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 33500

Article: 33500
Subject: Re: IOB FF in Synplicity
From: Rick Filipkiewicz <rick@algor.co.uk>
Date: Sat, 28 Jul 2001 20:33:25 +0100
Links: << >>  << T >>  << A >>


Rick Collins wrote:

> Rick Filipkiewicz wrote:
> > I've been working around this for some time by maintaining a list of what I
> > expect to see in the IOBs
> > and using a perl script to compare that with the last section of the MAP
> > report. The list, in fact, holds exceptions to the rule that everything is
> > registered onto & off the chip i.e. the default is to see INFF/OUTFF on an
> > input/output (or both on a bidir - the script parses the top level module to
> > get a list of the IOs).
>
> I am dealing with this problem right now. I can't seem to find the
> earlier posts on this topic that would have covered the details that can
> prevent an inferred FF from being moved to the IOB. I know you can't use
> the async S/R (other than via GSR) and you can't use the output for any
> other logic for an OFDX. A manager who has worked this design says you
> can't even source the FF from internal logic, but rather it has to be
> connected to another FF output. Anyone know about this?
>

That might have been true once, I vaguely remember it circa M1.4/5, but it
certainly isn't 2.1i+.

>
> I looked through a hundred or so results in the Xilinx data base, but
> could not find a doc that summarizes all of the things that are required
> to let a FF be placed in the IOB. Our parts are 4000XLA.
>
> Anyone care to take a stab at this list?
>

Here are the rules for IN/OUT FFs. The tri-state FFs are special and for XC4K
series parts that have them they have to be forced in some obscure way I cannot
remember.

R1: Out FF: No feedback from the output pin, no post output logic. For XC4K you
can do an inversion since the output polarity is selectable but that's all & even
that's dangerous since the synth tool might decide to merge some other logic in
with the inversion.

R2: In FF: No rules except the weak one that there must be no logic between the
pin & FF.

For BiDir IOs where both in & out are to be registered we have:

R3A:  Common clock. Shouldn't be a problem if you stick to the globals.

R3B:  Common initialisation signal. In can do different things to either FF e.g.
one can be an async set & the other a sync preset. A corollary to this is that if
one FF has an init condition so must the other & from the same signal. In other
words this sort of thing is not allowed:

always @(posedge clk)
    if (reset_myinput)
        reg_myinput <= 0;
    else
        reg_myinput <= myinput;

always @(posedge clk)
    reg_myoutput <= myoutput_src

I'm not sure whether this last is a an architectural restriction or a s/w bug
that's now a feature.

R3C: [XC4K only, maybe only some of them] Common clock enable. Note that as in R3A
if one FF has a CE then so must the other.

Now for the crunches aka what Synplify does wrong:

o If you have a bidir registered IO bus with say 64 pins then R3B says that all 64
FF pairs have to have the same init signal. The simplest solution to this is to
just have one common signal. However even if you ensure this in the code Synplify
may do some replication for you & break the rule because it doesn't have the
intelligence to make sure that after fanning out it sticks to R3B is it can.

Similarly for R3C & XC4K parts.

o Synplify considers e.g. a synchronous reset as a "free and gate" in this sense:
If it sees this sort of structure in the logic feeding an FF:

!foo & bar

it will quite cheerfully put foo on the `D' input & bar on the sync reset -
potentially violating R3B. This is where I first came across the IOB packing
problem. Its also possible that this is what your manager has seen as well. You
can hack around this with ``syn_keep'' directives or adding an async set/reset to
the FF concerned.

o If the register you want to pack into an IOB output FF is buried deep inside
some state machine there's no way of telling the Syn' optimiser not to violate R1.
This is the one I'm fighting at the moment. I have a couple of seriously ugly code
specific kludges that work but I don't yet quite understand *why* they work.
However this might just be a consequence of my design style - I don't like
designing SMs as separate combinatorial & registered parts. Even that doesn't help
without adding some extraneous signals to confuse the optimiser.

Of course if you have Syn-Pro with retiming switched on globally that's yet
another source of IOB FF rule breaking.

This is the story as far as I know it & as long as I can (i) stick to the rules at
HDL level, (ii) persuade Syn not to break them behind my back, then IOB packing
works fine [I just set the -pr b flag to MAP].

Generally things go along o.k. for a long time then I do some change & my MAP
report parser barfs & I end up with long & tedious head scratching session.

I might be behind the s/w times on some of these rules & its just possible that
I've overspec'ed the restrictions but I don't think by very much. Basically IOB FF
packing is easy & automagic when it goes right but a serious, Class 1++, PITA when
it doesn't. I tend to leave fixing it till right to the end, if I can, after the
logic works & the timing analysis is fine everywhere but some IOB failed-to-pack
FFs. I don't like doing it like this but the patches that need to be applied are
sometimes build specific.

Your mileage will probably vary by an order of magnitude or so ... Good luck &
happy packing.




Article: 33501
Subject: Re: PCI-Interface
From: Hong <lvhong@aol.com>
Date: Sat, 28 Jul 2001 13:22:53 -0700
Links: << >>  << T >>  << A >>
Hi Zimba,
I'm interested on the 10$ PLX9052, where/how do I get its specification? who's manufacture?

Best regards,

Hong
 

You Wrote:
>>>Just to let you know, the Xilinx PCI core for Spartan II costs about $5000,-
(I have one). A PLX9052 is about $10,-.

Clemens<<<

Article: 33502
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000
From: Rick Filipkiewicz <rick@algor.co.uk>
Date: Sat, 28 Jul 2001 21:25:09 +0100
Links: << >>  << T >>  << A >>


Dave Feustel wrote:

> The modelsim licensing program never asked me for the name
> of the file containing the license. I think the problem is more
> basic than that. And at this point I am pretty fed up with the
> entire licensing fiasco that I have been experiencing.

It would seem that radical measures are called for. (Tongue in cheek here) If you can afford the upfront outlay why
not travel to Xilinx headquarters bearing your machine & the eval CD, get into the support area, and chain yourself to
the release managers desk until they sort it & re-imburse the costs - if necessary also the ones associated with any
HD reformat that might be needed.

After succeeding there you move on to ModelSim. The lack of any sensible information about errors is absolutely
typical of anything that's been touched by Visual C/C++. If its not license.dat the only way to get any further is to
install some cheapo version VC/C++ and trigger the debugger on the error.

One possibility occurs. From time to time I've edited the license.dat file with wordpad to add a new year's license
that's arrived by email. If I select ``Save as plain text'' what happens is that a new file called ``license.dat.txt''
is created but Explorer cannot handle this so only displays ``license.dat''!?
You can only see the error via the NT shell's DIR command. No more problems now I'm using the Xemacs on both the PC &
the Unix box.

<Flame>
Does anyone but me find this idea utterly disgraceful that some s/w can do irreversible things to my HD that
effectively trashes it if I just want to set my clock a bit more accurately? This isn't license protection its the
rampant paranoia behind the ``Digital Millenium Copyright'' act that Adobe used recently on what appears to be a
relatively innocent Russian.

Xilinx I think don't do this,at the price of ``You don't buy the s/w you rent it''.

<Flame turned down - a bit>



Article: 33503
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000 Pro
From: "Dave Feustel" <dfeustel@mindspring.com>
Date: Sat, 28 Jul 2001 18:21:08 -0500
Links: << >>  << T >>  << A >>
Haven't done that as far as I know.

"Ben Franchuk" <bfranchuk@jetnet.ab.ca> wrote in message news:3B627063.4093183@jetnet.ab.ca...
> Dave Feustel wrote:
> >
> > The modelsim licensing program never asked me for the name
> > of the file containing the license. I think the problem is more
> > basic than that. And at this point I am pretty fed up with the
> > entire licensing fiasco that I have been experiencing.
>
> And if you ever set your real time clock ahead by accident, you practicaly
> have to format the drive and reinstall everything ( and get a new licence)
> if you want to run your computer at the correct time. The moment you
> set you clock back the software will refuse to run.
> Ben.



Article: 33504
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000 Pro
From: "Dave Feustel" <dfeustel@mindspring.com>
Date: Sat, 28 Jul 2001 18:24:36 -0500
Links: << >>  << T >>  << A >>
Haven't edited the license.dat file.

Did try rerunning the install program as administrator in the faint hope that perhaps the program
failed because I had downloaded and installed it as a restricted user. No Joy. Did discover that
modelsim.ini was read only. made it writeable and reinstalled. Same useless error message with
no info. This time the error message occurs during instal - I didn't have to run the licensing wizard
to get the bad news.

I think it's time to try Silos 3 again.


"Rick Filipkiewicz" <rick@algor.co.uk> wrote in message news:3B631FA5.3510C44F@algor.co.uk...
>
>
> Dave Feustel wrote:
>
> > The modelsim licensing program never asked me for the name
> > of the file containing the license. I think the problem is more
> > basic than that. And at this point I am pretty fed up with the
> > entire licensing fiasco that I have been experiencing.
>
> It would seem that radical measures are called for. (Tongue in cheek here) If you can afford the upfront outlay why
> not travel to Xilinx headquarters bearing your machine & the eval CD, get into the support area, and chain yourself to
> the release managers desk until they sort it & re-imburse the costs - if necessary also the ones associated with any
> HD reformat that might be needed.
>
> After succeeding there you move on to ModelSim. The lack of any sensible information about errors is absolutely
> typical of anything that's been touched by Visual C/C++. If its not license.dat the only way to get any further is to
> install some cheapo version VC/C++ and trigger the debugger on the error.
>
> One possibility occurs. From time to time I've edited the license.dat file with wordpad to add a new year's license
> that's arrived by email. If I select ``Save as plain text'' what happens is that a new file called ``license.dat.txt''
> is created but Explorer cannot handle this so only displays ``license.dat''!?
> You can only see the error via the NT shell's DIR command. No more problems now I'm using the Xemacs on both the PC &
> the Unix box.
>
> <Flame>
> Does anyone but me find this idea utterly disgraceful that some s/w can do irreversible things to my HD that
> effectively trashes it if I just want to set my clock a bit more accurately? This isn't license protection its the
> rampant paranoia behind the ``Digital Millenium Copyright'' act that Adobe used recently on what appears to be a
> relatively innocent Russian.
>
> Xilinx I think don't do this,at the price of ``You don't buy the s/w you rent it''.
>
> <Flame turned down - a bit>
>
>



Article: 33505
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000 Pro
From: "Dave Feustel" <dfeustel@mindspring.com>
Date: Sat, 28 Jul 2001 18:27:44 -0500
Links: << >>  << T >>  << A >>
Can someone post the results of running 'find /modelsim'
on a computer with a working version of the program
so that I can compare that listing with the output of a find
command run against /modelsim on my machine so I can
look for the name of a file missing on my machine?
(Just kidding! I'm going to Silos 3 now.)





Article: 33506
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000 Pro
From: "Dave Feustel" <dfeustel@mindspring.com>
Date: Sat, 28 Jul 2001 18:47:25 -0500
Links: << >>  << T >>  << A >>

"Rick Filipkiewicz" <rick@algor.co.uk> wrote in message news:3B631FA5.3510C44F@algor.co.uk...
>
>
> Dave Feustel wrote:
>
> > The modelsim licensing program never asked me for the name
> > of the file containing the license. I think the problem is more
> > basic than that. And at this point I am pretty fed up with the
> > entire licensing fiasco that I have been experiencing.
>
> It would seem that radical measures are called for. (Tongue in cheek here) If you can afford the upfront outlay why
> not travel to Xilinx headquarters bearing your machine & the eval CD, get into the support area, and chain yourself to
> the release managers desk until they sort it & re-imburse the costs - if necessary also the ones associated with any
> HD reformat that might be needed.

Actually this problem would pretty much seem to be Modelsim's doing, although I also gave up on installing
Xilinx Foundation 3.3i eval software because of (apparently) unfixable problems with Xilinx licensing.

> After succeeding there you move on to ModelSim. The lack of any sensible information about errors is absolutely
> typical of anything that's been touched by Visual C/C++. If its not license.dat the only way to get any further is to
> install some cheapo version VC/C++ and trigger the debugger on the error.

Using Visual C doesn't prevent the programmer from emiting useful and informative error messages.
Of course, neither does it guarantee that such error messages will be emitted.

> One possibility occurs. From time to time I've edited the license.dat file with wordpad to add a new year's license
> that's arrived by email. If I select ``Save as plain text'' what happens is that a new file called ``license.dat.txt''
> is created but Explorer cannot handle this so only displays ``license.dat''!?
> You can only see the error via the NT shell's DIR command. No more problems now I'm using the Xemacs on both the PC &
> the Unix box.
>
> <Flame>
> Does anyone but me find this idea utterly disgraceful that some s/w can do irreversible things to my HD that
> effectively trashes it if I just want to set my clock a bit more accurately? This isn't license protection its the
> rampant paranoia behind the ``Digital Millenium Copyright'' act that Adobe used recently on what appears to be a
> relatively innocent Russian.
>
> Xilinx I think don't do this,at the price of ``You don't buy the s/w you rent it''.
>
> <Flame turned down - a bit>
>
>

Silos2001demo installed in about 30 seconds and it runs. :-)




Article: 33507
Subject: Jitter Added by FPGA counter
From: "Andrew Bridger" <andrew.bridger@xtra.co.nz>
Date: Sun, 29 Jul 2001 12:59:29 +1200
Links: << >>  << T >>  << A >>
Hi,
In my application the FPGA is generating sampling clocks for various ADC's
and DAC's.  I was planning to use a divide by N counter then divide by 2 to
generate the 50% duty cycle clock.  I am wondering if there is any analysis
I can do to predict how much jitter the counter will add to my ADC clock?
It may be trivial for our application but I'm just not sure.

The ADC max sampling rate is 8MHz, the ADC device itself has 50 ps aperture
jitter.  (I notice a DLL adds 60ps max jitter, not that a DLL is part of the
clock chain in my design)

Chip is XC2S100/200, Foundation ISE.

Thanks
Andrew





Article: 33508
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000
From: Jjdur Ukfjg <nnvh@htkd.net>
Date: Sat, 28 Jul 2001 18:08:05 -0700
Links: << >>  << T >>  << A >>
> So it appears that both the Xilinx Foundation ISE install program
> and now the Modelsim licensing program like to let their customers guess
> what's wrong when the software won't install rather than just explicitly identify
> the cause of the failure.

I had a problem with running Xilinx Student Edition 2.1i under
Win2000Pro.
The application program worked fine, but the flexlm client didn't seem
to
know how to query Win2000 my ethernet-node address (the diagnostic tool
said my address was 'FF.FF.FF.FF.FF' ?!?)

That said, you can download the latest flexlm client package from
http://www.globetrotter.com.  Of course, you have to 'hunt down' the 
corresponding EXE/DLL files in the Modelsim/Xilinx install directories,
then manually replace them (I think.)  If they're embedded, or renamed,
then well, good luck...

The only other thing I can think of, which is probably NOT the cause,
but
you could check anyway...

make sure you have set the LM_LICENSE_FILE environment variable
from a MSDOS (not WinNT/2000) autoexec.bat file,

SET LM_LICENSE_FILE = <path to license.dat>

IN Win2000, I think you set these 'legacy' environment variables from
the
system setup control panel.  Spawning a DOS-shell, typing the line in,
then exiting, won't work because the variable is only active for that
DOS session.

When you download the flexlm-client from www.globetrotter.com, you
should run the LMUTIL program and verify that FLEXLM can find&read your
license file.

> Welcome to the Brave New World of Software Licensing
> 
> Any other alternatives to Modelsim?

Article: 33509
Subject: Re: Can't Install Modelsim - Alternatives for Verilog Simulation???
From: Jjdur Ukfjg <nnvh@htkd.net>
Date: Sat, 28 Jul 2001 18:11:08 -0700
Links: << >>  << T >>  << A >>
Give "Verilogger Pro" a try.  Same GUI front-end (with an HDL
text editor), timing waveform entry and all sorts of goodies!

http://www.syncad.com

Very competitive pricing, too.  Modelsim/PE was, what, $5000?

> Modelsim licensing refuses to work on my computer.
> 
> What alternatives to Modelsim are there for Verilog simulation
> on Windows 2000?

Article: 33510
Subject: Re: free VHDL and/or Verilog tools?
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sun, 29 Jul 2001 11:24:53 +1000
Links: << >>  << T >>  << A >>


"Steven K. Knapp" wrote:
> 
> You might find some useful information on the Programmable Logic Jump Station.
> 
> VHDL and Verilog Tutorials
> http://www.optimagic.com/tutorials.html
> 
> Free Software Page
> http://www.optimagic.com/lowcost.html
> 
> Philipp Krause <pkk@spth.de> wrote in message news:<3B594569.9000504@spth.de>...
> > Are there any free tools around? I'd like to learn VHDL, but don't want
> > to spend money on commercial software since I don't know which chip
> > family I'll use when it comes to implementing something.
> >
> > Philipp Krause

The xilinx webpack, or altera-leonardo spectrum.

--
   ___                                           ___
  /  /\                                         /  /\
 /  /__\ Russell Shaw, B.Eng, M.Eng(Research)  /  /\/\
/__/   / Victoria, Australia, Down-Under      /__/\/\/
\  \  /  http://home.iprimus.com.au/rjshaw    \  \/\/
 \__\/                                         \__\/


Article: 33511
Subject: Re: Laid-off worker needs software
From: sknapp@triscend.com (Steven K. Knapp)
Date: 28 Jul 2001 18:40:46 -0700
Links: << >>  << T >>  << A >>
Here's another link that may be of interest.  The Programmable Logic
Jump Station maintains a set of links to free evaluation software.  It
includes various CPLD/FPGA software, synthesis tools, tutorials, etc.

http://www.optimagic.com/lowcost.html

There are also synthesis tutorials at

http://www.optimagic.com/tutorials.html

All the best in finding a new position.  Keep the faith!  When the
eventual turn-around comes, it will be programmable logic leading the
way!

"Speedy Zero Two" <david@manorsway.freeserve.co.uk> wrote in message news:<9juqde$nua$1@news8.svr.pol.co.uk>...
> FH,
> 
> Try http://www.xilinx.com/sxpresso/webpack.htm
> It's all free but you may have to register.
> Dave
> 
> "feather head" <shyboy1@bellsouth.net> wrote in message
> news:jtm5mtsmvo1p2ums2on744nfreb3punhfh@4ax.com...
> > I was recently laid off from an employer due to a downturn in tech
> > industries.
> > During the time between jobs I would like to improve my base skills in
> > FPGA design and programming.
> > Could someone please post some xilinx design software.
> > thanks, This will help me find a job. I own a copy of OrCad 7.2
> >

Article: 33512
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000 Pro
From: "Dave Feustel" <dfeustel@mindspring.com>
Date: Sat, 28 Jul 2001 21:31:44 -0500
Links: << >>  << T >>  << A >>
If this is the problem why hasn't Modelsim been fixed to work with Windows 2000?

"Jjdur Ukfjg" <nnvh@htkd.net> wrote in message news:3B6361F5.3E4DB801@htkd.net...
> > So it appears that both the Xilinx Foundation ISE install program
> > and now the Modelsim licensing program like to let their customers guess
> > what's wrong when the software won't install rather than just explicitly identify
> > the cause of the failure.
>
> I had a problem with running Xilinx Student Edition 2.1i under
> Win2000Pro.
> The application program worked fine, but the flexlm client didn't seem
> to
> know how to query Win2000 my ethernet-node address (the diagnostic tool
> said my address was 'FF.FF.FF.FF.FF' ?!?)
>
> That said, you can download the latest flexlm client package from
> http://www.globetrotter.com.  Of course, you have to 'hunt down' the
> corresponding EXE/DLL files in the Modelsim/Xilinx install directories,
> then manually replace them (I think.)  If they're embedded, or renamed,
> then well, good luck...
>
> The only other thing I can think of, which is probably NOT the cause,
> but
> you could check anyway...
>
> make sure you have set the LM_LICENSE_FILE environment variable
> from a MSDOS (not WinNT/2000) autoexec.bat file,
>
> SET LM_LICENSE_FILE = <path to license.dat>
>
> IN Win2000, I think you set these 'legacy' environment variables from
> the
> system setup control panel.  Spawning a DOS-shell, typing the line in,
> then exiting, won't work because the variable is only active for that
> DOS session.
>
> When you download the flexlm-client from www.globetrotter.com, you
> should run the LMUTIL program and verify that FLEXLM can find&read your
> license file.
>
> > Welcome to the Brave New World of Software Licensing
> >
> > Any other alternatives to Modelsim?



Article: 33513
(removed)


Article: 33514
(removed)


Article: 33515
Subject: Re: The Continuing Saga of Installing Modelsim software on Windows 2000
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Sun, 29 Jul 2001 13:46:45 +1000
Links: << >>  << T >>  << A >>


Jjdur Ukfjg wrote:
> 
> > So it appears that both the Xilinx Foundation ISE install program
> > and now the Modelsim licensing program like to let their customers guess
> > what's wrong when the software won't install rather than just explicitly identify
> > the cause of the failure.
> 
> I had a problem with running Xilinx Student Edition 2.1i under
> Win2000Pro.
> The application program worked fine, but the flexlm client didn't seem
> to
> know how to query Win2000 my ethernet-node address (the diagnostic tool
> said my address was 'FF.FF.FF.FF.FF' ?!?)

Try installing IPX/SPX protocol in windows to fix that.
 
> That said, you can download the latest flexlm client package from
> http://www.globetrotter.com.  Of course, you have to 'hunt down' the
> corresponding EXE/DLL files in the Modelsim/Xilinx install directories,
> then manually replace them (I think.)  If they're embedded, or renamed,
> then well, good luck...
> 
> The only other thing I can think of, which is probably NOT the cause,
> but
> you could check anyway...
> 
> make sure you have set the LM_LICENSE_FILE environment variable
> from a MSDOS (not WinNT/2000) autoexec.bat file,
> 
> SET LM_LICENSE_FILE = <path to license.dat>
> 
> IN Win2000, I think you set these 'legacy' environment variables from
> the
> system setup control panel.  Spawning a DOS-shell, typing the line in,
> then exiting, won't work because the variable is only active for that
> DOS session.
> 
> When you download the flexlm-client from www.globetrotter.com, you
> should run the LMUTIL program and verify that FLEXLM can find&read your
> license file.
> 
> > Welcome to the Brave New World of Software Licensing
> >
> > Any other alternatives to Modelsim?

--
   ___                                           ___
  /  /\                                         /  /\
 /  /__\ Russell Shaw, B.Eng, M.Eng(Research)  /  /\/\
/__/   / Victoria, Australia, Down-Under      /__/\/\/
\  \  /  http://home.iprimus.com.au/rjshaw    \  \/\/
 \__\/                                         \__\/

Article: 33516
Subject: FPL 2001 - keynotes -- tutorials
From: hartenst@rhrk.uni-kl.de (Reiner Hartenstein)
Date: 28 Jul 2001 21:06:59 -0700
Links: << >>  << T >>  << A >>
don't miss FPL 2001 !

August 27 - 29, 2001 at Belfast, Northern Ireland

tutorials by Peter Alfke, Xilinx / Pat Mead, Altera:
Recent progress in FPGAs / System-On-a-Programmable-Chip (SOPC) 

homepage:
http://www.ee.qub.ac.uk/dsp/fpl2001/

links to keynotes, tutorials, highlights, also see:
http://www.fpl.uni-kl.de/fpl/index.html#fpl2001


Reiner Hartenstein

Article: 33517
Subject: Announcement: Flex 10K10 design kit
From: "Leon Heller" <leon_heller@hotmail.com>
Date: Sun, 29 Jul 2001 12:14:06 +0100
Links: << >>  << T >>  << A >>
I've decided to market the little Flex 10K10 design kit I originally
designed for my own amazement:

 http://www.leonheller.com/lfh.html

I'll also be putting the design files into the public domain.

Whilst it isn't big enough for any of the sexy FPGA CPUs that are
floating around, it is suitable for learning about programmable logic at
very little cost. I've implemented the small 8-bit CPU in Hamblen and
Furman's book on its precursor.

I'll be doing a similar design for one of the larger Spartan-II chips.

Leon




Article: 33518
(removed)


Article: 33519
(removed)


Article: 33520
(removed)


Article: 33521
Subject: Re: 3.3i service pack 8
From: Rick Collins <spamgoeshere4@yahoo.com>
Date: Sun, 29 Jul 2001 12:39:50 -0400
Links: << >>  << T >>  << A >>
Jan Martin wrote:
> 
> Hi,
> 
> I believe they only transfer through an FTP site from there. If I am not mistaken Xilinx used
> to have HTTP download possibilities on a 'documentation' page on their site, which used to be
> accessible from the downloads page of the updates.
> I also tried to download individual files and then tried to recombine using the batch file,
> but have never been able to get it to work. I now download using an FTP program. Advantage is
> that the 'resume' option of the FTP site does work; with some 80Mb to download on a telephone-
> line....
> 
> You may want to try installing a FTP program (CuteFtp for example) and use that to download
> the one large file.
> 
> Cheers,
> Jan Martin Wagenaar
> 
> In article <3B5C2513.7030409@amontec.com>, laurent.gauch@amontec.com says...
> > I am on
> > http://support.xilinx.com/support/techsup/sw_updates/31i/sw_f33i_pc.htm
> > trying to download the Foundation 3.3i Software Updates for PC (3.3i
> > service pack 8), but I cannot download these single files. Why?
> >
> >

Or you might try asking Xilinx to send you the update on CD. I think it
is rediculous that Xilinx thinks everyone has a high speed connection to
the web or that they have the patience to spend several nights trying to
download these huge files. Is there some reason that Xilinx can't split
them up into multiple smaller files? 5 MB chunks would be much easier to
download. With my FTP program I can select them all for download and it
will do them sequentially which maximizes the probability of sucess on
them all. 

-- 

Rick "rickman" Collins

rick.collins@XYarius.com
Ignore the reply address. To email me use the above address with the XY
removed.

Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design      URL http://www.arius.com
4 King Ave                               301-682-7772 Voice
Frederick, MD 21701-3110                 301-682-7666 FAX

Article: 33522
Subject: Re: Athlon 1.4 vs Pentium 4 1.7 for Foundation ISE/ModelSim?
From: Rick Collins <spamgoeshere4@yahoo.com>
Date: Sun, 29 Jul 2001 13:09:13 -0400
Links: << >>  << T >>  << A >>
cyber_spook wrote:
> 
> Hi,
> 
> Rick Filipkiewicz wrote:
> 
> > My conclusion from this small sample is that the P&R tool speed is
> > dominated by memory performance.
> 
> With this type of application (memory hugary!) the speed of the CPU is only
> any good when it has the data to process. In other words a 450, 600 or 1.4
> will just spend most of its time sitting on its "bum" waiting for memory.
> 
> This was proven when Cambridge Uni fitted 4Gb of SRAM (not DRAM) to a 40Mhz
> 386DX - Was the fastest thing you have ever seen!!
> 
> Cyber_Spook_Man

I find this hard to believe. I don't doubt that the general rule for
memory is faster is better. But I don't think a 40 MHz 386 will outrun a
1 GHz anything just because the 386 is running in zero wait state
memory. I also find it hard to believe you can fit 4 Gb of SRAM to a
processor without slowing down the SRAM. That is a lot of memory, about
512 chips (assuming the Gb means gigabits, not gigabytes)!!!

It would be useful to obtain benchmarks on some different platforms. I
can easily access a 1 GHz Intel machine (2 GB) and some slower Intel
machines (1 GB and less) with Foundation 3.3i. Anyone have a design that
they would like to use as a test case?

Of course, the other variable is the design. Even if we come up with a
common design to test, other designs may perform differently across the
different platforms. But this may only be a very slight difference. 
 
-- 

Rick "rickman" Collins

rick.collins@XYarius.com
Ignore the reply address. To email me use the above address with the XY
removed.

Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design      URL http://www.arius.com
4 King Ave                               301-682-7772 Voice
Frederick, MD 21701-3110                 301-682-7666 FAX

Article: 33523
Subject: Re: Xilinx/Altera "behavioral" verilog
From: "B. Joshua Rosen" <bjrosen@polybus.com>
Date: Sun, 29 Jul 2001 13:35:02 -0400
Links: << >>  << T >>  << A >>
The coverage that the FPGA vendors provide is less than perfect but it's
good enough for most applications. In my experience, writing FPGA test
programs for an ASIC emulator manufacturer, about 1 in 50 FPGAs have some
defect although the chance of any one pattern running into that defect is
in the neighborhood of 1 in 100. As a result the chance that any one
pattern will have a problem on any particular FPGA is about 1 in 5000. As
you raise the number of patterns that you run on a particular FPGA the
chances that you will run into a problem approached the 1 in 50 number.

The way that you test FPGAs is different then the way that you test an
ASIC. FPGAs are mostly interconnect which makes them harder to test. On
the other hand because they are soft you can run an unlimited number of
different test patterns on them, which simplifies the problem as compared
to an ASIC. If you are concerned about shipping an
FPGA with a hidden defect then the way that you solve it is to run a
large number of test patterns on your FPGAs and throw
out the bad ones. The FPGA vendors do some of this but they are limited
by the amount of time that they can tie up a chip tester for. Generally
the vendor tests runs in under a second, in my experience it takes about
20 minutes to really test an FPGA. If you are selling low volume high
value equiptment that contains a large number of FPGAs and loads many
different patterns into them, like and ASIC emulator, then this degree
of testing is necessary. If you are shipping high volume low priced
boards then it's cheaper to live with the one in 5000 fall out. Adding
scan logic to a particular circuit is useless because any change to the
pattern, including another place and route on the same design, will
change which resources are used inside of the FPGA.


In article <996308678.26850.0.nnrp-07.9e9832fa@news.demon.co.uk>, "Tim"
<tim@rockylogic.com.nospam.com> wrote:

> "Andy Botterill" <csm@plymouth2.demon.co.uk> wrote in message
> news:YNNvDUAmacY7EwX0@plymouth2.demon.co.uk...
> 
>> >Programmable parts have different test requirements to ASICs, so I
>> >guess the answer to your question is Yes.
>>
>> If you cannot insert scan into an FPGA how do you get a high fault
>> coverage? If you have a poor fault coverage you will be shipping
>> defective parts which is not good for your business.
> 
> The FPGA manufacturers do this for you.  They use a combination of the
> reprogrammability of the part and, presumably, a few undisclosed test
> structures.
> 
> The result is closer to 100% tested than just about any ASIC.  All you
> have to worry about is logical and timing errors :)
> 
> 
> 
> 
>

Article: 33524
(removed)




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