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 106675

Article: 106675
Subject: Re: Quartus and source control (continued)
From: Petter Gustad <newsmailcomp6@gustad.com>
Date: 17 Aug 2006 09:16:58 +0200
Links: << >>  << T >>  << A >>
Mark McDougall <markm@vl.com.au> writes:

> The *nice* thing about Quartus is that you only need to store *2*
> text files for the project, aside from your source - namely the .QPF
> and .QSF files.

I only store ONE file for Quartus, a tcl script to compile the design.
It includes all the pin assignments, global assignments to generate
SVF files etc. 

A typical build from a fresh checkout will look like:

cvs co designname
cd designname/impl/ep1c6t144c8
quartus_sh -t compile.tcl



Petter

-- 
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

Article: 106676
Subject: Re: Quartus and source control (continued)
From: Mark McDougall <markm@vl.com.au>
Date: Thu, 17 Aug 2006 17:31:55 +1000
Links: << >>  << T >>  << A >>
Martin Schoeberl wrote:

> And for the SOPC builder all information
> is in the .ptf. 

Ah yes, indeed! However, we'll strategically overlook mentioning NIOS
IDE projects and the !@#*@&^#%$% workspace... ;)

Regards,

-- 
Mark McDougall, Engineer
Virtual Logic Pty Ltd, <http://www.vl.com.au>
21-25 King St, Rockdale, 2216
Ph: +612-9599-3255 Fax: +612-9599-3266

Article: 106677
Subject: Re: Quartus and source control (continued)
From: Mark McDougall <markm@vl.com.au>
Date: Thu, 17 Aug 2006 17:35:39 +1000
Links: << >>  << T >>  << A >>
Petter Gustad wrote:

> A typical build from a fresh checkout will look like:
> cvs co designname
> cd designname/impl/ep1c6t144c8
> quartus_sh -t compile.tcl

Now there's someone who doesn't like documenting build instructions! ;)

Regards,

-- 
Mark McDougall, Engineer
Virtual Logic Pty Ltd, <http://www.vl.com.au>
21-25 King St, Rockdale, 2216
Ph: +612-9599-3255 Fax: +612-9599-3266

Article: 106678
Subject: Reinstalled Quartus + Nios II => cygwin1.dll hell :-(
From: Tommy Thorn <foobar@nowhere.void>
Date: Thu, 17 Aug 2006 00:36:06 -0700
Links: << >>  << T >>  << A >>
I used to use Cygwin, but when I installed Quartus II and Nios, my local 
installation stopped working; it couldn't find cygwin1.dll.  Rather than 
fight that, I remove my local installation of it, uninstalled Quartus II 
and Nios II and reinstalled everything.  To no avail the binaries under 
C:\altera\kits\nios2_60\bin\nios2-gnutools\H-i686-pc-cygwin\bin doesn't run.

Incidently, the Nios II Dev Board, reference manual seems out of sync. 
In "Appendix B: Restoring Factory Configuration, step 1" the directions 
to launch the SDK Shell doesn't match the menus -- in fact, there aren't 
any SDK Shell under the Altera menu.

I'd be much obliged for any help here.

Thanks,
Tommy

Article: 106679
Subject: Re: Problems about the synthesis(XST)
From: "mh" <moazzamhussain@gmail.com>
Date: 17 Aug 2006 01:12:52 -0700
Links: << >>  << T >>  << A >>
dear Zhaoyi
PAR tools are statistical tools and never generate the same results on
each run. Especially the case when the device utilization is high , You
will find some nets that will not be meeting timings. You fix them by
adding pipeline and next time after PAR some other nets will not be
meeting timings.

To solve this problem use efficient floor planning tools like planAhead
or make many runs of PAR and see results.

regards
MH


Article: 106680
Subject: Using XMD for memory dumps (speed)
From: "Martijn" <M.G.v.d.Horst@gmail.com>
Date: 17 Aug 2006 02:17:13 -0700
Links: << >>  << T >>  << A >>
Hello,

I am using the Xilinx XUPV2P board and connect to it using the USB
cable. My idea is to upload some data to the board, have the FPGA do
some processing and then downloading the result from the board.

To transfer the data I use the XMD utility and connect GDB to it. In
GDB I use the "restore" and "dump memory" commands to access the DDR
memory on the board and the "load" and "c" commands to load my program
and execute it.
The nice thing is that all of this is easy to set up, can be scripted,
and requires only the usb cable to connect to the board.

The bad thing is that dumping the memory takes a long time.
Downloading the data set to the board using the "restore" command has a
reasonable speed (for my purposes), about 60K/sec.
Uploading the memory from the board back to the PC with the "dump
memory" command, however, has a speed of approx. 2K/sec.

This was measured running the USB cable at 12 Mhz, and after the
program had finished so no other peripherals were using the bus. The
system consisted of a powerPC core, a serial interface and a 256 MB DDR
module, and everything the BSB adds to make them work together.

Can anybody tell me what causes this difference in speed?
Any hints on how to increase the speed would also be very welcome.

Currently I am considering using the serial port to transfer data back
to the PC, since it would get a higher throughput.
The ethernet port would also be an option, but since I have no license
for the Ethernet core from Xilinx it is kinda hard to get it working.

Thanks,

Martijn


Article: 106681
Subject: Re: Simple state machine in CUPAL
From: Jim Granville <no.spam@designtools.maps.co.nz>
Date: Thu, 17 Aug 2006 21:34:36 +1200
Links: << >>  << T >>  << A >>
logjam wrote:
> Thanks for the help!  Here is what I came up with.  It seems to work
> good (in the simulator!!!).  I hope there are no glitches in the 8080
> status signals...

 From here, you are on your own....

  I will add that CUPL can append the Sim vectors to the JED file, so
you can (on a good programmer) run a full HW vector test, after device PGM.
  That can be usefull, when you have a green design, and are not
sure if the PCB is valid.

-jg


Article: 106682
Subject: Re: Open-source JTAG software?
From: Kolja Sulimma <news@sulimma.de>
Date: Thu, 17 Aug 2006 12:12:01 +0200
Links: << >>  << T >>  << A >>
Antti schrieb:
> 1) JDrive does not support XCF ASFAIK

We use it on our time to digital converters to programm XCF02S.

Kolja Sulimma

Article: 106683
Subject: Re: Open-source JTAG software?
From: "Antti" <Antti.Lukats@xilant.com>
Date: 17 Aug 2006 03:14:04 -0700
Links: << >>  << T >>  << A >>
Kolja Sulimma schrieb:

> Antti schrieb:
> > 1) JDrive does not support XCF ASFAIK
>
> We use it on our time to digital converters to programm XCF02S.
> 
> Kolja Sulimma
then I was misinformed

Antti


Article: 106684
Subject: Re: Reinstalled Quartus + Nios II => cygwin1.dll hell :-(
From: antti.tyrvainen@luukku.com
Date: 17 Aug 2006 03:24:06 -0700
Links: << >>  << T >>  << A >>
Tommy Thorn kirjoitti:

> I used to use Cygwin, but when I installed Quartus II and Nios, my local
> installation stopped working; it couldn't find cygwin1.dll.  Rather than
> fight that, I remove my local installation of it, uninstalled Quartus II
> and Nios II and reinstalled everything.  To no avail the binaries under
> C:\altera\kits\nios2_60\bin\nios2-gnutools\H-i686-pc-cygwin\bin doesn't run.
>
> I'd be much obliged for any help here.
>
> Thanks,
> Tommy

Gygwin and Quartus should work nicely together. Are you sure you try to
start 'real cygwin' from Cygwin installation directory not from Quartus
installation.

My cygwin.bat says:

@echo off

C:
chdir C:\cygwin\bin

bash --login -i

I have cygwin1.dll under C:\cygwin\bin.

Antti


Article: 106685
Subject: Re: Open-source JTAG software?
From: "Antti" <Antti.Lukats@xilant.com>
Date: 17 Aug 2006 03:41:18 -0700
Links: << >>  << T >>  << A >>
Kolja Sulimma schrieb:

> Antti schrieb:
> > 1) JDrive does not support XCF ASFAIK
>
> We use it on our time to digital converters to programm XCF02S.
>
> Kolja Sulimma

Kolja

you are just lucky (more than I was)

1) at the time I needed to program the XCFxxP devices even the BSDL
files for the XCFxxP were not available (and as JDrive depends on them
it couldnt possible be used at time).

2) Secondly there is an errata on some XCFxxP parts saying that
programming only succeeds with JTAG clock above 3MHz, and JDrive
bitbang IO doesnt guarantee that clock rate. Ok it is possible to
desing high speed jtag interface and make JDrive to support it.

3) I still wonder how are you able to burn an XCF device from .BIT (or
binary) file using only the open-source code from JDrive? There is no
open-source tools to create ISC files for XCFxxS or XCFxxP or I am
misinformed again? XGEN1532 isnt even available any more (and I think
it wasnt fully opensource anyway)?

so I think it still stands that the XCF programming algorithm is not
public neither there is full source code (from xilinx) that implements
the algorithm. JDrive can playback ISC files but no opensource tool can
generate them, so Xilinx prop. tools are needed anyway

Antti


Article: 106686
Subject: Re: Power Supply Sequencing to V4 MGTs
From: Peter Mendham <petermendham@NOCANNEDMEAT.computing.dundee.ac.uk>
Date: Thu, 17 Aug 2006 12:38:05 +0100
Links: << >>  << T >>  << A >>
Austin Lesea wrote:
> There is no sequence sensitivity that I am aware of.

Fantastic.  Thank you.

> Perhaps there is something introduced as the device configures, and
> starts or tries to calibrate MGTs before all the supplies are present?
> 
> Since the device requires one Vcco (for the config bank), Vccaux, and
> Vccint to be above their power on reset thresholds, the device could
> configure and start before the MGT supplies are present or even while
> they are still changing (depending on how fast they power on in relation
> to the other three supplies I mentioned).
> 
> That might be the source of your problem?

No problem *yet*, I'm still at design stage, just want to aim for the 
golden egg of right first time :) The V4 specs require a soft-start for 
Vcco, Vccaux and Vccint which I'm giving them.  The MGTs require 
separate LDO supplies which will come up pretty instantaneously at power 
on, certainly way before the three main supplies.  I was just wondering 
whether this was a known no-no and if I needed to spec some kind of 
soft-start for these supplies too.  I have a V2 dev board which works 
fine with an LDO which comes up before the main supplies, I just wanted 
to make that the situation was the same with the V4.

-- Peter

> 
> Austin
> 
> Peter Mendham wrote:
>> Does anyone know if it's a problem if the LDO supplies to V4 MGTs come
>> up before supplies to the rest of the chip (i.e. before VCCINT)?
>>
>> Thanks,
>>
>> -- Peter

Article: 106687
Subject: Re: Using XMD for memory dumps (speed)
From: Siva Velusamy <siva.velusamy@xilinx.com>
Date: Thu, 17 Aug 2006 05:21:12 -0700
Links: << >>  << T >>  << A >>
> Can anybody tell me what causes this difference in speed?
> Any hints on how to increase the speed would also be very welcome.
> 
> Currently I am considering using the serial port to transfer data back
> to the PC, since it would get a higher throughput.
> The ethernet port would also be an option, but since I have no license
> for the Ethernet core from Xilinx it is kinda hard to get it working.
> 

I'd suggest you try using a download TCL script from within XMD. 
Basically write a TCL script which reads N bytes from X location, source 
it within XMD and then execute it. This will avoid the socket 
communication and remote protocol overhead between XMD and GDB.

However, I *think* you are mainly limited by the speed of the jtag 
protocol between the powerpc and XMD. In which case using Ethernet or 
RS232 might be the better option.

/Siva

Article: 106688
Subject: Re: Alternative for Mentor''s HDL Designer
From: "radarman" <jshamlet@gmail.com>
Date: 17 Aug 2006 05:41:51 -0700
Links: << >>  << T >>  << A >>

topweaver@hotmail.com wrote:
> Hi,
> Mentor's HDL Designer has many functions. But I really do not see any
> requirement to use a tool to generate a state machine (I have tested
> the tools from mentor and Xilinx ISE). Directly writing the code in a
> single always block is already very clear and easy. By using such tools
> usually cost more time.
> Many other functions in the HDL Designer are very useful. As a large
> company, Mentor's products indeed cover a very wide area, but at least
> in some cases, the free tools from Topweaver family are much more
> powerful.
> 1, Comparing TME with HDL Designer's Tabular IO, TME shows better
> performance in parameter/generic, dynamical adjustment of HDL code
> format, secure code synchronization, complex HDL code template and
> launch speed.
> 2, Like ISE, QUARTUS, ActiveHDL and other tools, HDL Designer's module
> integration function is based on traditional schematic method, which is
> hardly used in large projects. You can see these tools' demo all using
> only a few modules and wires. Topweaver can easily deal with hundreds
> of ports. A quick demo from http://www.topweaver.com/demo.htm can show
> how to fulfill the integration work within a few minutes. If you can
> use any other tool to get the same performance, please let me know.
> Now the new version of Topweaver is in the final test.
> www.topweaver.com

I agree - which is why I only produce state machines in HDL Designer
that need to be documented, or are designed by someone else. The tool
does a nice job producing pretty graphics, and reasonable HDL. The
documentation facility alone makes it worthwhile for those "committee"
jobs. While you could use Visio, or some other graphics package, HDL
Designer allows you to produce exactly what you drew - so the
documentation is immediately linked with the actual design -
eliminating documentation errors.

However, for the vast majority of my state machines, I just write them
out without any tools (other than UltraEdit).


Article: 106689
Subject: FFT on an FPGA
From: "Raymond" <raybakk@yahoo.no>
Date: 17 Aug 2006 06:12:54 -0700
Links: << >>  << T >>  << A >>
Hi there.

I am thinking about to use a Xilinx FPGA to take FFT on some data.

Xilinx provide a free FFT core that I most likely can use.

I need a windowing function (Like hamming) for that/a FFT function.
I have at the moment no idea of how to handle floatingpoints numbers in
an FPGA, further I have no ideas for any workarounds.

Does anyone have some Ideas/Solutions to that?

Raymond


Article: 106690
Subject: Re: Problems about the synthesis(XST)
From: "MM" <mbmsv@yahoo.com>
Date: Thu, 17 Aug 2006 10:05:02 -0400
Links: << >>  << T >>  << A >>
I think his problem has more to do with some paths being unconstrained since
the reported frequency itself meets the constraint. Another possible reason
for weird behaviour is an inadequate power supply.

/Mikhail


"mh" <moazzamhussain@gmail.com> wrote in message
news:1155802372.094594.62070@74g2000cwt.googlegroups.com...
> dear Zhaoyi
> PAR tools are statistical tools and never generate the same results on
> each run. Especially the case when the device utilization is high , You
> will find some nets that will not be meeting timings. You fix them by
> adding pipeline and next time after PAR some other nets will not be
> meeting timings.
>
> To solve this problem use efficient floor planning tools like planAhead
> or make many runs of PAR and see results.
>
> regards
> MH
>



Article: 106691
Subject: Re: FFT on an FPGA
From: "jens" <roden@rochester.rr.com>
Date: 17 Aug 2006 07:23:41 -0700
Links: << >>  << T >>  << A >>
Let's take a step back here... do you really need floating point?


Article: 106692
Subject: Re: FFT on an FPGA
From: "MM" <mbmsv@yahoo.com>
Date: Thu, 17 Aug 2006 10:43:52 -0400
Links: << >>  << T >>  << A >>
"Raymond" <raybakk@yahoo.no> wrote in message
news:1155820374.552982.238560@i3g2000cwc.googlegroups.com...
>
> Does anyone have some Ideas/Solutions to that?

Use fixed point. The core you mentioned is probably fixed point anyway.


/Mikhail



Article: 106693
Subject: Re: Open-source JTAG software?
From: "MM" <mbmsv@yahoo.com>
Date: Thu, 17 Aug 2006 10:59:19 -0400
Links: << >>  << T >>  << A >>
Antti,

The readme_xc18v00_xcf00s.txt file included in the downloadables for XAPP058
highlights at least some of the differences between programming algorithms
for XC18xx and XCF18xx parts. Don't know if it's of any use though...

/Mikhail



Article: 106694
Subject: Re: Power Supply Sequencing to V4 MGTs
From: Austin Lesea <austin@xilinx.com>
Date: Thu, 17 Aug 2006 08:18:39 -0700
Links: << >>  << T >>  << A >>
Peter,

Sounds just fine.

we are still characterizing switching power supplies for the MGTs, and I 
wish I could recommend one, but we haven't finished all our work.  Right 
now, it is best to treat the MGTs like the multi-GHz blocks they really are.

If you have any questions, feel free to email me directly,

Austin

Article: 106695
Subject: DCM and Maximum Frequency implied by XST
From: "Sandro" <sdroamt@netscape.net>
Date: 17 Aug 2006 09:00:07 -0700
Links: << >>  << T >>  << A >>
In a design:
case 1)
using a clock (divided whith a "vhdl process" 25Mhz
from 50Mhz) so that the remaining of the desing use the divided clock
and a minor part use the original clock...
...then the XST report show:

Device utilization summary:
---------------------------
Selected Device : 3s200ft256-4
 Number of Slices:                    1798  out of   1920    93%
 Number of Slice Flip Flops:           630  out of   3840    16%
 Number of 4 input LUTs:              3500  out of   3840    91%
    Number used as logic:             3244
    Number used as RAMs:               256
 Number of IOs:                        143
 Number of bonded IOBs:                143  out of    173    82%
    IOB Flip Flops:                     32
 Number of BRAMs:                        6  out of     12    50%
 Number of GCLKs:                        2  out of      8    25%

Timing Summary:
---------------
Speed Grade: -4
   Minimum period: 26.113ns (Maximum Frequency: 38.295MHz)
   Minimum input arrival time before clock: 13.848ns
   Maximum output required time after clock: 12.917ns
   Maximum combinational path delay: No path found

_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._.

In the same desing but:
case 2)
using a dcm to syntetize a frequency (35Mhz from 50Mhz) so that
the remaining of the design use the dcm output as clock and a minor
part use the original clock...
...then the XST report show:

Device utilization summary:
---------------------------
Selected Device : 3s200ft256-4
 Number of Slices:                    1610  out of   1920    83%
 Number of Slice Flip Flops:           657  out of   3840    17%
 Number of 4 input LUTs:              3137  out of   3840    81%
    Number used as logic:             2881
    Number used as RAMs:               256
 Number of IOs:                        143
 Number of bonded IOBs:                143  out of    173    82%
 Number of BRAMs:                        6  out of     12    50%
 Number of GCLKs:                        1  out of      8    12%
 Number of DCMs:                         1  out of      4    25%

Timing Summary:
---------------
Speed Grade: -4
   Minimum period: 63.448ns (Maximum Frequency: 15.761MHz)
   Minimum input arrival time before clock: 14.704ns
   Maximum output required time after clock: 12.091ns
   Maximum combinational path delay: No path found

============================================================

Anyone knows if it can be a normal behaviour (I'm pointing
in particular the Maximum Frequency 15.761MHz using the DCM vs
38.295MHz not using it)...
...or better ;-) I done some mistake somewhere?

thaks in advance
Sandro


Article: 106696
Subject: Re: FFT on an FPGA
From: "RCIngham" <robert.ingham@smiths-aerospace.com>
Date: Thu, 17 Aug 2006 11:04:44 -0500
Links: << >>  << T >>  << A >>
>Hi there.
>
>I am thinking about to use a Xilinx FPGA to take FFT on some data.
>
>Xilinx provide a free FFT core that I most likely can use.
>
>I need a windowing function (Like hamming) for that/a FFT function.
>I have at the moment no idea of how to handle floatingpoints numbers in
>an FPGA, further I have no ideas for any workarounds.
>
>Does anyone have some Ideas/Solutions to that?
>
>Raymond
>
>
I suspect that the Xilinx core is fixed-point, and handles the data growth
that happens during a FFT in some manner described in its documentation. If
it doesn't, then don't use it.

You could use a number representation such as Q1.15 that handles numbers
in the range -1.0 to +1.0, and do scaling at each pass of the FFT.

True floating-point on a FPGA is difficult.

Cheers!





Article: 106697
Subject: Re: DCM and Maximum Frequency implied by XST
From: "Peter Alfke" <peter@xilinx.com>
Date: 17 Aug 2006 09:28:06 -0700
Links: << >>  << T >>  << A >>
How do your two clock domains, 35 MHz and 50 MHz communicate with each
other?
I hope you are aware of their asynchronous relationship, and the many
problems that poses...
Peter Alfke, Xilinx
============================
Sandro wrote:
> In a design:
> case 1)
> using a clock (divided whith a "vhdl process" 25Mhz
> from 50Mhz) so that the remaining of the desing use the divided clock
> and a minor part use the original clock...
> ...then the XST report show:
>
> Device utilization summary:
> ---------------------------
> Selected Device : 3s200ft256-4
>  Number of Slices:                    1798  out of   1920    93%
>  Number of Slice Flip Flops:           630  out of   3840    16%
>  Number of 4 input LUTs:              3500  out of   3840    91%
>     Number used as logic:             3244
>     Number used as RAMs:               256
>  Number of IOs:                        143
>  Number of bonded IOBs:                143  out of    173    82%
>     IOB Flip Flops:                     32
>  Number of BRAMs:                        6  out of     12    50%
>  Number of GCLKs:                        2  out of      8    25%
>
> Timing Summary:
> ---------------
> Speed Grade: -4
>    Minimum period: 26.113ns (Maximum Frequency: 38.295MHz)
>    Minimum input arrival time before clock: 13.848ns
>    Maximum output required time after clock: 12.917ns
>    Maximum combinational path delay: No path found
>
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._.
>
> In the same desing but:
> case 2)
> using a dcm to syntetize a frequency (35Mhz from 50Mhz) so that
> the remaining of the design use the dcm output as clock and a minor
> part use the original clock...
> ...then the XST report show:
>
> Device utilization summary:
> ---------------------------
> Selected Device : 3s200ft256-4
>  Number of Slices:                    1610  out of   1920    83%
>  Number of Slice Flip Flops:           657  out of   3840    17%
>  Number of 4 input LUTs:              3137  out of   3840    81%
>     Number used as logic:             2881
>     Number used as RAMs:               256
>  Number of IOs:                        143
>  Number of bonded IOBs:                143  out of    173    82%
>  Number of BRAMs:                        6  out of     12    50%
>  Number of GCLKs:                        1  out of      8    12%
>  Number of DCMs:                         1  out of      4    25%
>
> Timing Summary:
> ---------------
> Speed Grade: -4
>    Minimum period: 63.448ns (Maximum Frequency: 15.761MHz)
>    Minimum input arrival time before clock: 14.704ns
>    Maximum output required time after clock: 12.091ns
>    Maximum combinational path delay: No path found
>
> ============================================================
>
> Anyone knows if it can be a normal behaviour (I'm pointing
> in particular the Maximum Frequency 15.761MHz using the DCM vs
> 38.295MHz not using it)...
> ...or better ;-) I done some mistake somewhere?
> 
> thaks in advance
> Sandro


Article: 106698
Subject: Re: FFT on an FPGA
From: Ray Andraka <ray@andraka.com>
Date: Thu, 17 Aug 2006 12:30:42 -0400
Links: << >>  << T >>  << A >>
Raymond wrote:

> Hi there.
> 
> I am thinking about to use a Xilinx FPGA to take FFT on some data.
> 
> Xilinx provide a free FFT core that I most likely can use.
> 
> I need a windowing function (Like hamming) for that/a FFT function.
> I have at the moment no idea of how to handle floatingpoints numbers in
> an FPGA, further I have no ideas for any workarounds.
> 
> Does anyone have some Ideas/Solutions to that?
> 
> Raymond
> 

The first thing you need to address is whether or not you really need 
floating point.  Floating point requires quite a bit more complexity for 
  dynamically managing the scaling.  Unless your data has a very wide 
dynamic range, you may find that a fixed point implementation is 
suitable. If you really need floating point, there is floating point FFT 
IP out there, but not for free as far as I know.  I've got, for example, 
a floating point FFT core for Virtex 4 that does IEEE single precision 
FFT sizes from 32 to 4096 points at 400MS/sec continuous.  Three engines 
fit in a V4SX55 for a composite throughput of over 1GS/sec.

Article: 106699
Subject: Using an FPGA as USB HOST without PHY
From: "bm" <nospam@nospam.fr>
Date: Thu, 17 Aug 2006 18:36:50 +0200
Links: << >>  << T >>  << A >>
Hi everybody,
Isi it possible to connect directly FPGA ouptuts to USB line like this :
http://osainto.free.fr/USBHOST/usbDirect.pdf
(Pull downs are necessary to be able to act as USB Host.)
And then  to use a dedicated IP for USB host role (some on opencores )

Thanks 





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