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 7425

Article: 7425
Subject: Re: HELP: FIFO's on an FPGA
From: Ray Andraka <no_spam_randraka@ids.net>
Date: Tue, 09 Sep 1997 12:03:37 -0400
Links: << >>  << T >>  << A >>

> Ray Andraka's reply (see other thread to original posting). I just need
> to figure out the logic to latch the write request, process the message,
> and then negate the write request until the next one comes through.
> Hmmm...
> 
Use a T flip flop clocked by the write clock and enabled by the write
enable. The data should go into a register clocked by the write clock
and enabled by the write enable.  The T flop will toggle each time a
write is performed.  The output of this feeds a four state state
machine:


present state   toggle   next state
 00		  0        00
 00               1        01
 01               X        11
 11               0        10
 11               1        11
 10               X        00


decode states 10 and 01 (XOR or the state bits) to create a write pulse
that is synchronized to the RAM clock.  The state machine described
needs no additional synchronizing flops for most systems since only one
bit is enabled for toggling at any time and the toggle sense states are
separated by intermediate states.  The only real restriction is the RAM
clock has to be a lot faster than the write rate to ensure the holding
register is emptied before the next write occurs.     
 

-Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email randraka @ ids . net
http://users.ids.net/~randraka
Article: 7426
Subject: Re: HELP: FIFO's on an FPGA
From: Bryn Wolfe <bwolfe@hypercon.com>
Date: Tue, 09 Sep 1997 10:00:57 -0700
Links: << >>  << T >>  << A >>



Thanks for your input, Ray. Your suggestion woke me up to the simplicity
of this process. For my design the read processing will occur at a
higher frequency, so I can run the FIFO at that rate, and latch the
write requests for subsequent processing by the FIFO at FIFO rates. The
write pulser is basically just a JKFF followed by a pair of DFFs, with
an XOR of the outputs of the two D "flops", as you call them, to
generate the synchronized write request. With J and K tied high and the
input write request as a clock, the JKFF does the toggle, while the
subsequent DFF pair translates the signal to the FIFO time domain. The
XOR generates a write request that is only one clock period long, which
is what the FIFO needs to write the data.


If you see any bugs with this, please say so. I think your suggestion
pulled me out of a logical rut.


-- 
Bryn Wolfe - Robotics Engineer
Metrica TRAC Labs
Article: 7427
Subject: Job-USA; Charlotte, NC; Electrical engineer; FPGA, DSP; 50 Mghtz
From: richard_steinman@cmagroup.com
Date: 9 Sep 1997 18:32:11 GMT
Links: << >>  << T >>  << A >>



USA; Charlotte, NC; Electrical engineer; FPGA, DSP; 50 Mghtz


Experience in embedded systems design using: FPGAs (Altera preferred), 
DSPs and PLDs; Systems level architecture; Sensitive high speed analog 
designs; and signal distribution. Control systems familiarity highly desirable. 




Please refer to JO# 2793RJS in your response.






Richard Steinman
Team Leader
rjs@cmagroup.com
IT & Software Solutions Team
Career Marketing Associates
http://www.cmagroup.com/IT.html
Article: 7428
Subject: Re: HELP: FIFO's on an FPGA
From: Bryn Wolfe <bwolfe@hypercon.com>
Date: Tue, 09 Sep 1997 12:04:55 -0700
Links: << >>  << T >>  << A >>



Peter Alfke wrote:
> 
> Maybe standard terminology is lacking, but I would be interested how a
> "non-clocked FIFO" could possibly work. Something has to tell the device
> to accept data, and something else has to tell it to spit out the next
> word. I call those signals clocks. You can call them "handshake".
> 
> Just semantics.
> 
> Peter Alfke, Xilinx


The best example I have of a non-clocked FIFO is a COTS chip, the 
TI SN74ACT7203 2048x9 Asynchronous FIFO. It has no clock, but depends on
separate read and write signals to indicate when a transaction should
occur. Granted, the rising (or in this case, falling) edge "clocks" the
circuit, but there is no reference clock on either side to say whether
the read or write signal lines are valid. Consequently, with this
asynchronous operation, the read signal must transition (high to low) to
indicate a read, whereas a synchronous FIFO could leave the read enable
signal asserted and let the read clock initiate independed read
operations.


As to which one is better, that depends on the use of the circuit. TI
produces both kinds. In fact, they have Asynchronous, clocked, strobed
and synchronous FIFOs, all of which say that the read and write
operations can be asynchronous or coincident. So, let me think,
asynchronous refers to the relationship of a read or write operation to
a reference clock AND asynchronous refers to the timing relationship
between independent read and write operations. Aye caramba! 


Hope you can understand my confusion with the words. Thanks for your
input.


-- 
Bryn Wolfe - Robotics Engineer
Metrica TRAC Labs
Article: 7429
Subject: Re: HELP: FIFO's on an FPGA
From: jhallen@world.std.com (Joseph H Allen)
Date: Tue, 9 Sep 1997 19:44:08 GMT
Links: << >>  << T >>  << A >>

In article <341580C9.EBCE1109@hypercon.com>,
Bryn Wolfe  <bwolfe@hypercon.com> wrote:


>Thanks for your input, Ray. Your suggestion woke me up to the simplicity
>of this process. For my design the read processing will occur at a
>higher frequency, so I can run the FIFO at that rate, and latch the
>write requests for subsequent processing by the FIFO at FIFO rates. The
>write pulser is basically just a JKFF followed by a pair of DFFs, with
>an XOR of the outputs of the two D "flops", as you call them, to
>generate the synchronized write request. With J and K tied high and the
>input write request as a clock, the JKFF does the toggle, while the
>subsequent DFF pair translates the signal to the FIFO time domain. The
>XOR generates a write request that is only one clock period long, which
>is what the FIFO needs to write the data.


This is dangerous.  Your write pulse might be shortened if the first D
flip-flop is clocked right at the edge where the toggler transistions (this
is called a meta-state instability problem).  To fix it, insert another
D-flip-flop before the two D-flip-flops.  If this flip-flop captures during
the transistion, it's output will settle before the next flip-flop sees the
signal (actually it may not, but the probability of hitting the second
flip-flop's capture window is very small, which is just as good).
-- 
/*  jhallen@world.std.com (192.74.137.5) */               /* Joseph H. Allen */
int a[1817];main(z,p,q,r){for(p=80;q+p-80;p-=2*a[p])for(z=9;z--;)q=3&(r=time(0)
+r*57)/7,q=q?q-1?q-2?1-p%79?-1:0:p%79-77?1:0:p<1659?79:0:p>158?-79:0,q?!a[p+q*2
]?a[p+=a[p+=q]=q]=q:0:0;for(;q++-1817;)printf(q%79?"%c":"%c\n"," #"[!a[q-1]]);}
Article: 7430
Subject: Re: Which FPGA ?
From: s_clubb@netcomuk.co.uk (Stuart Clubb)
Date: Tue, 09 Sep 1997 21:32:28 GMT
Links: << >>  << T >>  << A >>



On 8 Sep 1997 18:01:52 GMT, "Rich K." <rich.katz@gsfc.nasa.gov> wrote:


>at&t data book shows att3030 in either 44 or 68 pin and 3,000 gates and 100
>clb's.  each clb has two f-f's according to the picture and f-f's in the
>i/o blocks also which are real handy either directly or by unused pins. 
>i'm sure if i get this wrong we'll hear from our good friend across the
>pond!


Does he mean me? ;-)


You need a new _Lucent_ book Rich.


Stuart


>(vendors arranged in alphabetical order from lower to higher - next one
>will be reverse order :-)


-- Joke Press Release --


Zwitterion Systems - High quality FPGA vapourware, available
soon(ish). Real expensive now, but Zwitterion expects them to be so
cheap, they project volume pricing of under a buck for by Q4/2000.


(Add safe harbour BS get-out clause here)


-- Ends --


Sorry, bad day at the office, needed some light relief.
Article: 7431
Subject: Re: HELP: FIFO's on an FPGA
From: s_clubb@netcomuk.co.uk (Stuart Clubb)
Date: Tue, 09 Sep 1997 21:32:29 GMT
Links: << >>  << T >>  << A >>

On Tue, 09 Sep 1997 12:04:55 -0700, Bryn Wolfe <bwolfe@hypercon.com>
wrote:


>The best example I have of a non-clocked FIFO is a COTS chip, the 
>TI SN74ACT7203 2048x9 Asynchronous FIFO. It has no clock, but depends on
>separate read and write signals to indicate when a transaction should
>occur. Granted, the rising (or in this case, falling) edge "clocks" the
>circuit, but there is no reference clock on either side to say whether
>the read or write signal lines are valid. Consequently, with this
>asynchronous operation, the read signal must transition (high to low) to
>indicate a read, whereas a synchronous FIFO could leave the read enable
>signal asserted and let the read clock initiate independed read
>operations.


Both Lucent and Xilinx DPRAM elements should be able to implement such
an asynchronous FIFO as you describe.


The write signal would be a processor wr_bar line, and would feed the
clock input of the DPRAM and the clock input of your write address
counter. Both would run on the same rising edge so you would setup
your data relative to the rising edge of wr_bar, and at the edge, the
write would be synchronised to write into the RAM (self clocking), and
at the same time increment the counter ready for the next write cycle.


Reading is done asynchronously so a counter feeds the read address
port, and is incremented on the rising edge of your rd_bar signal.
rd_bar asserted (low) would turn on your output buffer etc. so the
data is instantly readable. Throw in a bit of logic to do full/empty
flags and "Hey presto", a truly asynchronous FIFO in all senses of the
description (I think).


Only fly in the ointment for Xilinx is the 1mS maximum high time on
their RAM clock input.


Stuart


Article: 7432
Subject: Re: HELP: FIFO's on an FPGA
From: s_clubb@netcomuk.co.uk (Stuart Clubb)
Date: Tue, 09 Sep 1997 21:32:30 GMT
Links: << >>  << T >>  << A >>

On Tue, 09 Sep 1997 10:00:57 -0700, Bryn Wolfe <bwolfe@hypercon.com>
wrote:


>If you see any bugs with this, please say so. I think your suggestion
>pulled me out of a logical rut.


If you are reading (asynchronously?) on every clock cycle, when do you
plan to write to the "synchronous" SRAM that runs on your read clock?
You would have to run the read clock at least double the read rate, so
that you can fit in any pending write requirement so that they appear
to be transparant to the system.


Stuart
Article: 7433
Subject: wanted: experiences with VCC "HOTworks"
From: Andy Wilson <andyw@ibeam.intel.com>
Date: Tue, 09 Sep 1997 15:30:35 -0700
Links: << >>  << T >>  << A >>

greetings.  i'd be interested in hearing people's experiences,
good, bad, or indifferent, with the Virtual Computer Corporation
"HOTworks" development system.  {This is a $995 PCI board for
reconfigurable computing based on the Xilinx 6216 RPU.}  How
well do the development tools work?


thanks


atw
not speaking for intel....
Article: 7434
Subject: Re: Which FPGA ?
From: "Richard B. Katz" <stellere_nospam@erols.com>
Date: 9 Sep 1997 23:21:56 GMT
Links: << >>  << T >>  << A >>



Stuart Clubb <s_clubb@netcomuk.co.uk> wrote in article
<3415bb76.4920844@nntp.netcruiser>...
> 
> On 8 Sep 1997 18:01:52 GMT, "Rich K." <rich.katz@gsfc.nasa.gov> wrote:
> 
> >at&t data book shows att3030 in either 44 or 68 pin and 3,000 gates and
100
> >clb's.  each clb has two f-f's according to the picture and f-f's in the
> >i/o blocks also which are real handy either directly or by unused pins. 
> >i'm sure if i get this wrong we'll hear from our good friend across the
> >pond!
> 
> Does he mean me? ;-)


****
**** you answered, didn't you? ;^)
****


> 
> You need a new _Lucent_ book Rich.


****
**** er, no offense intended!
****
**** but did get a new job (in other life, this from home 'puter)
****




> 
> Stuart
> 
> >(vendors arranged in alphabetical order from lower to higher - next one
> >will be reverse order :-)




****
**** after i do higher to lower alphabetical order, i planned from going
**** from the middle out to be fair.  by this time, my new *lucent* book
**** will be here and they'll be right up there; good timing stu!
**** but you'll be behing motorola.
****


> 
> Sorry, bad day at the office, needed some light relief.
> 


****
**** as our leader says, "i feel your pain"
****
Article: 7435
Subject: Re: HELP: FIFO's on an FPGA
From: Ray Andraka <no_spam_randraka@ids.net>
Date: Wed, 10 Sep 1997 00:02:32 -0400
Links: << >>  << T >>  << A >>

Joseph H Allen wrote:
> 
> This is dangerous.  Your write pulse might be shortened if the first D
> flip-flop is clocked right at the edge where the toggler transistions (this
> is called a meta-state instability problem).  To fix it, insert another
> D-flip-flop before the two D-flip-flops.  If this flip-flop captures during
> the transistion, it's output will settle before the next flip-flop sees the
> signal (actually it may not, but the probability of hitting the second
> flip-flop's capture window is very small, which is just as good).
> --


I am assuming synchronous logic, such as what you get using the Altera
EAB as a synchronous RAM.  In that case, as long as the shortened pulse
meets the setup time of the write enable flop on the EAB there is not a
problem (the pulse shortening effectively delays the leading edge of the
synchronized toggle pulse due to the metastable recovery time).  The
probability of a metastable upset in a given period of time is a
function of the clock frequencies of both clock domains and of the
metastable recovery characteristic of the flop (which is related to the
max toggle rate).  The metastability characteristics of most FPGAs are
pretty good, so in most applications the single level synchronization is
sufficient (assuming a synchronous target for the generated pulse).
However, in cases where clock latency is not an issue the extra flop
certainly won't hurt. For the FIFO, the extra flop will eat into the
minimum write cycle.  In this case, I believe you will get more bang for
the buck by lowering the clock frequency on the RAM domain instead of
adding an extra latency stage. That can be accomplished by using a
fraction of the RAM clock to clock the synchronizing circuit


-Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email randraka @ ids . net
http://users.ids.net/~randraka
Article: 7436
Subject: Large FPGA
From: Martin Vorbach <Martin.Vorbach@SCRAP.de>
Date: Wed, 10 Sep 1997 10:23:46 +0200
Links: << >>  << T >>  << A >>



> When I served a booth at EuroDac '96, across from us there
> was a small European FPGA company that was touting a 20x30mm
> FPGA die.
> 
> I would like a contact if anyone knows who I am talking about.
> Anyone knowing of an alternative largish source would
> be greatly appreciated as well.
> 
> Many thanks.
> 
> 
	[M.Vorbach]  The last thing I heard was, that this company
doesn´t exist any longer. 
	Their chip vendor was not able to implement the design.








Article: 7437
Subject: Re: HELP: FIFO's on an FPGA
From: jhallen@world.std.com (Joseph H Allen)
Date: Wed, 10 Sep 1997 09:48:49 GMT
Links: << >>  << T >>  << A >>



In article <34161BD7.42B9@ids.net>,
Ray Andraka  <no_spam_randraka@ids.net> wrote:
>Joseph H Allen wrote:


>> This is dangerous.  Your write pulse might be shortened if the first D
>> flip-flop is clocked right at the edge where the toggler transistions (this
>> is called a meta-state instability problem).  To fix it, insert another
>> D-flip-flop before the two D-flip-flops.  If this flip-flop captures during
>> the transistion, it's output will settle before the next flip-flop sees the
>> signal (actually it may not, but the probability of hitting the second
>> flip-flop's capture window is very small, which is just as good).


>I am assuming synchronous logic, such as what you get using the Altera
>EAB as a synchronous RAM.  In that case, as long as the shortened pulse
>meets the setup time of the write enable flop on the EAB there is not a
>problem (the pulse shortening effectively delays the leading edge of the
>synchronized toggle pulse due to the metastable recovery time).  The
>probability of a metastable upset in a given period of time is a
>function of the clock frequencies of both clock domains and of the
>metastable recovery characteristic of the flop (which is related to the
>max toggle rate).  The metastability characteristics of most FPGAs are
>pretty good, so in most applications the single level synchronization is
>sufficient (assuming a synchronous target for the generated pulse).
>However, in cases where clock latency is not an issue the extra flop
>certainly won't hurt. For the FIFO, the extra flop will eat into the
>minimum write cycle.  In this case, I believe you will get more bang for
>the buck by lowering the clock frequency on the RAM domain instead of
>adding an extra latency stage. That can be accomplished by using a
>fraction of the RAM clock to clock the synchronizing circuit


You are probably right, but I would still be careful.  Feeding one flip flop
into another has a very low probably for a double meta-instability event
since the second flip-flop's capture window is very short (.1ps if I
remember correctly).  If you instead feed the shortened pulse to multiple
inputs, the clock-skew between different ram bits becomes your window: on an
fpga this is in the 400ps range (for xilinx if you use a global clock
buffer), so things are vastly worse.  Still if you are running at 30MHz or
less, you're probably ok.


There are other games you can play to reduce the pipeline delay: clock the
inserted flip-flop with the negative clock edge.


-- 
/*  jhallen@world.std.com (192.74.137.5) */               /* Joseph H. Allen */
int a[1817];main(z,p,q,r){for(p=80;q+p-80;p-=2*a[p])for(z=9;z--;)q=3&(r=time(0)
+r*57)/7,q=q?q-1?q-2?1-p%79?-1:0:p%79-77?1:0:p<1659?79:0:p>158?-79:0,q?!a[p+q*2
]?a[p+=a[p+=q]=q]=q:0:0;for(;q++-1817;)printf(q%79?"%c":"%c\n"," #"[!a[q-1]]);}
Article: 7438
Subject: Re: Which FPGA ?
From: "Rich K." <rich.katz@gsfc.nasa.gov>
Date: 10 Sep 1997 13:44:40 GMT
Links: << >>  << T >>  << A >>

Stuart Clubb <s_clubb@netcomuk.co.uk> wrote in article
<3415bb76.4920844@nntp.netcruiser>...
> On 8 Sep 1997 18:01:52 GMT, "Rich K." <rich.katz@gsfc.nasa.gov> wrote:
> 
> >at&t data book shows att3030 in either 44 or 68 pin and 3,000 gates and
100
> >clb's.  each clb has two f-f's according to the picture and f-f's in the
> >i/o blocks also which are real handy either directly or by unused pins. 
> >i'm sure if i get this wrong we'll hear from our good friend across the
> >pond!
> 
> Does he mean me? ;-)
> 
> You need a new _Lucent_ book Rich.
> 
> Stuart




**** got on the *lucent* www site last nite, couldn't find where to order
one.
**** can you supply a www address, e-mail address, or a phone # where i can
order one
**** up?


**** thanks,


**** rk
Article: 7439
Subject: Re: hdtv interpolation and decimation
From: ltest@rte9-sun_5.5.1 (news_check.py)
Date: Wed, 10 Sep 1997 14:10:21 GMT
Links: << >>  << T >>  << A >>

In article <340D7061.1078@compuserve.com> Chad Bearden <ebeard@compuserve.com> wrote:
> Hello,
> 
> I'm looking for some info/parts to do the job of interpolating and


Hello,


I'm looking for some info/parts to do the job of interpolating and
decimating a US/Euro HDTV digital stream.  In the past I've used the
Raytheon TMC2242 to do this on CCIR streams.  any suggestions?


chad.
Article: 7440
Subject: Re: fpga configuration over PCI
From: ltest@rte9-sun_5.5.1 (news_check.py)
Date: Wed, 10 Sep 1997 14:16:07 GMT
Links: << >>  << T >>  << A >>



In article <EFrLGD.AF6@world.std.com> jhallen@world.std.com (Joseph H Allen) wrote:
> In article <5u7v21$5d8@neptune.myri.com>,
> Wen-King Su <wen-king@myri.com> wrote:
> >In a previous article jhallen@world.std.com (Joseph H Allen) writes:


In article <5u7v21$5d8@neptune.myri.com>,
Wen-King Su <wen-king@myri.com> wrote:
>In a previous article jhallen@world.std.com (Joseph H Allen) writes:
>:
>;It would be really nice if PCI controller FPGAs could be configured over the
>:PCI bus.  One way of doing it is to use configuration writes with the fpga
>;data encoded in the address.


>Why not just do it through JTAG.


I don't know anything about JTAG or JTAG on the PC.  Is it board by board
addressable?  I'll have to learn more about it.
-- 
/*  jhallen@world.std.com (192.74.137.5) */               /* Joseph H. Allen */
int a[1817];main(z,p,q,r){for(p=80;q+p-80;p-=2*a[p])for(z=9;z--;)q=3&(r=time(0)
+r*57)/7,q=q?q-1?q-2?1-p%79?-1:0:p%79-77?1:0:p<1659?79:0:p>158?-79:0,q?!a[p+q*2
]?a[p+=a[p+=q]=q]=q:0:0;for(;q++-1817;)printf(q%79?"%c":"%c\n"," #"[!a[q-1]]);}
Article: 7441
Subject: Re: Which FPGA ?
From: s_clubb@netcomuk.co.uk (Stuart Clubb)
Date: Wed, 10 Sep 1997 14:47:59 GMT
Links: << >>  << T >>  << A >>



On 10 Sep 1997 13:44:40 GMT, "Rich K." <rich.katz@gsfc.nasa.gov>
wrote:


>**** can you supply a www address, e-mail address, or a phone # where i can
>order one


email me a shipping address, and I will have it mailed to you.


Regards
Stuart


Article: 7442
Subject: Re: University FPGA Project
From: "Steven K. Knapp" <sknapp @ optimagic.com>
Date: 10 Sep 1997 15:35:02 GMT
Links: << >>  << T >>  << A >>

Because you are working with Xilinx FPGAs, the Xilinx web site is probably
a good start 'http://www.xilinx.com'.  However, you may also find various
interesting and applicable information on The Programmable Logic Jump
Station at 'http://www.optimagic.com', in particular, the Research pages at
'http://www.optimagic.com/research.html'.
-- 
Steven Knapp
OptiMagic, Inc.
E-mail:  sknapp @ optimagic.com
Programmable Logic Jump Station:  http://www.optimagic.com


KITT <gillette@gillette.demon.co.uk> wrote in article
<3413126D.6B50@gillette.demon.co.uk>...
| Hi, my names Phillip and I'm going to be starting a final year
| university project involving the Xilinx FPGA within the next few months. 
| 
| I don't have much knowledge of FPGA's at the moment, would someone take
| the time to either copy some information to this newgroup, or point me
| to a site on the web please?
| 
| Thanks in advance,
| 
| Phil Renfield
| 
Article: 7443
Subject: Reading Viewlogic files
From: Simon Bacon <SimonBacon@tile.cut_this.demon.co.uk>
Date: Wed, 10 Sep 97 16:00:23 GMT
Links: << >>  << T >>  << A >>

I have been given some design files in Viewlogic format.  I only
have OrCAD and Protel.  Is there a utility around which can convert
the Viewlogic stuff to something I can read?




Path: ix.netcom.com!news.webspan.net!newsfeed.internetmci.com!newsfeed.internetmci.com!206.229.87.25!news-peer.sprintlink.net!news.sprintlink.net!Sprint!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!newsfeed.ecrc.net!news.new-york.net!news.columbia.edu!osiris.giss.nasa.gov!newsfeed.gsfc.nasa.gov!usenet

Article: 7444
Subject: Text Book and VHDL Simulator $50.00
From: Richard Schwarz <aaps@erols.com>
Date: Wed, 10 Sep 1997 12:49:14 -0400
Links: << >>  << T >>  << A >>



For those looking for a solid VHDL Text Book and a useable VHDL
Simulator,
APS is now selling the Prentice Hall text by Pellerin and Taylor, along
with a limited
version of PeakVHDL simulator (perfect for learning) for $49.99. Some
details on the
text are shown below see the full details at:


http://www.associatedpro.com/aps


VHDL Made Easy! is over 400 pages packed full of VHDL tips and examples,
and includes:


     Strong focus on synthesis, and on verification of synthesizable
designs.
     Clear, understandable descriptions of topics such as concurrency,
hierarchy, and abstraction levels.
     Straightforward, tutorial approach to VHDL; without the dull,
sleep-inducing descriptions and obscure syntax rules found in most
     other VHDL books.
     Dozens of examples, large and small, to help you get started with
this powerful hardware description language.
     Information about the latest VHDL standards, including IEEE
1076-1993, 1164, 1076.3 and 1076.4 (VITAL).
     Tips on how to get the most out of synthesis, including
descriptions of the most common synthesis coding errors.
     Descriptions of how to write test benches, including how to use
VHDL's text I/O features to read and write files.
     A complete, plain English VHDL keyword reference.


CD-ROM Bonus!


Also included with VHDL Made Easy! is a CD-ROM containing the following:


     A limited edition version of the PeakVHDL Personal Edition
simulator, allowing you to write and test small VHDL designs.
     Perfect for learning!
     Demonstration versions of other commercial VHDL products, including
products from Accolade Design Automation, Visual
     Software Solutions, SynaptiCAD, Capilano Computing and others.
     Dozens of VHDL source files that you can copy and modify for your
own needs.


--
__/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/


Richard Schwarz, President              EDA & Engineering Tools
Associated Professional Systems (APS)   http://www.associatedpro.com
3003 Latrobe Court                      richard@associatedpro.com
Abingdon, Maryland 21009
Phone: 410.569.5897                     Fax:410.661.2760


__/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/




Article: 7445
Subject: Re: Reading Viewlogic files
From: "Rich K." <rich.katz-nospam@gsfc.nasa.gov>
Date: 10 Sep 1997 17:11:35 GMT
Links: << >>  << T >>  << A >>

hi,


viewoffice (pc version is what i have) has an edif schematic writer - does
that help?


rk


Simon Bacon <SimonBacon@tile.cut_this.demon.co.uk> wrote in article
<873907223snz@tile.CUT_THIS.demon.co.uk>...
> I have been given some design files in Viewlogic format.  I only
> have OrCAD and Protel.  Is there a utility around which can convert
> the Viewlogic stuff to something I can read?


Article: 7446
Subject: Re: HELP: FIFO's on an FPGA
From: Bryn Wolfe <bwolfe@hypercon.com>
Date: Wed, 10 Sep 1997 10:50:08 -0700
Links: << >>  << T >>  << A >>



Stuart Clubb wrote:
> 
> On Tue, 09 Sep 1997 10:00:57 -0700, Bryn Wolfe <bwolfe@hypercon.com>
> wrote:
> 
> >If you see any bugs with this, please say so. I think your suggestion
> >pulled me out of a logical rut.
> 
> If you are reading (asynchronously?) on every clock cycle, when do you
> plan to write to the "synchronous" SRAM that runs on your read clock?
> You would have to run the read clock at least double the read rate, so
> that you can fit in any pending write requirement so that they appear
> to be transparant to the system.
> 
> Stuart
Actually, my FIFO implementation uses interleaved memory blocks, so that
reads and writes can occur simultaneously. A pair of staging registers
on the read side allow me to instantly read the current first-out (FO
part of FIFO) value. When the read switches to the other staging
register, the one just read is being loaded with the next value. This is
kind of a prefetch operation, in caching terms. It's impossible for a
read and a write to collide. There's a fairly good writeup of the logic
in Altera's application note AN 66 (Implementing FIFO Buffers in FLEX
10K Devices) which you can get from their web site at
http://www.altera.com/html/literature/litprod1.html#f10k. It's pretty
slick!


Did I answer your question? Hope so.


-- 
Bryn Wolfe - Robotics Engineer
Metrica TRAC Labs
Article: 7447
Subject: Opportunities for FPGA Designers
From: Dave NewKirk <dave.newkirk@xaqti.com>
Date: Wed, 10 Sep 1997 11:00:51 -0700
Links: << >>  << T >>  << A >>



XaQti Corporation:


If you think that putting "Systems-on-a-Chip" is the direction the
Electronics Industry is headed; and if you believe that the major future
growth of the Semiconductor Industry is in networking; then you will
want
to work at XaQti Corporation.


XaQti is seeking a few key individuals with strong networking
architecture
backgrounds; hardware ASIC designers, systems and board engineers and
Software Engineers who will help us fulfill the promise of putting the
"Network-on-a-CHIP"(TM).


XaQti is privately held and well financed.  We offer competitive
salaries,
excellent benefits and employee incentive stock options.  If you are
interested in helping to bring the next generation of Ethernet prodcuts
to
the market, please email your resume to hr@xaqti.com








David C. Newkirk
XaQti Corporation
Web: http://www.xaqti.com
"Network-on-a-CHIP" TM
Article: 7448
Subject: Re: daisy-chained bitstreams
From: Andreas Kugel <andreas.kugel@informatik.uni-mannheim.de>
Date: Thu, 11 Sep 1997 07:04:05 GMT
Links: << >>  << T >>  << A >>
John Archambeault schrieb:

>         Oh yeah.  If forgot.  If you are TRULY masochistic, you could
> convert
> the bitstream to a text representation of the binary.  (so it looks on
> a
> normal text viewer as 001110011100, etc).  Then manually append the
> bitstream
> for each fpga, removing the header for each and modifying the header
> for the
> first to include to total number of bits (of the total bitstream) and
> then
> either covert the "text 0110011" to real binary or just leave it in
> the "text
> 0110011" and download it.
>         With patience you could get a small script (program) to do it,
> but (for
> us at least) it was much simpler to do what I said before.  (copied
> below)
>
>         Enjoy,
>         John
>
>
> I had the same problem only with 3100A.  The way around it is a bit
> convaluted,
> but it does work.  Use the prom makeing tool to set it up as if a
> monster PROM
> was going to perform a serial download.  Then, juse feed that PROM
> file into
> the XCHECKER (name?) PC download cable.  It will accept PROM files.
>
> Good Luck,
> John
>
> BTW:  I documented this a while back, but am going from memory right
> now.  If
> you still can't get it, e-mail and I'll check the old documents.
> --
> John Archambeault
> End of article 7946 (of 7957) -- what next? [npq]
>
> --
> John Archambeault


I didn't read the original message, however if it concerns assembling
daisy-chain
bitstreams for serial download (slave serial or JTAG) for XC4000 (or
3000)
from a CPU, we have a small tool to do that.Code is in C.
Mail if you need it.

Andreas
--
Andreas Kugel - University of Mannheim - Dept. of Computer Science V
B6,26 - 68131 Mannheim - Germany
Phone:+(49)621 292 1634 - Fax:+(49)621 292 5756
mailto:kugel@mp-sun1.informatik.uni-mannheim.de
http://www-mp.informatik.uni-mannheim.de


Article: 7449
Subject: VHDL/Verilog Trainers Required in US and/or UK
From: Simon Moreton <simon@esperan.com>
Date: Thu, 11 Sep 1997 15:55:56 +0100
Links: << >>  << T >>  << A >>



Esperan specialize in supplying independent, application orientated VHDL 
and
Verilog based training. We have very close relationships with most of 
the
main EDA and FPGA vendors, and our courses are recommended or re-sold by
many of them.


We are rapidly expanding our training services in the US and Europe, and 
we
are seeking further trainers who we can outsource the delivery of 
training
courses to on a week by week basis.


Suitable candidates need the following qualifications:


* Real world experience of VHDL and/or Verilog based design using 
synthesis.
Use of FPGAs would be a distinct advantage.


* Good communication skills and the ability to explain complex topics in 
a
clear, concise manner.


* Availability for periods of up to one week for between four and ten 
weeks
per year.


* A willingness to travel.


In return, we can help to heighten the awareness of yourself and your
organisation within the electronic design community and broaden your
knowledge of HDL based design techniques.


If you are interested in this opportunity then please contact me by 
phone,
email or fax.


==============================
Simon Moreton
Esperan
Phone: (800) 947 6116
Fax:   (800) 766 3895
Email: simon@esperan.com
WWW:   http://www.esperan.com
==============================


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