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
2019JanFebMar2019

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 44475

Article: 44475
Subject: Multiply by 8 with DLL in Spaertan-II.
From: "XU QIJUN" <qijun677@oki.com>
Date: Fri, 21 Jun 2002 10:38:18 +0800
Links: << >>  << T >>  << A >>
Hi,

I am using a Spartan-II 100K to do some experiments on clock multiplication.
I successfully made a X2 and X4, then problem occurs when I coded one extra
DLL to make it 8. Following is the bug,
Whats your comments?

ERROR:Place:1726 - Could not find an automatic placement for the following
   components:
    CLKIN of type GCLK IOB is placed at P77.
    dll2x of type DLL is unplaced.
    clk2xg of type GCLK BUFFER is unplaced.
    dll4x of type DLL is unplaced.
    clk4xg of type GCLK BUFFER is unplaced.
    dll8x of type DLL is unplaced.
    clk8xg of type GCLK BUFFER is unplaced.
    lckpad of type GCLK BUFFER is unplaced.
ERROR:Place:1727 - Xilinx requires using locate constraints to preplace such
   connected GCLK/GCLKIO/DLL components.
Total REAL time to Placer completion: 2 secs
Total CPU time to Placer completion: 1 secs

-----------------------------------------------------------------

Following are the codes:

module dll_standard (CLKIN, CLK2X, CLK4X, CLK8X, LOCKED);
input CLKIN;
output CLK2X, CLK4X, CLK8X, LOCKED;

wire RESET;
wire CLKIN_w, RESET_w, CLK2X_dll, CLK4X_dll, LOCKED2X, LOCKED4X;
wire LOCKED2X_delay, RESET4X;
wire LOCKED4X_delay, RESET8X;
wire logic1;
assign RESET = 1'b0;

assign logic1 = 1'b1;

assign CLKIN_w = CLKIN;
assign RESET = RESET_w;

CLKDLL dll2x (.CLKIN(CLKIN_w), .CLKFB(CLK2X), .RST(RESET_w),
              .CLK0(), .CLK90(), .CLK180(), .CLK270(),
              .CLK2X(CLK2X_dll), .CLKDV(), .LOCKED(LOCKED2X));
BUFG   clk2xg (.I(CLK2X_dll),  .O(CLK2X));
SRL16  rstsrl (.D(LOCKED2X), .CLK(CLK2X), .Q(LOCKED2X_delay),
               .A3(logic1), .A2(logic1), .A1(logic1), .A0(logic1));
assign RESET4X = !LOCKED2X_delay;

CLKDLL dll4x (.CLKIN(CLK2X), .CLKFB(CLK4X), .RST(RESET4X),
              .CLK0(), .CLK90(), .CLK180(), .CLK270(),
              .CLK2X(CLK4X_dll), .CLKDV(), .LOCKED(LOCKED4X));
BUFG   clk4xg (.I(CLK4X_dll),  .O(CLK4X));
SRL16  rstsr2 (.D(LOCKED4X), .CLK(CLK4X), .Q(LOCKED4X_delay),
               .A3(logic1), .A2(logic1), .A1(logic1), .A0(logic1));
assign RESET8X = !LOCKED4X_delay;
// assign CLK4X_B = ~ CLK4X;

CLKDLL dll8x (.CLKIN(CLK4X), .CLKFB(CLK8X), .RST(RESET8X),
              .CLK0(), .CLK90(), .CLK180(), .CLK270(),
              .CLK2X(CLK8X_dll), .CLKDV(), .LOCKED(LOCKED_dll));
BUFG   clk8xg (.I(CLK8X_dll),  .O(CLK8X));

BUFG   lckpad (.I(LOCKED_dll), .O(LOCKED));

endmodule




Article: 44476
Subject: design cycle metrics
From: pdemos@aol.com (PDemos)
Date: 21 Jun 2002 02:40:06 GMT
Links: << >>  << T >>  << A >>
anyone have ballpark info on how long it takes to design, develop, debug  a 2
million gate design.  complexity and throughput both high.

Article: 44477
Subject: Re: Multiply by 8 with DLL in Spaertan-II.
From: "XU QIJUN" <qijun677@oki.com>
Date: Fri, 21 Jun 2002 11:08:45 +0800
Links: << >>  << T >>  << A >>
Forgot to mention, my CLKIN = 16.0 MHz. CLK8X will yield 128 MHz.




"XU QIJUN" <qijun677@oki.com> wrote in message
news:3d12919d$1@news.starhub.net.sg...
> Hi,
>
> I am using a Spartan-II 100K to do some experiments on clock
multiplication.
> I successfully made a X2 and X4, then problem occurs when I coded one
extra
> DLL to make it 8. Following is the bug,
> Whats your comments?
>
> ERROR:Place:1726 - Could not find an automatic placement for the following
>    components:
>     CLKIN of type GCLK IOB is placed at P77.
>     dll2x of type DLL is unplaced.
>     clk2xg of type GCLK BUFFER is unplaced.
>     dll4x of type DLL is unplaced.
>     clk4xg of type GCLK BUFFER is unplaced.
>     dll8x of type DLL is unplaced.
>     clk8xg of type GCLK BUFFER is unplaced.
>     lckpad of type GCLK BUFFER is unplaced.
> ERROR:Place:1727 - Xilinx requires using locate constraints to preplace
such
>    connected GCLK/GCLKIO/DLL components.
> Total REAL time to Placer completion: 2 secs
> Total CPU time to Placer completion: 1 secs
>
> -----------------------------------------------------------------
>
> Following are the codes:
>
> module dll_standard (CLKIN, CLK2X, CLK4X, CLK8X, LOCKED);
> input CLKIN;
> output CLK2X, CLK4X, CLK8X, LOCKED;
>
> wire RESET;
> wire CLKIN_w, RESET_w, CLK2X_dll, CLK4X_dll, LOCKED2X, LOCKED4X;
> wire LOCKED2X_delay, RESET4X;
> wire LOCKED4X_delay, RESET8X;
> wire logic1;
> assign RESET = 1'b0;
>
> assign logic1 = 1'b1;
>
> assign CLKIN_w = CLKIN;
> assign RESET = RESET_w;
>
> CLKDLL dll2x (.CLKIN(CLKIN_w), .CLKFB(CLK2X), .RST(RESET_w),
>               .CLK0(), .CLK90(), .CLK180(), .CLK270(),
>               .CLK2X(CLK2X_dll), .CLKDV(), .LOCKED(LOCKED2X));
> BUFG   clk2xg (.I(CLK2X_dll),  .O(CLK2X));
> SRL16  rstsrl (.D(LOCKED2X), .CLK(CLK2X), .Q(LOCKED2X_delay),
>                .A3(logic1), .A2(logic1), .A1(logic1), .A0(logic1));
> assign RESET4X = !LOCKED2X_delay;
>
> CLKDLL dll4x (.CLKIN(CLK2X), .CLKFB(CLK4X), .RST(RESET4X),
>               .CLK0(), .CLK90(), .CLK180(), .CLK270(),
>               .CLK2X(CLK4X_dll), .CLKDV(), .LOCKED(LOCKED4X));
> BUFG   clk4xg (.I(CLK4X_dll),  .O(CLK4X));
> SRL16  rstsr2 (.D(LOCKED4X), .CLK(CLK4X), .Q(LOCKED4X_delay),
>                .A3(logic1), .A2(logic1), .A1(logic1), .A0(logic1));
> assign RESET8X = !LOCKED4X_delay;
> // assign CLK4X_B = ~ CLK4X;
>
> CLKDLL dll8x (.CLKIN(CLK4X), .CLKFB(CLK8X), .RST(RESET8X),
>               .CLK0(), .CLK90(), .CLK180(), .CLK270(),
>               .CLK2X(CLK8X_dll), .CLKDV(), .LOCKED(LOCKED_dll));
> BUFG   clk8xg (.I(CLK8X_dll),  .O(CLK8X));
>
> BUFG   lckpad (.I(LOCKED_dll), .O(LOCKED));
>
> endmodule
>
>
>



Article: 44478
Subject: Re: Power supply caps on PCB
From: John Larkin <John@0.com>
Date: Thu, 20 Jun 2002 20:13:40 -0700
Links: << >>  << T >>  << A >>
On Mon, 10 Jun 2002 12:57:14 +0200, "Noddy" <g9731642@campus.ru.ac.za>
wrote:

>Hi,
>
>I know that one should place 0.1uF and 0.01uF caps on all power supply pins
>on the FPGA... does this include both Vcco and Vccint, or can one get away
>with just Vccint and then put only 0.1uF on Vcco. I am seriously running out
>of room on my PCB, and increasing the PCB size is not an option!
>
>adrian
>
>

I've done some pretty serious TDR testing of multilayer boards with
different bypass caps installed. Most of my boards have a solid ground
plane and a solid Vcc plane separated by a thin (0.005 inch or so)
dielectric. Sometimes the power plane is chopped up: say, +5 mostly
but a few islands of 3.3 or whatever for FPGAs. I generally use four
bypass caps *per FPGA*. At high frequencies, the planes themselves are
the cap, and the ceramic caps just help out at lower frequencies and
are the gross charge storage reservoir. Actual cap placement doesn't
seem to matter as long as the planes are solid, the dielectric is
thin, and the caps are somewhere - basicly anywhere - on the planes. 

0.1 uf 0805 or 0603 is probably as good as any.

I've done lots of boards with way too many bypass caps, but so far
I've never done a multilayer board that had too few.

John


Article: 44479
Subject: Re: 12 years experience in Digital HW/ FPGA design, looking for job in the US
From: kayrock66@yahoo.com (Jay)
Date: 20 Jun 2002 22:46:04 -0700
Links: << >>  << T >>  << A >>
Ya, but the more people enter the job pool in your specific field, the
cheaper the salaries for everyone in that field will be.  Just simple
supply and demand.  Lets just call H1-B what it is: a labor subsidy
for high tech, approved by corrupt politicians.   Good for employers,
bad for employees (unless of course you're the one getting the visa). 
And don't buy that stupid ad in the paper BS.  The company placing the
ad already knows who they want to hire, and are just playing the game
so they don't get in trouble with a govt agency that doesn't give a
squat.

I don't blame the applicants, heck, I'd be first line if I lived in
the second world and someone offered me a way out.

I'm all for open borders and free trade, but its got to be for
everyone, not just MY specific field.  Otherwise my salary is bid down
(or equivalently held steady when it should be rising) while I have to
purchase goods and services in the higher priced controlled market. 
(e.g. If I have to compete with a guy from India at my job, when I go
home I should be able to hire a house keeper from Mexico for $5/hr.)

"jakab tanko" <jtanko@ics-ltd.com> wrote in message news:<aet1vn$msc$1@news.storm.ca>...
> Relax, Hater, if you are as qualified as this guy is
> you will get the job.....
> 
> jakab
> 
> Spam Hater <spam_hater_7@email.com> wrote in message
> news:3d101eab.4925791@64.164.98.7...
> >
> > Notice to potential employers:
> >
> > You will be displacing a US citizen who is qualified, and willing to
> > work for that salary.
> >
> > Make sure that you put a statement to that effect on the H1-B transfer
> > application.
> >
> > Nothing personal Farhad.  I have a family to feed.
> >
> >
> >
> > On 17 Jun 2002 21:18:46 -0700, farhad@everdream.com (Farhad Abdolian)
> > wrote:
> >
> > >Hi,
> > >I am currntly on H1-B visa, and since my current employer has decided
> > >to close our office, I am looking for a new job, and a company to take
> > >over my H1-B visa while my green card application goes through (my
> > >wife is American).
> > >
> >

Article: 44480
Subject: Re: fpga and ultra highspeed counters
From: kayrock66@yahoo.com (Jay)
Date: 20 Jun 2002 23:05:10 -0700
Links: << >>  << T >>  << A >>
It may be overpriced but can you design, build, debug ONE card for
that price?  As engineers I think we get caught up sometimes in part
cost or tend to be optimists on our ability to finish a task quickly. 
You kind of have to be to tackle some of the problems we get handed.

kolja@bnl.gov (Kolja Sulimma) wrote in message news:<25c81abf.0206161242.261776e4@posting.google.com>...
> This card will likely do the job for you. 
> But for your simple application it is extremely overpriced.
> 
> If I understand your application correctly, 24 counters of 24 Bits
> should fit into a Spartan-II 200 Device together with a PCI core and a
> FIFO.
> 
> Kolja Sulimma
> 
> 
> 
> "Pat Ford" <pat.ford@nrc.ca> wrote in message news:<aeco9f$oi8$1@moonstone.imsb.nrc.ca>...
> > They do BUT each card will only do 8 channels and the cost is high, and
> > they don't support the range of OS's that we are looking at.
> >  We are looking at the Nallatech
> >  Strathnuey kit with the XCV1000 fpga, any have opinions on this card?
> > http://www.nallatech.com/products/dime_select/strathnuey/index.asp
> > thanks for your help so far
> > Pat
> > 
> > "Jay" <kayrock66@yahoo.com> wrote in message
> > news:d049f91b.0206131216.10a6fa2d@posting.google.com...
> > > I don't want to spoil your fun but this sounds like something that
> > > might already be available.  Look at those PC intrumentation guys like
> > > National Instruments and the like, they may have something you can use
> > > or that can be gated.

Article: 44481
Subject: Re: How to get Unisims netlist?
From: "Anthony Ellis" <anthony@ams.co.za>
Date: Fri, 21 Jun 2002 08:06:33 +0200
Links: << >>  << T >>  << A >>
Yes tt is this output that I wan't but from the ISE4.2 tool kit.

Using the simprims libraries I do not have any timing info turned on so I
should get the same functional simulation as per the VHDL code.

"Ray Andraka" <ray@andraka.com> wrote in message
news:3D12109A.4460F72B@andraka.com...
> YOu can get a mapped synthesis output from the synthesizer which is
> constructed of unisim library elements.  In Synplify you turn on mapped
> output and it produces a .vhm (mapped VHDL or .vm mapped verilog output)
> which is a unisim primitive only realization of your design.  This can be
> useful to check the results of synthesis before going into the XIlinx
> tools, and it runs a bunch faster than the simprim xilinx output files.
>
> newman wrote:
>
> > Anthony,
> >   Good questions.  I have not thought too deeply about Unisim, but my
> > general belief is that they are there to allow instantiation of
> > functionality into your VHDL/ Verilog/ Schematic design, and provide
> > functional simulation capabilities.  After synthesis, one uses the
> > Simprim libraries, and I have never questioned that approach.  I
> > suspect that this library is required in order to back-annotate timing
> > delays from the Standard Delay Format (SDF) file for timing simulation
> > purposes.
> >
> >   Getting the X's out of a synthesized design may frusterate the new
> > user, and sometimes the veteran user as well.  One needs to be
> > cognisant of the timing constraints placed on the design by the UCF
> > file.  When one does functional simulations, one is almost dealing
> > with an ideally fast part.  After place and route, there are real
> > world timing delays, and the design may not work at 200 Mhz any more.
> >    Other issues to consider are:
> >    Are your memory storage elements getting initialized at the start
> > of your simulation?
> >    Is the setup and hold time of your testbench compliant with the UCF
> > file?
> >    Is your simulator default resolution set correctly?
> >    Are the Xilinx tools optimizing away portions of your design
> > because of coding errors?
> >
> >    One thing you could try would be to put in a divide by two test
> > circuit, and see if you can make that simulate after synthesis, and
> > gradually debug the circuits that are giving you problems.
> >
> > Hope this helps,
> >
> > Newman
>
> --
> --Ray Andraka, P.E.
> President, the Andraka Consulting Group, Inc.
> 401/884-7930     Fax 401/884-7950
> email ray@andraka.com
> http://www.andraka.com
>
>  "They that give up essential liberty to obtain a little
>   temporary safety deserve neither liberty nor safety."
>                                           -Benjamin Franklin, 1759
>
>



Article: 44482
Subject: Re: Multiple Nios CPU's on Altera PLD?
From: "Ryan" <ryans@cat.co.za>
Date: Fri, 21 Jun 2002 08:10:19 +0200
Links: << >>  << T >>  << A >>
HI

Thanks for the pointer. I have downloaded the file but I am having
difficulty opening each Nios core, ie SOPC won't load up. Any suggestions?

Thanks
Ryan

"Vincent JADOT" <vjadot@digitalsurf.fr> wrote in message
news:3d119950$0$24013$4d4eb98e@read.news.fr.uu.net...
> Hi
>
> there is an example to download of 2 nios on a chip (excalibur kit) at
this
> address:
> http://cnfm.cnfm.fr/ALTERA/ExcaliburN/Arbitrage.zip
>
> it 's on a french site, http://cnfm.cnfm.fr/ALTERA/Excalibur.htm, it's for
> examples and update for the ALTERA french university program. So it's in
> french, but i think it's understandable for english people.
>
> Excuse my english, i'm a french student.
>
>
>
>
>
> "Ryan" <ryans@cat.co.za> a écrit dans le message news:
> 3d12031a.0@obiwan.eastcoast.co.za...
> > Hi
> >
> > I am working with an Altera Excalibur Nios development board version
1.1.
> Is
> > it possible to configure the PLD (20K200) with more than 1 Nios cpu?
What
> I
> > would like to do is configure 1 cpu to be a master and the others as
> slaves.
> > The master will then control the processess executed on the slaves etc.
> > Ideally the slaves will not have to make use of external SRAM of FLASH.
Is
> > this concept possible and how can it be implemented? If not, have you
any
> > suggestions on what other route to consider?
> >
> > Thanks
> > Ryan
> >
> >
>
>



Article: 44483
Subject: Xilinx's 4.1i's Lastest webpack
From: muthu_nano@yahoo.co.in (Muthu)
Date: 20 Jun 2002 23:39:50 -0700
Links: << >>  << T >>  << A >>
Hi,

where can i get the xilinx's 4.1i's Latest webpack? and how to updatae the webpack?

Thanks and regards,
Muthu

Article: 44484
Subject: Retiming option in synplify pro
From: muthu_nano@yahoo.co.in (Muthu)
Date: 20 Jun 2002 23:56:30 -0700
Links: << >>  << T >>  << A >>
Hi,
In synplify_pro, to disable retiming option of a particular register i
did the below:


########################################################

reg       [1:0] data/* syn_allow_retiming=0 */;

#########################################################

But still synplify_pro is retiming this register data.

How can i avoid this?

Thanks and regards,
Muthu

Article: 44485
Subject: Re: Does anyone have experience with HandelC and Celoxica's RC1000 with VirtexE
From: ash@ashandnath.com (Ash)
Date: 21 Jun 2002 02:07:39 -0700
Links: << >>  << T >>  << A >>
Hi Antonio,

By default, DK1 will build a RAM out of distributed RAM, which it
appears to be doing here. If you want to use block RAM, then you need
to say so explicitly:

ram unsigned char RAM[160*120] with {block = 1};

That should solve your problem.

Ashley

Antonio Martínez Álvarez <amartinez@atc.ugr.es> wrote in message news:<3D0DC7A6.1010108@atc.ugr.es>...
> Hello.
> I can't work with a ram defined as:
>     ram unsigned char RAM[160*120];
> 
> dk1.1 gives an enormous EDIF an uses 10e6 logical gates. ???
> Xilinx Foundation crushes when attempting to do the mapping.
> We have VirtexE 2000ebg560-6 FPGA on a RC1000, and we use
> DK1.1.
> 
> I really need to use the internal FPGA RAM. I know how to use
> the external RC1000 SRAM (4 modules)... but don't know what's
> wrong with my code.
> 
> Could you please contact me or give me an example for doing that.
> I work on image processing. I can send my little code.
> 
> What I attempt to do:
>     1.- I have a frame in the host program. a 160*120 matrix if unsigned
> bytes.
>     2.- I send this matrix to SRAM
>     3.- I want to read the SRAM and write the matrix to the FPGA's RAM
>     4.- Manipulate the frame... for example adding 1 to every pixel...
> whatever (then we need to write a filter)
>     5.- Pass the new rewritten frame to the SRAM
>     6.- Then the host program loades the frame.
> 
> I've do that with SRAM. But FPGA's RAM doesn't work. Any suggestion
> 
> Thank you.

Article: 44486
Subject: Re: Multiple Nios CPU's on Altera PLD?
From: "Vincent JADOT" <vjadot@digitalsurf.fr>
Date: Fri, 21 Jun 2002 11:27:08 +0200
Links: << >>  << T >>  << A >>
Hi,

I can open each NIOS core with SOPC Builder on my system (win 2k and
quartus1.1).
I have install nios111 update (13 Mo) and quartus1.1 sp1full (91Mo) .
Probably, it's a nios 1.11 conflict (i'm not sure). You can download the
nios1.11 update on ALTERA site:
https://www.altera.com/support/software/download/sof-download_center.html





Article: 44487
Subject: StrongARM - 110 Model request
From: "Ulises Hernandez" <ulises@britain.agilent.com>
Date: Fri, 21 Jun 2002 11:38:45 +0100
Links: << >>  << T >>  << A >>
Hello.

Does anyone know about an StrongARM - 110 VHDL behavioral model? If possible
a free model and available from the web.
We have one from 1996 and it doesn't seem to be doing what I was expecting
according to the SA-110 datasheet. I checked on Intel
(http://appzone.intel.com/toolcatalog/arch.asp?architecture=10) but couldn't
find what I was looking for.

Thank you very much.

--
Ulises Hernandez




Article: 44488
Subject: Re: xilinx, jtag vs. serial parallel mode
From: "Cyrille de Brébisson" <cyrille_de-brebisson@hp.com>
Date: Fri, 21 Jun 2002 08:00:56 -0600
Links: << >>  << T >>  << A >>
Hello,

One of the major difference is going to be the programming speed. The
parallel mode would be faster...

regards, Cyrille


"emanuel stiebler" <emu@ecubics.com> wrote in message
news:3D127698.39F5AE01@ecubics.com...
> Hi,
> I was planning to download the bitstream from a uController
> via JTAG. (XAPP058). Because the uController has the databus (d0..d7)
> already connected to the d0..d7 pins of the FPGA (Spartan2e), would the
> slave parallel mode be the better solution ?
>
> Any problems I should expect from this ?
>
> cheers & thanks



Article: 44489
Subject: Re: Retiming option in synplify pro
From: Ken McElvain <ken@synplicity.com>
Date: Fri, 21 Jun 2002 07:03:11 -0700
Links: << >>  << T >>  << A >>


Muthu wrote:

> Hi,
> In synplify_pro, to disable retiming option of a particular register i
> did the below:
> 
> 
> ########################################################
> 
> reg       [1:0] data/* syn_allow_retiming=0 */;


Should have been:
   reg       [1:0] data/* synthesis syn_allow_retiming=0 */;

You could also apply this in the constraint file.



> 
> #########################################################
> 
> But still synplify_pro is retiming this register data.
> 
> How can i avoid this?
> 
> Thanks and regards,
> Muthu
> 


Article: 44490
Subject: Self upgrading Data I/O programmers?
From: dfnr2@yahoo.com (David)
Date: 21 Jun 2002 07:31:46 -0700
Links: << >>  << T >>  << A >>
Hello,

I have an Autosite programmer acquired on Ebay.  Because of their
business model, Data I/O is fairly tight with technical information. 
I am sure that there are a lot of autosites out there (it is the
programmer component of the the ProMaster production models).

Has anyone ever successfully self-upgraded one of these for (1) more
memory, or (2) a hard disk (MSM=mass storage module)?

I would imagine that these are standard parts.  I'd be very interested
in information anyone may have.

Thanks for any reply,

David.

Article: 44491
Subject: Re: Multiple Nios CPU's on Altera PLD?
From: "Wolfgang Loewer" <wolfgang.loewer@elca.de>
Date: Fri, 21 Jun 2002 16:43:51 +0200
Links: << >>  << T >>  << A >>
Hi,

you should switch to NIOS V2.0 or V2.1, the board is the same, you just need
the new SW. The newer versions have full SOPC builder support for
integrating multiple NIOS processors into a single system. Check out
application note 184 which is available on the Altera website. It discusses
the implementation of multiple masters into a NIOS system. These multiple
masters can also be multiple NIOS processors. For each processor you get
your own directory structure for SW development. With some mouse clicks you
can add multiple processors and specify what slaves they should share. Slave
side arbitration gets automatically built in for those slaves that can be
accessed by multiple masters and the arbitration scheme is configurable.
Furhtermore each master gets it's own Avalon bus segment and so mutliple
masters can simultaneously talk to different slaves. Arbitration only takes
place if multiple masters try to access the same slave in the same cycle.
When using V2.0 or newer and once you understood the concept it's really
just a question of minutes to put together a complex system with multiple
processors.

For your application you could have one master CPU that runs from external
memory. The slave CPUs could run from on-chip RAM or ROM. You could
implement a common on-chip memory that can be accessed from all CPUs in the
system and that's  being used to exchange data between the CPUs.
Alternatively you could also make the data memories from the slave CPUs
accessible to the master. The possible combinations are infinite and what
you described is certainly possible. V2.0 and newer supports a multi master
avalon bus structure and so it's a lot easier to implement than with 1.1 or
1.1.1.

Regards
Wolfgang
http://www.elca.de

"Ryan" <ryans@cat.co.za> schrieb im Newsbeitrag
news:3d12031a.0@obiwan.eastcoast.co.za...
> Hi
>
> I am working with an Altera Excalibur Nios development board version 1.1.
Is
> it possible to configure the PLD (20K200) with more than 1 Nios cpu? What
I
> would like to do is configure 1 cpu to be a master and the others as
slaves.
> The master will then control the processess executed on the slaves etc.
> Ideally the slaves will not have to make use of external SRAM of FLASH. Is
> this concept possible and how can it be implemented? If not, have you any
> suggestions on what other route to consider?
>
> Thanks
> Ryan
>
>



Article: 44492
Subject: Re: xilinx, jtag vs. serial parallel mode
From: Laurent Gauch <laurent.gauch@amontec.com>
Date: Fri, 21 Jun 2002 17:08:45 +0200
Links: << >>  << T >>  << A >>
Using slave parallel, you have to make sure your uc has the control on 
the other specific slave parallel mode signals like program signal, cclk 
signal ... (done to check the download). Without these signals you will 
be not able to build a slave parallel configuration access.

Laurent
www.amontec.com

emanuel stiebler wrote:

> Hi,
> I was planning to download the bitstream from a uController 
> via JTAG. (XAPP058). Because the uController has the databus (d0..d7)
> already connected to the d0..d7 pins of the FPGA (Spartan2e), would the 
> slave parallel mode be the better solution ?
> 
> Any problems I should expect from this ?
> 
> cheers & thanks
> 


Article: 44493
Subject: Xpower accuracy
From: Steven Derrien <sderrien@irisa.fr>
Date: Fri, 21 Jun 2002 17:39:22 +0200
Links: << >>  << T >>  << A >>
Hi,

	I would like to have some feed-bak regarding of the accuracy of 
the XPower tool from Xilinx. I have been working on a high-level power 
estimator for a few month, and I wish I could get some figure to compare
with Xpower with my model (I remember someone who said that the accuracy 
was about 5-10% but I can't find any reference).


Thanks,


Steven

Article: 44494
Subject: Coolrunner Orcad, Pads ChipScale packages?
From: "steve synakowski" <srs@twcny.rr.com>
Date: Fri, 21 Jun 2002 16:02:35 GMT
Links: << >>  << T >>  << A >>
Hi, Does anyone know if there are premade packages for Orcad capture
chipscale CPLD's XCR3064XL-CP56 and XCR3128XL-CS144. I couldn't find any,
but could just make them.
Pads footprints would be nice too, but again could make them, just saves
time.
Thanks,
Steve



Article: 44495
Subject: Re: How to get Unisims netlist?
From: Rick Filipkiewicz <rick@algor.co.uk>
Date: Fri, 21 Jun 2002 17:08:39 +0100
Links: << >>  << T >>  << A >>


newman wrote:

> Oops, I meant to say: After MAP.
>
> > functional simulation capabilities.  After synthesis, one uses the
>
> Newman

In fact you can, with 3.3i and later, use NGD2VER or NGD2VHDL to get a
simprims based simulation netlist from the .ngd file produced by
NGDBUILD. No SDF of course.

You can also get simprims based netlists for any of the submodules
linked together by NGDBUILD by going

NGDBUILD: .edn -> .ngo -> .ngd
NGD2<VER | VHDL>: .ngd -> <.v |.vhd>

but I've never tried using them. This might also work with XST's
brain-dead-binary output as well as the - 10% - sane EDIF.



Article: 44496
Subject: Re: Xilinx's 4.1i's Lastest webpack
From: Rick Filipkiewicz <rick@algor.co.uk>
Date: Fri, 21 Jun 2002 17:23:12 +0100
Links: << >>  << T >>  << A >>


Russell wrote:

> Muthu wrote:
> >
> > Hi,
> >
> > where can i get the xilinx's 4.1i's Latest webpack? and how to updatae the webpack?
>
> Isn't it 4.2i on the xilinx site?

Its a bit point & click tedious to get all the different packages so you might want to
consider using ``wget'' [available on Linux some modern Unix systems, its an FSF tool] to
do the HTTP downloads via a script (many thanks to the original poster on this NG who
suggested it)

wget --http-user <your user ID> --http-passwd <your passwd>
http://direct.xilinx.com/direct/webpack/FILE

where:

FILE = WebPACK_42wp20_installer.exe

or

FILE = WebPACK_42wp20_<a | b | c | ...>.xwp



Article: 44497
Subject: adding timing constraints
From: "cfk" <cfk_alter_ego@pacbell.net>
Date: Fri, 21 Jun 2002 16:54:11 GMT
Links: << >>  << T >>  << A >>
I guess this is really a design flow question. I am using ISE and
implementing a PCI interface in a VirtexE. I have now gotten to the stage
where I need to add some timing constraints. The question is, should those
timing constraints be added to the .UCF file under "Design Entry
Utilities>User Constraints>Edit Implementation Constraints File" or some
other file. I have tried using Constraints Editor, somewhat unsuccessfully
to add constraints so maybe part of the issue is I dont really know how to
add them properly. I can add IOSTANDARD's and LOC's with no problem to the
UCF, but now timing has become an issue. As always, any suggestions will be
greatly appreciated.

Charles



Article: 44498
Subject: Re: Help!I can't use the programmer of Max-plus II on windows XP.
From: Marcin E. Hamerla <mehamerla@pro.onet.pl>
Date: Fri, 21 Jun 2002 19:18:32 +0200
Links: << >>  << T >>  << A >>
Jim napisal(a):

>But I can't find the driver.
>
>I try to use the setup application under the directory "drivers" in the path
>where I installed MaxPlus+II
>but it looked useless.
>
>any other ideas? or Maxplus just doesn't support WinXP?

Try searching Altera website. I had the same problem with ByteBlaster
and I found good info on their www.

-- 
Pozdrowienia, Marcin E. Hamerla

"Watch carefully, I will do it only once."

Article: 44499
Subject: Logic Minimization in Max+Plus II compiler
From: jfu1650@hotmail.com (Johnny Fu)
Date: 21 Jun 2002 10:41:19 -0700
Links: << >>  << T >>  << A >>
Hi,

I am trying to find a way to prevent the Max+Plus II 9.4 compiler from
performing any logic minimization on a VHDL design similar to the one
below:

       A         B           C
input ----|>*-------|>*--------|>*----- output

Using Synplify 7.1, I can synthesize the design and keep all the
appropriate nets.  However, even after specifying a WYSIWYG global
logic synthesis compilation option in Max+PlusII; taps A, B, and C are
no longer connected together and appear as separate input and output
paths.  The logic for B and C are also minimized as B is driven purely
by the input (and the internal and-or-xor structure) without the
presence of the two inverters before it and C is driven by a single
inverter instead of three.  Therefore, A, B, and C in a time
simulation all change value at the same time without any delay between
them (since they are no longer connected). Are there any other options
that I can try for preventing logic minimization?  I am compiling for
an Altera Max7000AE chip.  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
2019JanFebMar2019

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