
> HT 
> 
> 

Article: 103830
Subject: Re: Virtex4 DCM in DRP mode
From: "Peter Alfke" <peter@xilinx.com>
Date: 12 Jun 2006 15:51:32 -0700
Links: << >>  << T >>  << A >>
I do not know what's wrong, but I would have divided 100 MHz by 25 in a
simple 5-bit counter, and then multiplied the 4 MHz by M/2 in a DCM.
Seems simpler to me.
Note that in synthesis mode, the input frequency can be as low as 1
MHz, as long as the output frequency is above 25 MHz.
Peter Alfke
==================
Guru wrote:
> Hi all!
>
> I made a clocking module for Virtex4FX12 that accepts 66MHz from DCM0
> (FX output at 6/9 ratio from 100MHz). Output is giving frequencies
> ranging from 35 to 66 MHz in 2MHz increments by changing Multiplier
> value (while D is fixed at 32) using DCM_ADV's Dynamic Reconfiguration
> (DRP). The output drives CMOS imaging chip MT9V403. The structure of
> this module is very simple and works perfectly in simulation and on
> oscilloscope, but I am not satisfied with the performance in real
> usage. When I change the clock settings the imaging chips fails to
> respond. The only setting that works perfectly is 66MHz (M=D=32). I
> also added BUFGCE at the output to enable clock only when DCM lock and
> syncronised to full clock pulse. That doesn't help either. Does anyone
> has the idea what can be wrong?   
> 
> Thanks in advance
> 
> Guru


Article: 103831
Subject: Re: Virtex4 DCM in DRP mode
From: "Peter Alfke" <peter@xilinx.com>
Date: 12 Jun 2006 16:21:35 -0700
Links: << >>  << T >>  << A >>
Other choices are:
divide 100 MHz by 5 in a counter, then multiply by N/10, or
divide 100 MHz by10 in a counter, then multiply by N/5
Peter Alfke

Peter Alfke wrote:
> I do not know what's wrong, but I would have divided 100 MHz by 25 in a
> simple 5-bit counter, and then multiplied the 4 MHz by M/2 in a DCM.
> Seems simpler to me.
> Note that in synthesis mode, the input frequency can be as low as 1
> MHz, as long as the output frequency is above 25 MHz.
> Peter Alfke
> ==================
> Guru wrote:
> > Hi all!
> >
> > I made a clocking module for Virtex4FX12 that accepts 66MHz from DCM0
> > (FX output at 6/9 ratio from 100MHz). Output is giving frequencies
> > ranging from 35 to 66 MHz in 2MHz increments by changing Multiplier
> > value (while D is fixed at 32) using DCM_ADV's Dynamic Reconfiguration
> > (DRP). The output drives CMOS imaging chip MT9V403. The structure of
> > this module is very simple and works perfectly in simulation and on
> > oscilloscope, but I am not satisfied with the performance in real
> > usage. When I change the clock settings the imaging chips fails to
> > respond. The only setting that works perfectly is 66MHz (M=D=32). I
> > also added BUFGCE at the output to enable clock only when DCM lock and
> > syncronised to full clock pulse. That doesn't help either. Does anyone
> > has the idea what can be wrong?
> > 
> > Thanks in advance
> > 
> > Guru


Article: 103832
Subject: Re: Anyone with Xilinx SP305-board ?
From: Daniel O'Connor <darius@dons.net.au>
Date: Tue, 13 Jun 2006 08:55:13 +0930
Links: << >>  << T >>  << A >>
John Smith wrote:
> When I discover en error in one of your products, and there is no way what
> so
> ever tha I can get in touch with you, I gett pissed off !  You must have
> some sort of customer service, anything else is just stupid..

http://www.xilinx.com
Click on 'Support'
Click on 'WebCase' (scroll down a bit to see it)
Login and enter your complaint.

If you're going to spit the dummy over the fact you have to use a web
browser instead of an email client, well.. grow up.

Also, as someone else said, if you don't supply any useful information it's
impossible to help you.

-- 
Daniel O'Connor software and network engineer
for Genesis Software - http://www.gsoft.com.au
"The nice thing about standards is that there
are so many of them to choose from."
  -- Andrew Tanenbaum
GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C

Article: 103833
Subject: Re: How to get lowest price for a ModelSim license?
From: "bart" <bart.borosky@latticesemi.com>
Date: 12 Jun 2006 16:58:10 -0700
Links: << >>  << T >>  << A >>
oneweek wrote:
> you can try lattice semi for a fraction of cost, only restriction is
> you have to use their FPGA, but not bad at all.
>
ModelSim is included with shipments of Lattice's ispLEVER base
software, which you can purchase (Windows part number:
LS-HDL-BASE-PC-N) from the Lattice online store for $495 at
http://www.latticesemi.com/store/software.cfm

Regards,
Bart Borosky, Lattice


Article: 103834
Subject: Re: Looking for patent attorney specialized in programmable logic
From: Jim Granville <no.spam@designtools.co.nz>
Date: Tue, 13 Jun 2006 12:26:00 +1200
Links: << >>  << T >>  << A >>
> HT Chang wrote:
> 
>>Looking for a qualified patent attorney/agent specialized in programmable 
>>logic technology to file FPGA architecture and software algorithm related 
>>patents.
>>Worked with a patent agent advertized in IEEE magazine but this guy don't 
>>get it.
>>I'm in Bay area but the attorney/agent doesn't have to be in Bay area. Any 
>>recommendations?
>>Thanks in advance,
>>HT 
Austin Lesea wrote:
> HT,
> 
> Sure.  Email me at xilinx.
> 
> Austin

Wot - no smiley ? :)
-jg


Article: 103835
Subject: Re: Looking for patent attorney specialized in programmable logic
From: "Rob" <robnstef@frontiernet.net>
Date: Tue, 13 Jun 2006 01:24:30 GMT
Links: << >>  << T >>  << A >>
Maybe he moonlights?

"Jim Granville" <no.spam@designtools.co.nz> wrote in message 
news:448e05d3$1@clear.net.nz...
>> HT Chang wrote:
>>
>>>Looking for a qualified patent attorney/agent specialized in programmable 
>>>logic technology to file FPGA architecture and software algorithm related 
>>>patents.
>>>Worked with a patent agent advertized in IEEE magazine but this guy don't 
>>>get it.
>>>I'm in Bay area but the attorney/agent doesn't have to be in Bay area. 
>>>Any recommendations?
>>>Thanks in advance,
>>>HT
> Austin Lesea wrote:
>> HT,
>>
>> Sure.  Email me at xilinx.
>>
>> Austin
>
> Wot - no smiley ? :)
> -jg
> 



Article: 103836
Subject: Re: Looking for patent attorney specialized in programmable logic
From: Austin Lesea <austin@xilinx.com>
Date: Mon, 12 Jun 2006 18:41:30 -0700
Links: << >>  << T >>  << A >>
Rob,

Ha ha.  No.  Hardly.  But I do work with consultants who do good work.

Austin

Rob wrote:

> Maybe he moonlights?

Article: 103837
Subject: Re: xc3sprog -- any updates?
From: Daniel O'Connor <darius@dons.net.au>
Date: Tue, 13 Jun 2006 12:09:45 +0930
Links: << >>  << T >>  << A >>
Eric wrote:
> I was about to start hacking on xc3sprog after e-mailing the author and
> not hearing back, but I was curious if anyone else has done the same.
> I'm interested in adding programmatic support for the spartan-3
> USER1/USER2 instructions, as well as getting it to program my
> coolrunner-II CPLD and Virtex-4 LX. Before I reinvent the wheel, has
> anyone else done this?

I've modified it to run under FreeBSD (pretty trivial changes) but nothing
more (since I'm only using Spartan 3's :)

Perhaps you could put it up on Sourceforge or something similar?

-- 
Daniel O'Connor software and network engineer
for Genesis Software - http://www.gsoft.com.au
"The nice thing about standards is that there
are so many of them to choose from."
  -- Andrew Tanenbaum
GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C

Article: 103838
Subject: Can i use "burstcount" in my userlogic while using Altera SOPC builder 5.1?
From: "xuweijun1983@gmail.com" <xuweijun1983@gmail.com>
Date: 12 Jun 2006 22:28:29 -0700
Links: << >>  << T >>  << A >>
Hi, All! When adding my userlogic in component editor, i can't find the
singal type "burstcount".
Does Altera SOPC builder 5.1 support it?


Article: 103839
Subject: Re: Rumor Control:: Will Quartus phase out supporting AHDL?
From: "Subroto Datta" <sdatta@altera.com>
Date: Tue, 13 Jun 2006 06:22:32 GMT
Links: << >>  << T >>  << A >>
Jim, you are correct. With Quartus you can mix AHDL, VHDL, Verilog and 
schematics (bdf's) in the same project. As was pointed out earlier, if you 
are doing a functional simulation prior to place and route, or a timing 
simulation after place and route you will be able to use a VHDL or Verilog 
simulator, even if the sources are of different types. Quartus writes out a 
VHDL or Verilog netlist for the project in both cases. The thing that you 
cannot do in the mixed language with AHDL case is  behavioral simulation 
prior to place and route.

- Subroto Datta
Altera Corp.

"Jim Granville" <no.spam@designtools.co.nz> wrote in message 
news:448dc77d$1@clear.net.nz...
> Andy wrote:
>> I can think of several reasons Altera's own customers would want to do
>> this, and that should mean that Altera should want to too.
>>
>> For enterprises that have standardized on vhdl/verilog design
>> methodology, yet have legacy AHDL code for which they don't want to
>> throw away and completely start over.
>
>  I've assumed Altera allows mixed-source projects : so you do not have
> to re-code the AHDL much at all.
> Someone in Altera can correct that, if I am wrong ?
>
>  Of course, new projects should always have a language review, to choose
> the best tools to get the job done.
>
> -jg
>
> 



Article: 103840
Subject: Re: xc3sprog -- any updates?
From: Uwe Bonnes <bon@hertz.ikp.physik.tu-darmstadt.de>
Date: Tue, 13 Jun 2006 07:44:43 +0000 (UTC)
Links: << >>  << T >>  << A >>
Daniel O'Connor <darius@dons.net.au> wrote:
> Eric wrote:
> > I was about to start hacking on xc3sprog after e-mailing the author and
> > not hearing back, but I was curious if anyone else has done the same.
> > I'm interested in adding programmatic support for the spartan-3
> > USER1/USER2 instructions, as well as getting it to program my
> > coolrunner-II CPLD and Virtex-4 LX. Before I reinvent the wheel, has
> > anyone else done this?

> I've modified it to run under FreeBSD (pretty trivial changes) but nothing
> more (since I'm only using Spartan 3's :)

> Perhaps you could put it up on Sourceforge or something similar?

Yes, Sourceforge would be a good place.

I have changes to recognise different adapters ( Xilinx parallel, Altera
Byteblaster and the MPSSE machine of an FT2232  USB Adapter).

-- 
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Article: 103841
Subject: Re: How to get lowest price for a ModelSim license?
From: "Francesco" <francesco.poderico@trendcomms.com>
Date: 13 Jun 2006 00:57:59 -0700
Links: << >>  << T >>  << A >>

bart wrote:
> oneweek wrote:
> > you can try lattice semi for a fraction of cost, only restriction is
> > you have to use their FPGA, but not bad at all.
> >
> ModelSim is included with shipments of Lattice's ispLEVER base
> software, which you can purchase (Windows part number:
> LS-HDL-BASE-PC-N) from the Lattice online store for $495 at
> http://www.latticesemi.com/store/software.cfm
>
> Regards,
> Bart Borosky, Lattice

if you use verilog I suggest you to try icarus verilog
:http://www.icarus.com/eda/verilog.
and the wave viewer you can use wave 1.0 : www.iss-us.com
I think they are quite good.
They are free and in my opinion they are veru good.

Francesco


Article: 103842
Subject: Re: How to get lowest price for a ModelSim license?
From: "Nial Stewart" <nial@nialstewartdevelopments.co.uk>
Date: Tue, 13 Jun 2006 09:35:10 +0100
Links: << >>  << T >>  << A >>
> Quartus FIXEDPC $2k license covers windows modelsim
> for vhdl or verilog.
> One year limit only applies to updates -- tools keep working.


Are you sure about that Mike, I thought they changed the licencing
model a while ago so Modelsim only works for 18 months now.


Nial. 



Article: 103843
Subject: FSM state minimization with ISE?
From: backhus <nix@nirgends.xyz>
Date: Tue, 13 Jun 2006 10:54:32 +0200
Links: << >>  << T >>  << A >>
Hi everybody,
I just tried to minimize a simple statemachine with ISE8.1 and wondered 
that there is no effect, while the machine itself is an ideal target for 
state minimization. (It's a lecture example fur just that purpose)

I tried several synthesis properties with no effect.

Anyone knows how to do state minimizing with ISE, or what the reasons 
are for not minimizing the given FSM? It works with Design Compiler!

Best regards
   Eilert

-- example FSM source
-- analyzes a serial bitstream and
-- detects non-BCD values in 4-bit segments (tetrades)



LIBRARY ieee;
USE ieee.std_logic_1164.ALL;

ENTITY pseudotetradedetector IS
   PORT (
     X     : IN  STD_LOGIC;
     Reset : IN  STD_LOGIC;
     CLK   : IN  STD_LOGIC;
     Y     : OUT STD_LOGIC
     );
END pseudotetradedetector;

ARCHITECTURE ptd_2p OF pseudotetradedetector IS

   TYPE state IS (S0, S1, S2, S3, S4, S5, S6, S7,
                  S8, S9, S10, S11, S12, S13, S14);
   SIGNAL cs,ns : state;

BEGIN

  PROCESS (Reset, CLK)
   BEGIN
     IF Reset = '1' THEN
       cs <= S0;
     ELSIF rising_edge(CLK) THEN
       cs<= ns;
     END IF;
   END PROCESS;

   process(cs,x)
   BEGIN
     CASE cs IS
         WHEN S0 => IF (X = '0') THEN
			Y <= '0';
			ns<= S1;
                    ELSE
			Y <= '0';
			ns<= S2;
                    END IF;
         WHEN S1 => IF (X = '0') THEN
			Y <= '0';
			ns<= S3;
                    ELSE
			Y <= '0';
			ns<= S4;
                    END IF;
         WHEN S2 => IF (X = '0') THEN
			Y <= '0';
			ns<= S5;
                    ELSE
			Y <= '0';
			ns<= S6;
                    END IF;
         WHEN S3 => IF (X = '0') THEN
			Y <= '0';
			ns<= S7;
                    ELSE
			Y <= '0';
			ns<= S8;
                    END IF;
         WHEN S4 => IF (X = '0') THEN
			Y <= '0';
			ns<= S9;
                    ELSE
			Y <= '0';
			ns<= S10;
                    END IF;
         WHEN S5 => IF (X = '0') THEN
			Y <= '0';
			ns<= S11;
                    ELSE
			Y <= '0';
			ns<= S12;
                    END IF;
         WHEN S6 => IF (X = '0') THEN
			Y <= '0';
			ns<= S13;
                    ELSE
			Y <= '0';
			ns<= S14;
                    END IF;
         WHEN S7 => IF (X = '0') THEN
			Y <= '0';
                    ELSE
			Y <= '0';
                    END IF;
			ns <= S0;
         WHEN S8 => IF (X = '0') THEN
			Y <= '0';
                    ELSE
			Y <= '1';
                    END IF;
			ns <= S0;
         WHEN S9 => IF (X = '0') THEN
			Y <= '0';
                    ELSE
			Y <= '1';
                    END IF;
			ns <= S0;
         WHEN S10 => IF (X = '0') THEN
			Y <= '0';
                     ELSE
			Y <= '1';
                     END IF;
			ns <= S0;
         WHEN S11 => IF (X = '0') THEN
				Y <= '0';
                     ELSE
			Y <= '0';
                     END IF;
			  ns <= S0;
         WHEN S12 => IF (X = '0') THEN
				Y <= '0';
                     ELSE
			Y <= '1';
                     END IF;
			  ns <= S0;
         WHEN S13 => IF (X = '0') THEN
   		      Y <= '0';
                     ELSE
		      Y <= '1';
                     END IF;
	              ns <= S0;
         WHEN S14 => IF (X = '0') THEN
		      Y <= '0';
                     ELSE
    		      Y <= '1';
                     END IF;
		      ns <= S0;
         WHEN OTHERS => Y <= '0';
		       ns <= S0;
       END CASE;
   END PROCESS;
END ptd_2p;


Article: 103844
Subject: Re: Virtex4 DCM in DRP mode
From: "Guru" <ales.gorkic@email.si>
Date: 13 Jun 2006 03:28:08 -0700
Links: << >>  << T >>  << A >>
Thanks Peter, I will try this route instead solving an unsolvable
problem, which took me about a month without any progress.

Cheers,

Guru


Article: 103845
Subject: Re: Xilinx timing viloations
From: "prav" <praveen.kantharajapura@gmail.com>
Date: 13 Jun 2006 04:00:28 -0700
Links: << >>  << T >>  << A >>
Hi austin,

Generally what is the budget we give for Logic and Routing  delays
respectively?

Regards,
Prav
Aurelian Lazarut wrote:
> Prav,
> The approach is very dependent of the root cause.
> you have to investigate where your time budget is spent, logic or routing
> if is on logic try to reduce the number of logic levels, pipeline,
> re-balance registers, etc.
> if is on routing try to floor plan, use area constraints, different
> switches in map/par etc.
>
> Aurash
>
> prav wrote:
> > Hi all,
> >
> > I am facing timing violations in my FPGA with Max frequency 125 Mhz(Set
> > up viloations).
> > What are the steps to be taken to meet the frequency other than doing a
> > pipeline in RTL.
> > I am using Xilinx FPGA's.
> > 
> > regards
> >


Article: 103846
Subject: Re: Xilinx timing viloations
From: Aurelian Lazarut <aurash@xilinx.com>
Date: Tue, 13 Jun 2006 12:28:43 +0100
Links: << >>  << T >>  << A >>
prav wrote:

>Hi austin,
>
>Generally what is the budget we give for Logic and Routing  delays
>respectively?
>
>  
>
Prav,
the budget is your clock period minus setup and hold (assuming SDR)

Aurash
 

>Regards,
>Prav
>Aurelian Lazarut wrote:
>  
>
>>Prav,
>>The approach is very dependent of the root cause.
>>you have to investigate where your time budget is spent, logic or routing
>>if is on logic try to reduce the number of logic levels, pipeline,
>>re-balance registers, etc.
>>if is on routing try to floor plan, use area constraints, different
>>switches in map/par etc.
>>
>>Aurash
>>
>>prav wrote:
>>    
>>
>>>Hi all,
>>>
>>>I am facing timing violations in my FPGA with Max frequency 125 Mhz(Set
>>>up viloations).
>>>What are the steps to be taken to meet the frequency other than doing a
>>>pipeline in RTL.
>>>I am using Xilinx FPGA's.
>>>
>>>regards
>>>
>>>      
>>>
>
>  
>


-- 
 __
/ /\/\ Aurelian Lazarut
\ \  / System Verification Engineer
/ /  \ Xilinx Ireland
\_\/\/
 
phone:	353 01 4032639
fax:	353 01 4640324
    
     

Article: 103847
Subject: Re: xc3sprog -- any updates?
From: "Eric" <jonas@mwl.mit.edu>
Date: 13 Jun 2006 05:50:22 -0700
Links: << >>  << T >>  << A >>
> Yes, Sourceforge would be a good place.
>

I've created the sf project as xc3sprog -- I think with all these
additions we've expanded the program a bit from what its original
purpose was, but there seems to be no harm in keeping the name.

Evidently SF will be contacting me in a day or two saying they've set
up the project, at which point I'll upload the code and we can start
accepting patches. Does anyone else want admin bits on the site?
   ...Eric


Article: 103848
Subject: Re: Xilinx timing viloations
From: John_H <johnhandwork@mail.com>
Date: Tue, 13 Jun 2006 12:54:26 GMT
Links: << >>  << T >>  << A >>
The general rule for Xilinx devices (others may be similar) is 50% 
routing, 50% logic.  If you're 70% routing, chances are the place & 
route can give better results.  If you have 43% routing, there's a 
strong chance you won't get better results, at least easily.

prav wrote:
> Hi austin,
> 
> Generally what is the budget we give for Logic and Routing  delays
> respectively?
> 
> Regards,
> Prav
> Aurelian Lazarut wrote:

Article: 103849
Subject: Re: Rumor Control:: Will Quartus phase out supporting AHDL?
From: "Andy" <jonesandy@comcast.net>
Date: 13 Jun 2006 05:57:12 -0700
Links: << >>  << T >>  << A >>
Most projects I've been on that used legacy code still had to modify
the code somewhat to suit the new use.  Modifying legacy code in AHDL
is not a good solution for us, since the reviewers are not AHDL
literate (not that its that hard to understand), and it cannot be
simulated on 3rd party simulators, except at the gate level, which is
painfully slow.

I don't see where Altera would lose customers if they had the
capability to convert AHDL to VHDL/verilog. They could put the same
type of copyright restrictions in the code that Xilinx does if they
wish.

Andy


Jim Granville wrote:
> Andy wrote:
> > I can think of several reasons Altera's own customers would want to do
> > this, and that should mean that Altera should want to too.
> >
> > For enterprises that have standardized on vhdl/verilog design
> > methodology, yet have legacy AHDL code for which they don't want to
> > throw away and completely start over.
>
>   I've assumed Altera allows mixed-source projects : so you do not have
> to re-code the AHDL much at all.
> Someone in Altera can correct that, if I am wrong ?
>
>   Of course, new projects should always have a language review, to choose
> the best tools to get the job done.
> 
> -jg




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
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

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