Custom Search
Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
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
peter: : :There are lots of stupid things in NT but none of them make it in any : ;way unsuitable for EDA. wen: : Oh yes. For one, it prevents me from using the tools if I am not sitting : in front of of the machine. rk: solution: sit in front of the machine. i think nt (and probably linux) is winner here over unix since with nt, you can simply fold up the machine, put it in your backpack, hop on the plane, wait for the announcement that you can pull out your 'puter, and get back to work. i travel a modest amount on business, 6-8 times per year, and i do some personal travelling. i see a lot of pc's with windows and mac's being used everywhere. never saw anybody doing unix on the plane. and i frequently pack up a pc, stick it in a box, get to where i'm going, take it out, turn it on, don't miss a beat. NOT being tied to the network makes it the portable computer and i don't lose any of the environment i'm used to. can you do this with unix box? can't with the way ours are setup but not a unix guru, just a key tapper. also, what about the remote tools for the pc? they're not very expensive although i don't have any first-hand experience with them. perhaps someone else can comment on their performance. then again, tools for pc are cheap. got a set for home pc. don't gotta go to work. and can upload/download files no prob over the internet. hooking up drag and drop from home pc to work pc now, it'll be even easier to communicate. -------------------------------------------------------------- rk "there's nothing like real data to screw up a great theory" - me (modified from original, slightly more colorful version) --------------------------------------------------------------Article: 9301
Thanks for the reply, Bob. What I'm looking for is to use object oriented portion of C++ (classes, inheritance, etc.) to build my models using PLI for verilog or FLI for VHDL. The existing standard PLI only supports procedure oriented C programming and C++ classes are not allowed. Furthermore, I am also hopping to be able to present simulation output using some kind of GUI. I know there are tools like i-Logix' HDL express, etc doing things like that. But I need to use standard simulators such as Modelsim, VSS, etc as my primary simulation environment and I need to have more control to the code and using standard interfaces like PLI/FLI. Although I am hardware designer, I started using Java as my GP programming language early last year and have been hooked since. Jian Zhang Petaluma, CA P.S. I am not doing hardware/software co-simulations. I am an ASIC RTL designer and need to verify only the RTL designs I did. > > I don't know of anyone currently using Java (sort of a back burner > project > of mine :), but people DO use C++ for PLI/FLI code (also for system test > and verification). > > There will be presenting a paper one method of "hooking up" tests > written in C (or C++) to Verilog at the upcoming IVC conference in > Santa Clara. > > There's also a commercial product available that pretty much does the > same thing (VCPU from SimTech). > > Is this what you're talking about? > > --Bob > > -- > Bob Beckwith > To reply, remove NOSPAM. from the email address above.Article: 9302
The xilinx 4000EX data sheet says that there is an input hysteresis of "about 300mV", however there is no guaranteed figure in the DC characteristics section. Does anyone know what is guaranteed, and if it applies to all input pins in CMOS or TTL modes? I have a slow slewing signal (~5V/30ns) that I wish to use as a clock. I don't wish to incur the cost of an external buffer. TIA, Allan.Article: 9303
I am looking for an OTP FPGA at the size of about 30,000 to 50,000 gates. Anybody knows which manufacturer provide the lowest cost parts for this sort of FPGA? I've got a rough quote for Actel's A32300DX series. It costs about $500. Considering it is an OTP part, it doesn't seem very attractive to me in terms of pricing. C. FungArticle: 9304
cf: : I am looking for an OTP FPGA at the size of about 30,000 to 50,000 gates. : Anybody knows which manufacturer provide the lowest cost parts for this : sort of FPGA? I've got a rough quote for Actel's A32300DX series. It costs : about $500. Considering it is an OTP part, it doesn't seem very attractive : to : me in terms of pricing. rk: hmmm ... haven't bought any a32300dx's yet. but here's a comment or two, perhaps it could help. one, you have to define 'gates.' the a32300dx family's gate count may be higher than some other manufacturers (without trying to start a gate count war). of course, if you use their new 42mx family, you can double your gate count for free! (to be fair, they quote it two ways, as pld gates and gate array gates. but it's a good trick to double my design efficiency in gates per day). don't know the pricing for the 42mx stuff, but it is a smaller die, and presumably cheaper than the dx. 42mx = 0.45 um. dx = 0.6 um. might want to try some different packaging options. sometimes that helps. an option for some of the actel parts is the -F version. you can guess what the 'f' stands for. but, if you are not pushing the technology for speed, and these parts are say 30-40% slower than standard speed, you may be able to get a bargain. also, you can look at the pasic3 series from quicklogic. these parts are 0.35 um, 4 layer metal. should be cheap. and, relative to another thread/brewing flame war, the die is small. i opened up some ql3025's (quoted at 25kgates, iirc) and the die is tiny, no doubt about that. if you'll be purchasing in volume, you may want to prototype with otp fpga. and then plunk down a quick turn asic such as the cx2041 from chipx. i know that you can get the pinouts to match up as well as some of the footprints. of course, gate count is in the idea of the beholder. hope this helps, good luck! -------------------------------------------------------------- rk "there's nothing like real data to screw up a great theory" - me (modified from original, slightly more colorful version) --------------------------------------------------------------Article: 9305
Jian_Zhang wrote:
>
> Thanks for the reply, Bob.
>
You're welcome!
> What I'm looking for is to use object oriented portion of C++ (classes,
> inheritance, etc.) to build my models using PLI for verilog or FLI for
> VHDL. The existing standard PLI only supports procedure oriented C
> programming and C++ classes are not allowed.
I know of people that have done this with verilog. You may have to
resort
to some fairly sophisticated hackery (an oxymoron?) in order to make it
work, but it definitely can be done. This biggest problem is static
object
constructors (and there may be some issues if you try to mix C stdio
with
C++ streams) but you can certainly write C++ classes to encapulate the
PLI functionality in a straightforward manner. What you can NOT do with
the PLI is modify the netlist topology (you could write even more
classes
that allow you do implement "hardware" widgets, but where they connect
in with your HDL code would presumably be "fixed")
>
> Furthermore, I am also hopping to be able to present simulation output
> using some kind of GUI. I know there are tools like i-Logix' HDL
> express, etc doing things like that. But I need to use standard
> simulators such as Modelsim, VSS, etc as my primary simulation
> environment and I need to have more control to the code and using
> standard interfaces like PLI/FLI.
>
You should look into the MVC (model/view/controller) methodology that
evolved with Smalltalk. In a nutshell, each entity (i.e. object) has
3 parts (listed above). The model - which is the actual object, a view
which is its visual representation and a controller (a way of making the
model do things and also vary model parameters). You could adopt this
and implement in parts. Once you've got all three going, you'll have
your
GUI.
> Although I am hardware designer, I started using Java as my GP
> programming language early last year and have been hooked since.
>
> Jian Zhang
> Petaluma, CA
>
> P.S. I am not doing hardware/software co-simulations. I am an ASIC RTL
> designer and need to verify only the RTL designs I did.
>
I am a verification engineer, and I won't speak for co-simulation at
this
point, but I'd MUCH rather write my tests in C (or C++) than in an HDL.
I
know there are those who would disagree with me (and some rather
strongly,
I'm sure), but this is my preference. They're welcome to theirs.
--
Bob Beckwith
To reply, remove NOSPAM. from the email address above.
Article: 9306
Hello,
I read in the Xilinx documentation that the 4000XL series allows
mapping an internal signal to an available BUFGLS.
(page 4-43 9/96 databook). I have not been able to do this, however.
In Leonardo, I have tried assigning a BUFGLS to the signal and
even tried mapping a component in. However, when it comes time for
the Xilinx tools to place and route, I get the following error:
--------------------------------------------------
Processing global clock buffers...
ERROR:baste:263 - The LOC constraint "P175" (a IOB location) is not valid for
IPAD symbol `ix2282' (pad signal=dsptx_wr), which is being mapped to the
following site types:
CLKIOB
--------------------------------------------------
Can anyone lend some advice? Thanks!
Cheers,
Jake
--
janovetz@uiuc.edu | Once you have flown, you will walk the earth with
University of Illinois | your eyes turned skyward, for there you have been,
| there you long to return. -- da Vinci
PP-ASEL | http://www.ews.uiuc.edu/~janovetz/index.html
Article: 9307
janovetz@ews.uiuc.edu (Jacob W Janovetz) writes:
> I read in the Xilinx documentation that the 4000XL series allows
>mapping an internal signal to an available BUFGLS.
>(page 4-43 9/96 databook). I have not been able to do this, however.
>In Leonardo, I have tried assigning a BUFGLS to the signal and
>even tried mapping a component in. However, when it comes time for
>the Xilinx tools to place and route, I get the following error:
>--------------------------------------------------
>Processing global clock buffers...
>ERROR:baste:263 - The LOC constraint "P175" (a IOB location) is not valid for
> IPAD symbol `ix2282' (pad signal=dsptx_wr), which is being mapped to the
> following site types:
> CLKIOB
>--------------------------------------------------
> Can anyone lend some advice? Thanks!
Well, I think I just answered my own question. I had to map the
pin to an IBUF first, then to a BUFGLS. This helped....
Cheers,
Jake
--
janovetz@uiuc.edu | Once you have flown, you will walk the earth with
University of Illinois | your eyes turned skyward, for there you have been,
| there you long to return. -- da Vinci
PP-ASEL | http://www.ews.uiuc.edu/~janovetz/index.html
Article: 9308In comp.arch.fpga Thomas Pornin <pornin@news.ens.fr> wrote: : In article <6dn451$b0c$1@nntp.Stanford.EDU>, : Mark Willey <willey@etla.ml.org> wrote: :>The tools I refer to are: : [...several useful tools...] :>One of the most important tools I use is X windows. It does me absolutely :>no good to have an NT box or server across campus. However, if I have a :>UNIX box, I just log in remotely and run my CAD tool while displaying the :>application GUI on my local X server. : Actually, all (or almost all) the tools you describe are available for NT : for free; the "export DISPLAY" one exists also and is called NTrig (or : something like that)(and it costs much !)(there is also something free I : saw once, but it was not very efficient in terms of network consumption). The tools that are available for NT form that list (and *no*, not all of them are) are often quite buggy. Perl under NT, for example has a lot of problems. I used many of these tools under NT last summer and was thoroughly disgusted. : So you can get many posix tools and make your NT look like a Unix. The You can get them, but they are often very buggy - and don't fit together as seamlessly as they do under UNIX. POSIX under NT always has been a sick joke. : point is: if you want something like Unix, why do you use NT ? Get Linux, : it is faster and cheaper, and at least as reliable as NT. Agreed. : Anyway, I once met a guy who was doing some development on FPGAs using : only NT -- and he was quite happy and efficient, more than he could be : with a Unix shell. I think I represent the opposite: I am painfully slow Um... yes, more than with a UNIX shell... but who just uses a UNIX shell? People use the GUIs that run on UNIX. : To sum up: some people prefer NT, other prefer Unix. So it is a pity that : tools are not as available under Unix as they are under NT. Well, the tools are coming to NT. They really aren't there as of yet. The shame is that they are missing the easier target of Linux. MarkArticle: 9309
In comp.arch.fpga rk <stellare@erols.com.NOSPAM> wrote: : peter: : : :There are lots of stupid things in NT but none of them make it in any : : ;way unsuitable for EDA. : wen: : : Oh yes. For one, it prevents me from using the tools if I am not sitting : : in front of of the machine. : rk: : solution: sit in front of the machine. Uh... who wants to do that? I'd rather do my work from wherever I want to be, such as on the train, at home, or out in the garden. Not in the office breathing recycled air and sitting in uncomfortable ergo chairs. : i think nt (and probably linux) is winner here over unix since with nt, you : can simply fold up the machine, put it in your backpack, hop on the plane, : wait for the announcement that you can pull out your 'puter, and get back : to work. i travel a modest amount on business, 6-8 times per year, and i : do some personal travelling. i see a lot of pc's with windows and mac's : being used everywhere. never saw anybody doing unix on the plane. and i : frequently pack up a pc, stick it in a box, get to where i'm going, take it : out, turn it on, don't miss a beat. NOT being tied to the network makes it : the portable computer and i don't lose any of the environment i'm used to. : can you do this with unix box? can't with the way ours are setup but not a : unix guru, just a key tapper. I think the thing that you're missing here is that Linux *is* a UNIX. NetBSD *is* a UNIX. And it runs on more hardware than MS Windows and Mac combined, including PCs. So any claims about the portability of the PC is even more true for the UNIX world. Case in point, you can disconnect from the network on a Mac or UNIX, move to a new location, and reconnect to the network w/o rebooting. Last time I checked, Windows could not do that. : also, what about the remote tools for the pc? they're not very expensive : although i don't have any first-hand experience with them. perhaps someone : else can comment on their performance. They... suck. ;-) : then again, tools for pc are cheap. got a set for home pc. don't gotta go : to work. and can upload/download files no prob over the internet. hooking : up drag and drop from home pc to work : pc now, it'll be even easier to communicate. Who wants to upload/download files? Ugh. Welcome to the 80's, man... ;-) Just NFS mount the work drive or telnet in and run the job on the remote machine. The key point of the message: PC != Windows. Linux/*BSD = UNIX. UNIX != workstation. UNIX runs on a PC, too. Only faster/better. :) Mark "Not a fanatic, really! ;-)" WilleyArticle: 9310
Hello, You may want to post your announcement at the TechExpo Calendar of Science & Technology events at http://www.techexpo.com (select TechCALENDAR). Over 1,000 conferences are currently posted, updated daily, the site is visited daily by 1,000 to 3,000 scientists, engineers, and technical managers. Posting is free to qualified organizations. Regards, N. Gat, Ph.D. (I apologise if this message reached you twice, or if I miised your prior posting at TechExpo)Article: 9311
I have had a couple of email responses from people who are interested in schematic version control. One is working on his own custom solution as a background task. I thank those that responded. I encourage others to post follow-up messages to this news group, however, so the venders can see there is a demand. I also had a phone call from the PVCS people. They claim that PVCS can automatically check in and out multiple dirs, and that PVCS will automatically recognize extra files being checked in, and will properly include them in the version. So one main function I was looking for is already there. I won't need .zip or TAR. PVCS also claims they can provide the rest of my whish list, but at a cost of me paying for a PVCS expert to come to my company for two days to write the necessary scripts. I'm looking into it. That's it for this update. Dave Decker Diablo Research Co. LLC Please use only one 'h' in mush. I'm trying to reduce the spam. "Animals . . . are not brethren they are not underlings; they are other nations, caught with ourselves in the net of life and time, fellow prisoners of the splendor and travail of the earth." Henry Beston - The Outermost HouseArticle: 9312
I should have phrased it differently. NT is easy to crash if there are hardware problems, and of course hardware problems can be caused by software. Let's say you have a normal user program. It is very hard for that program to crash NT in the traditional way: by crapping into memory which doesn't belong to it - although reportedly this is just about possible, as someone showed by writing a DOS app which writes 0x00 into every single address in its 1MB address space. But let's say that program sends an invalid command to the SCSI controller, and messes it up. Then NT dies with it, too. And of course a DOS or win16 app can just do cli loop: jmp loop and you have a dead machine. A win32 app cannot do this however. Peter. Return address is invalid to help stop junk mail. E-mail replies to zX80@digiYserve.com but remove the X and the Y.Article: 9313
>CVS >diff >gcc >emacs/vi >make >perl >shell scripts >screen >grep It's funny - I use none of the above, except make. VI is for masochists :) And I have designed some 300 different board-level products, a dozen FPGAs, a few ASICs. It just shows how far apart the "workstation" and "PC" worlds can be, even today. Peter. Return address is invalid to help stop junk mail. E-mail replies to zX80@digiYserve.com but remove the X and the Y.Article: 9314
thomas: : : Anyway, I once met a guy who was doing some development on FPGAs using : : only NT -- and he was quite happy and efficient, more than he could be : : with a Unix shell. I think I represent the opposite: I am painfully slow mark: : Um... yes, more than with a UNIX shell... but who just uses a UNIX shell? : People use the GUIs that run on UNIX. rk: a lot of people don't use the gui and are command line oriented. a big plus when you want to operate, say, via telnet, and don't have access to a high speed link to support a gui system. and, for hdl-based eda, you want all of the cpu cycles doing real work, not drawing pictures.Article: 9315
> rk: > solution: sit in front of the machine. > > i think nt (and probably linux) is winner here over unix since with nt, you > can simply fold up the machine, put it in your backpack, hop on the plane, > wait for the announcement that you can pull out your 'puter, and get back > to work. i travel a modest amount on business, 6-8 times per year, and i > do some personal travelling. i see a lot of pc's with windows and mac's > being used everywhere. never saw anybody doing unix on the plane. and i > frequently pack up a pc, stick it in a box, get to where i'm going, take it > out, turn it on, don't miss a beat. NOT being tied to the network makes it > the portable computer and i don't lose any of the environment i'm used to. > can you do this with unix box? can't with the way ours are setup but not a > unix guru, just a key tapper. > There is a significant advantage of Linux over Windows concerning portables. I was very surprised when I touched the processor inside my desktop when it was running Linux. It was COLD. So was the power regulator. Under Windows95 the processor is always hot, even when OSR2 System Monitor shows 1% CPU utilisation for an hour. Same about NT. Of course, when CPU-intensive application runs under Linux, the processor warms up. But for typing text on the portable Linux should give much longer battery life. Misha.Article: 9316
peter: : : : :There are lots of stupid things in NT but none of them make it in any : : : ;way unsuitable for EDA. wen: : : : Oh yes. For one, it prevents me from using the tools if I am not sitting : : : in front of of the machine. rk: : : solution: sit in front of the machine. mark: : Uh... who wants to do that? I'd rather do my work from wherever I want to : be, such as on the train, at home, or out in the garden. Not in the : office breathing recycled air and sitting in uncomfortable ergo chairs. rk: perhaps you should have read the next line of my post. can you do your eda on a train w/ unix? sorry, haven't seen that, my experience (somewhat limited) usually had unix on fairly big machines. and i haven't seen the standard eda applications running on mac or pc hardware running unix os's. is that available? rk eloquently said: : : i think nt (and probably linux) is winner here over unix since with nt, you : : can simply fold up the machine, put it in your backpack, hop on the plane, : : wait for the announcement that you can pull out your 'puter, and get back : : to work. i travel a modest amount on business, 6-8 times per year, and i : : do some personal travelling. i see a lot of pc's with windows and mac's : : being used everywhere. never saw anybody doing unix on the plane. and i : : frequently pack up a pc, stick it in a box, get to where i'm going, take it : : out, turn it on, don't miss a beat. NOT being tied to the network makes it : : the portable computer and i don't lose any of the environment i'm used to. : : can you do this with unix box? can't with the way ours are setup but not a : : unix guru, just a key tapper. mark: : I think the thing that you're missing here is that Linux *is* a UNIX. rk: if Linux *is* a UNIX, can i just go down to the local synopsys store, order me up a copy of UNIX synopsys, drop in the cd onto my linux/x86 portable machine, and start running? mark: : NetBSD *is* a UNIX. And it runs on more hardware than MS Windows and Mac : combined, including PCs. So any claims about the portability of the PC is : even more true for the UNIX world. rk: hmmm ... just a little old engineer here, but i seem to remember, let me know if it's not true, that s/w needs to be written or tweaked for different versions of unix. can i take code from an hp box and run it on a sun box, or a dec box, or a ibm box, or an cray box, or a silicon graphics box (haven't followed these two too close since merger) or a xenix box? anyways, that's s/w portability. mark: : Case in point, you can disconnect from : the network on a Mac or UNIX, move to a new location, and reconnect to the : network w/o rebooting. Last time I checked, Windows could not do that. rk: while i like all three environments, unix, mac, and win, for particular uses, being able to change physical location, disconnect and reconnect to a network without rebooting, doesn't seem to be a very important criteria for determining whether an operating system wins the $ if you have to choose the three. in fact, for a machine that is being moved or is portable, i would want to turn it off. in fact, i would insist on it. i think, rather than the cost of a reboot, which just takes a minute or so, i would worry more about those high-speed rotating pie tins covered with rust that we store our information on. reloading that is a pain. and losing info is worse. rk: : : also, what about the remote tools for the pc? they're not very expensive : : although i don't have any first-hand experience with them. perhaps someone : : else can comment on their performance. mark: : They... suck. ;-) rk: haven't tried them myself, so i'll take your word for it. but, i do have experience with the laptops and their fine, easy to carry around. and with the move to win '95/nt, we've been writing remote capability into our software, can operate it from anywhere there's an internet connection. it's GREAT! ;) <you're supposed to chime in, 'less filling.'> rk: : : then again, tools for pc are cheap. got a set for home pc. don't gotta go : : to work. and can upload/download files no prob over the internet. hooking : : up drag and drop from home pc to work : : pc now, it'll be even easier to communicate. mark: : Who wants to upload/download files? Ugh. Welcome to the 80's, man... ;-) : Just NFS mount the work drive or telnet in and run the job on the remote : machine. rk: me. don't want to tie up the phone line running programs. just let the machine whirrr away and be happy. and for a lot of graphical-based work, doing that over a 28.8 kbps modem is sooooooooooooooooooooooooooo sloooooooooooooooooooooooooooooooow. yuch. welcome to the '90s, you can take your computing cycles with you. anywhere. i'm thinking the model of the processor in a room where remote users use a terminal to get access. what is that, 1962? my memory fades ;) telnet is nicer for some applications, haven't seen it for windows. mark: : The key point of the message: PC != Windows. Linux/*BSD = UNIX. UNIX != : workstation. UNIX runs on a PC, too. Only faster/better. :) rk: if this logic is true, then i can go down to the synopsys store, on the way back stop off at the linux store, and then load linux and synopsys on the pc hardware and run, no muss, no fuss. correct? and all of the other unix programs, right? mark: : Mark "Not a fanatic, really! ;-)" Willey rk: really? -------------------------------------------------------------- rk "there's nothing like real data to screw up a great theory" - me (modified from original, slightly more colorful version) --------------------------------------------------------------Article: 9317
Peter wrote: > > >CVS > >diff > >gcc > >emacs/vi > >make > >perl > >shell scripts > >screen > >grep > > It's funny - I use none of the above, except make. No revision control ? No batch mode ? Most of the cpu and time consuming jobs here are done by batch (shell scripts, etc) overnight, e.g. Synthesis, ATPG, Simulations, Analysis, etc. Don't tell me that you turn your PC off overnight (perhaps to get your NT free off memory leaks ;)) and thus waste a lot of time ... > VI is for masochists :) Or for those who need a small and fast editor and don't want to lose time by moving their hands to the mouse and back during touch typing, e.g. to copy just one line :). > It just shows how far apart the "workstation" and "PC" worlds can be, > even today. Yes, I agree here :)). Stephan.Article: 9318
INTERNATIONAL SYMPOSIUM ON PHYSICAL DESIGN 1998
Embassy Suites, Monterey, CA
April 6-8, 1998
http://www.ee.iastate.edu/~ispd98
***************************************************
** HOTEL RESERVATION DEADLINE: Mar 9 **
** Early registration deadline: Mar 10 **
***************************************************
ADVANCE PROGRAM
The International Symposium on Physical Design provides a high-quality forum
for the exchange of ideas and results in critical areas related to the physical
design of VLSI systems. This meeting evolved from the ACM/SIGDA Physical
Design Workshops held during the years 1987-1996. The first Symposium in 1997
was highly successful and drew such a large number of attendees that
registration had to be closed a month early. The scope of this symposium
includes all aspects of physical design, from interactions with behavior- and
logic-level synthesis, to back-end performance analysis and verification.
MONDAY, April 6
0915-0930 Welcome
M. Sarrafzadeh, General Chair (Northwestern)
D. F. Wong, Program Chair (UT-Austin)
0930-1030 Keynote Address
"Perspectives on Systems at 1 GHz and beyond"
Dave LaPotin (IBM Austin Research Laboratory)
1030-1100 BREAK
1100-1230 Session 1: Floorplanning and Placement
Chairs: C.K. Cheng (UCSD) and Jochen Jess (Eindhoven)
"On Wirelength Estimations for Row-Based Placement"
A.B. Kahng, S. Mantik, I.L. Markov, A. Zelikovsky (UCLA)
"Performance-Driven Soft-Macro Clustering and Placement by
Preserving HDL Design Hierarchy"
H.-P. Su, A.C.-.H. Wu, Y.-L. Lin (Tsing Hua)
"Nostradamus: A Floorplanner of Uncertain Design"
K. Bazargan, S. Kim, M. Sarrafzadeh (Northwestern)
1230-1430 Lunch (Pinot Noir Room)
Special Address: "Impact of Web Technologies on EDA System
Architectures"
A. R. Newton (UCB)
1430-1600 Tutorial: "Timing Metrics for Physical Design of Deep
Submicron Technologies "
Presenter: L. Pileggi (CMU)
Panelists: J. Cong (UCLA)
S. Otto (Intel)
A. Yang (Washington)
1600-1630 BREAK
1630-1730 Special Address: "Moore's Law and Physical Design of ICs"
W. Maly (CMU)
1730-1830 Session 2: Interconnect Optimization
Chairs: M. Alexander (Washington State) and
Y.-L. Lin (Tsing Hua)
"Greedy Wire-Sizing is Linear Time"
C.C.N. Chu, D.F. Wong (UT-Austin)
"An Efficient Technique for Device and Interconnect
Optimization in Deep Submicron Designs"
J. Cong, L. He (UCLA)
1900-2100 Dinner (Pinot Noir Room)
Special Address:
"Consorting with the Consortia: Cooperative Research
For Fun and Profit"
W. H. Joyner (SRC)
TUESDAY, April 7
0830-0930 Session 3: Layout Methodologies for RF Circuits
Chairs: M. Pedram (USC) and W. Dai (UCSC)
"Device-Level Early Floorplanning Algorithms for RF Circuits"
M. Aktuna, R.A. Rutenbar, L. R. Carley (CMU)
"A Layout Approach to Monolithic Microwave IC"
A. Nagao, T. Kambe (SHARP); I. Shirakawa (Osaka)
0930-1030 Session 4: Framework and Benchmarks
Chairs: D. Hill (Synopsys) and L. Jones (Motorola)
"CHDStd--Application Support for Reusable Hierarchical
Interconnect Timing Views"
S. Grout, G. Ledenbach, R. G. Bushroe, P. Fisher,
A. Chokhavtia (Sematech); D. Cottrell, D. Mallis
(Silicon Integration Initiative); S. DasGupta,
J. Morrell (IBM)
"The ISPD Circuit Benchmark Suite"
C.J. Alpert (IBM)
1030-1100 BREAK
1100-1230 Panel: "Given that SEMATECH is levelling the semiconductor
technology playing field, will corporate CAD (in
particular, PD) tools be an enabler/
differentiator of technology in the future?"
Organizer: S. DasGupta (IBM)
Panelists:
B. Beers, IBM
M. Khaira, Intel
R. Abrishami, Fujitsu Microelectronics
J. Hutt, Synopsys
L. Sheffer, Cadence
D. Guiou, Mentor Graphics
V. Kulkarni, Avant!
1230-1330 Lunch (Atrium Court)
1330-1430 Session 5: "PD for Manufacturability"
Chairs: M. Weisel (Intel), R. Rutenbar (CMU)
"Critical Area Computation--A New Approach"
E. Papadopoulou (IBM), D.T. Lee (Northwestern)
"Filling and Slotting: Analysis and Algorithms"
G. Robins, A. Singh (Virginia); H. Wang (UCLA);
A. Zelikovsky (Virginia)
1430-1530 Special Address: "Global Wires: Harmful?"
R. Otten (Delft)
1530-1600 BREAK
1600-1645 Session 6: Poster Presentations
Chairs: S. Sapatnekar (Minnesota) and G. Robins (Virginia)
"Partioning Using Second-Order Information and Stochastic-Gain
Functions"
S. Dutt (UIC), H. Theny (Intel)
"A Parallel Algorithm for Zero Skew Clock Tree Routing"
Z. Xing, P. Banerjee (Northwestern)
"On Convex Formulation of the Floorplan Area Minimization
Problem"
T. Chen, M. Fan (Georgia Tech)
"A Pattern Matching Algorithm for Verification and Analysis
of Very Large IC Layouts"
M. Niewczas, W. Maly, A. Strojwas (CMU)
"LIBRA--A Library-Independent Framework for Post-Layout
Performance Optimization"
R. Huang (UCSB), Y. Wang (Avant!), K.-T. Cheng (UCSB)
"Estimation of Maximum Current Envelope for Power Bus
Analysis and Design"
S. Bobba, I.N. Hajj (UIUC)
"New Efficient Algorithms for Computing Effective Capacitance"
S. Muddu (SGI)
"Calculation of Ramp Response of Lossy Transmission Lines Using
Two-Port Network Functions"
P. Heydari, M. Pedram (USC)
"Switch-Matrix Architecture and Routing for FPDs"
G.-M. Wu, Y.-W. Chang (Chiao-Tung)
1645-1745 Poster Session
1900-2200 Banquet (Rancho Canada Golf Club)
WEDNESDAY, April 8
0830-1000 Session 7: Efficient Representation in Placement
Chairs: R. Otten (Delft) and C. Sechen (Washington)
"Sequence-Pair Based Placement Method for Hard/Soft/Pre-placed
Modules"
H. Murata, E.S. Kuh (UCB)
"Rectilinear Block Placement using Permutation-Pair"
J. Xu, C.K. Cheng (UCSD)
"Topology Constrained Rectilinear Block Packing for Layout
Reuse"
M. Kang, W. Dai (UCSC)
1000-1030 BREAK
1030-1200 Panel: "Process development and its impact on Physical Design"
Moderator: N. Sherwani (Intel)
Panel members: J. Cong (UCLA)
D. Lapotin (IBM)
J. Rey (Cadence)
1230-1400 Lunch (Pinot Noir Room)
1400-1530 Tutorial: "Why Clustering is the Key to Partitioning"
Presenter: A. Kahng (UCLA)
Panelists: C. Alpert (IBM)
G. Janac (Cadence)
J. Lillis (UIC)
1530-1700 Session 8: Routing Algorithms
Chairs: J. Cong (UCLA) and J. Fishburn (Lucent)
"Chip-Level Area Routing"
L.-C. Liu, H.-P. Tseng, C. Sechen (Washington)
"Routing Tree Topology Construction to Meet Interconnect
Timing Constraints"
H. Hou (Iowa State), S. Sapatnekar (Minnesota)
"Analysis, Reduction and Avoidance of Crosstalk on VLSI Chips"
T. Stoehr, M. Alt (IBM); A. Hetzel (Bonn), J. Koehl (IBM)
1700 Symposium Closes
%%==========================================================================%%
%% Symposium Organization %%
%%==========================================================================%%
General Chair
M. Sarrafzadeh (Northwestern)
Past Chair
A. B. Kahng (UCLA)
Steering Committee
J. P. Cohoon (Virginia)
S. DasGupta (IBM)
M. Marek-Sadowska (UC Santa Barbara)
B. Preas (Xerox PARC)
E. Yoffa (IBM)
Technical Program Chair
D. F. Wong (UT-Austin)
Technical Program Committee
M. J. Alexander (Washington State) C. K. Cheng (UC San Diego)
J. Cong (UCLA) W. W.-M Dai (UC Santa Cruz)
J. Fishburn (Lucent) D. Hill (Synopsys)
J. A. G. Jess (Eindhoven) L. Jones (Motorola)
S. M. Kang (UIUC) Y.-L. Lin (Tsing Hua)
M. Pedram (USC) R. Rutenbar (CMU)
C. Sechen (Washington) M. Wiesel (Intel)
T. Yoshimura (NEC)
Publication Chair
D. Hill (Synopsys)
Panel Chair
N. Sherwani (Intel)
Local Arrangements Chair
R.-S. Tsay (Axis Systems)
Publicity Chair
S. Sapatnekar (Minnesota)
Treasurer
S. Souvannavong
Sponsors
ACM Special Interest Group on Design Automation
in cooperation with
IEEE Circuits and Systems Society and IEEE Computer Society
Additional Support From:
Avant! Corporation
Ambit Design Systems
%%==========================================================================%%
%% Hotel Accommodations and Travel %%
%%==========================================================================%%
ISPD-98 is being held at the Embassy Suites Monterey Bay in Monterey,
California, located on the beautiful Monterey Peninsula, two blocks from
the beach, at the intersection of Canyon Del Rey and Del Monte Boulevard.
The address is
Embassy Suites Monterey Bay Hotel & Conference Center
1441 Canyon Del Rey
Seaside, California 93955
Tel: (408) 393 1115
Fax: (408) 393 1113
For hotel reservations, call (408) 393 1115 or (800)362 2779.
A block of rooms is being held for the nights of Sunday through Wednesday
(April 5 through April 8). Room rates are $125 per night for a single room
and $145 per night for a double room. Any individual cancellations within 72
hours from the date of arrival will be billed for (1) night's stay, plus tax.
+---------------------------------------------------------+
| Please make room reservations directly with the hotel |
| at either 1-408-393-1115 or 1-800-362-2779, mentioning |
| ``ISPD'' to get the special rate. |
+---------------------------------------------------------+
The number of rooms available at this rate is limited, and are only being
held through March 9. Early room reservation is highly recommended.
%%==========================================================================%%
%% ISPD-98 Advance Registration Form %%
%%==========================================================================%%
Name: _______________________________________________________
Company/University: _________________________________________
Responsibility/Title: _______________________________________
Address: ____________________________________________________
City: _______________________________ State: ________________
Country: ______________________ Postal Code: ________________
Phone: ________________________ Fax: ________________________
Email: ______________________________________________________
Food Choices:
[ ] Vegetarian meals
[ ] Non-vegetarian meals
[ ] Either one is fine
Advance Late
(Through March 10) (After March 10)
ACM/IEEE Members [ ] $350 [ ] $425
Non-Members [ ] $425 [ ] $500
Full-Time Students [ ] $175 [ ] $225
Student ID is required if registering as a student.
ACM or IEEE Member No. _____________________________
Registration fee includes meals and Banquet. The Banquet will be held at
the Rancho Can~ada Golf Club and ISPD will provide a bus service to the site.
Payment may be submitted via personal or company check in US funds only and
drawn on a US bank, made payable to ``ACM/International Symposium on
Physical Design''. Payment may also be made with credit card (circle):
Mastercard Visa American Express
Credit Card # _______________________________________________
Expiration Date: ______________ Total Payment: ______________
Name as it appears on credit card: __________________________
Signature: ___________________________ Date: ________________
Please mail or FAX (credit card only) your completed registration form to:
ISPD-98 Symposium Registration
Sally Souvannavong, Treasurer
P.O. Box 395
Pullman, WA 99163-0395
FAX: 1-509-332-6118
Email registration will not be accepted. Cancellations must be in writing
and must be received by March 24, 1998. Questions concerning symposium
registration should be directed to Sally Souvannavong at 1-509-334-3162,
Email: ispd98@eecs.wsu.edu.
Article: 9319Jian_Zhang <jzhang@optilink.dsccc.com> writes: > Does anyone know if there exist tools or even it's possible to use Java > as verilog PLI? How about using C++ classes in the code for PLI. I > need them to model some complex telecomm devices and it's very AWKWARD > to use C and downright impossible to use vhdl/verilog. Basically you can use any language as long as it support C calling convention. E.g. in C++ you have to declare the PLI interface routines as extern "C" (I had to make veriuser.h C++ aware), but you can still use classes etc. in the bulk of your code. This is quite a few years ago (I know use the method described below) and I think I had to play some tricks to make sure the constructors was called since the main program is not written in C++. What we did where I work was to write a small PLI routine with a socket interface. We have the different pattern generators and verifiers running as separate UNIX (or they could potentially run on a PC using INET style sockets) processes and communicating with the Verilog model using sockets. Another advantage of this is that you don't have to link your Verilog simulator each time you change your test, you only compile and link the smaller C++ pattern generators. The disadvantage is that the socket communication could become an bottleneck, but we have limited data sent over the socket and I don't think this is a problem. If your company has an excessive number of simulator licenses and computers you could use this method to do parallel simulation as well. Petter -- ________________________________________________________________________ Petter Gustad 8'h2B | (~8'h2B) - Hamlet http://home.sol.no/~pegu Remove the obvious from my reply address to send mail -- sed s/nospam//gArticle: 9320
In a previous article "rk" <stellare@erols.com.NOSPAM> writes: : ;peter: :: :There are lots of stupid things in NT but none of them make it in any ;: ;way unsuitable for EDA. : ;wen: :: Oh yes. For one, it prevents me from using the tools if I am not sitting ;: in front of of the machine. : ;rk: :solution: sit in front of the machine. Sorry, I can't do that. The machine I sometimes need to use would have so many processors, disks, and memory and powersupply that it would take a forklift to carry. I also routinely use manage jobs that runs on a farm of workstations and I need to do it remotely and service them at odd hours of the day. :also, what about the remote tools for the pc? they're not very expensive ;although i don't have any first-hand experience with them. perhaps someone :else can comment on their performance. They don't work very well, and when I am running jobs on a farm of 20 work stations, I don't want to have 20 virtual consoles on my screen where I have to locate and click menu buttons individually. The problem is not with NT, the problem is with how Microsoft is having an effect on how EDA programs were written. There is absolutely no reason why EDA vendors couldn't compile their program on NT but link them with X-window libraries so that the NT machine would be useful as a remote compute engine within a network of unix workstations. NT still has a long way to go to match Unix for networking and user interface. By making it so that engineers either has to choose between inexpensive EDA software and versatile and mature networked user environment, Microsoft is causing us a lot of pain. I am happy to take NT if Microsoft would release X-window library to the EDA vendors and encourage them them to build programs the way I described. But Microsoft won't do that because they don't care about you. They care about market share. They are a monopoly.Article: 9321
R. Mark Gogolewski wrote: > > In article <34FEE93E.9EC4786F@ix.netcom.com>, Rick Kwan > <rkwan@ix.netcom.com> wrote: > > >Pricing. This is an interesting artificial, obsolete separator > >between UNIX and PC platforms. I am arguing here for equivalent > >application pricing for UNIX and NT systems. > > <snip> > > >From the software vendor's point of view, pricing should reflect > >development and support costs. If it is the same on both NT > >and UNIX, then it should be priced the same. If it is not, > >you are getting ripped off on one platform or the other. > > That is definitely the crux. As to "rk"'s points, there is certainly > a market out there to lower performance/cheaper software. That > makes the development/support costs much less. You should get what > you pay for. Agreed. And those apps didn't justify a traditional workstation. What has happened in the fast few years, however, is that the workstation (POSIX) APIs have come down to the PC in reasonably robust implementations. Getting all the pieces into a common design flow becomes much easier. > >(Oh, BTW, compilers and debuggers come free on Linux and > >FreeBSD, not to mention the OS. :-) > > Yeah, and that represents < 5% of development costs. I would > find it difficult to argue that any specific platform significantly > changes the development costs vis-a-vis compilers, debuggers, > etc... You are absolutely correct. Having made the point about costs, I just couldn't resist the term "free". (Note above smiley face.) I repent. > > Mark --Rick KwanArticle: 9322
"rk" <stellare@erols.com.NOSPAM> writes:
> rk:
> agreed. but today, the unix s/w is generally more expensive. and the pc
> s/w is generally cheaper. and, the high end tools tend to run on unix.
> and the low end tools tend to run on pc. but the goals are, as you say,
> the key, not the OS. but with the volume of win machines in general as
> compared to unix boxes, it's simple economics that the $5k package can make
> more $ than the $90k if they both can do the job. and here do the job
> means process in a reasonable amount of time with a reasonable amount of
> hassle. and, actually getting back to my last msg, the small to medium
> sized designs are not as sensitive to performance as the million gate asic.
>
Great theroy; and it applies quite well to word processors,
spreadsheets and the like.
But your theory assumes completely elastic demand.
Something like 10% of Unix boxes (my guess) are used to run
EDA tools. I will go out on a limb here: No matter how many quality
EDA tools are ported to the intel platform (Linux or Win/NT), and no
matter how low the price, you will see far less than 1/100 of 1% of
the total number of PC's that will ever run EDA software.
So, vendor XYZ figures they can sell 100 copies of their
combination simulator, synthesis and place & route tool on UNIX in one
year. They can change $100k for the package; so $10M total sales!
Now if they port to the PC, and hence cut the price to $10k
for the package, can they sell 1000 copies? Sorry, no. They may be
able to sell 200 copies.. and hence only do $2M in sales.
That is the problem.
--
/\ Michael McNamara <mac@surefirev.com>
/\// SureFire Verification Inc.
/\///\ <http://www.surefirev.com>
_\///\/ Formerly Silicon Sorcery
\//\/ Get my verilog emacs mode from
\/ <http://www.surefirev.com/verilog-mode.html>
Article: 9323In article <01bd4897$2b62d1c0$2985accf@homepc>, rk <stellare@erols.com.NOSPAM> wrote: >peter: >: :There are lots of stupid things in NT but none of them make it in any >: ;way unsuitable for EDA. > >wen: >: Oh yes. For one, it prevents me from using the tools if I am not sitting >: in front of of the machine. > >rk: >solution: sit in front of the machine. Not always possible. For example if you can't make it into work because of an ice storm, you can (if you're running Linux at home) establish a ppp connection with a system at work and set your display to your home system (assuming you're running some kind of Unix at work) and do a reasonable amount of work. > >i think nt (and probably linux) is winner here over unix since with nt, you >can simply fold up the machine, put it in your backpack, hop on the plane, >wait for the announcement that you can pull out your 'puter, and get back >to work. i travel a modest amount on business, 6-8 times per year, and i >do some personal travelling. i see a lot of pc's with windows and mac's >being used everywhere. never saw anybody doing unix on the plane. and i >frequently pack up a pc, stick it in a box, get to where i'm going, take it >out, turn it on, don't miss a beat. NOT being tied to the network makes it >the portable computer and i don't lose any of the environment i'm used to. >can you do this with unix box? can't with the way ours are setup but not a >unix guru, just a key tapper. Sure, not being tied to the network might make it easier to take you laptop on the plane and work on a design, but the need for this is only 2-3 times a year for most of us. The disadvantages of not being tied to a network greatly outweigh the so called advantages (especially if you have multiple people on a project). BTW: Linux does run very nicely on most laptops, so you could do the same if you wanted. As far as the environment issues you mention - you would have the same problem if your laptop running NT was connected to a network at the office. > >also, what about the remote tools for the pc? they're not very expensive >although i don't have any first-hand experience with them. perhaps someone >else can comment on their performance. > >then again, tools for pc are cheap. got a set for home pc. don't gotta go >to work. and can upload/download files no prob over the internet. hooking >up drag and drop from home pc to work >pc now, it'll be even easier to communicate. > Well, the tools for running remotely with Linux are free. You get XWindows which allows you to do this sort of thing. Now there is something called VNC (Virtual Network Computer) which allows PCs running NT, Suns running Solaris, and Linux boxen to display a desktop remotely - and VNC is free (GPLed). philArticle: 9324
Given the following code for a control register where a single bit has to have a seperate async reset : (target device = Xilinx 4ke) process ( strobe, reset, clear_bit0) begin if reset = '0' then cntl_reg <= DEFAULT_REG; elsif clear_bit0 = '1' then cntl_reg(0) <= DEFAULT_REG(0); elsif strobe'event and strobe = '1' then cntl_reg <= data; end if; end process; I assign reset to GSR so I expect clear_bit0 to drive the reset line of a dff on bit 0 and the reset of the register bit to only have the GSR reset. My synthesis tool (Leonardo) says that since I have nothing defined for the upper bits under the clear_bit0 condition it defaults to a preset which is not what I inteded. I have pulled out bit 0 into its own process and everythings happy, I just don't understand why this method doesn't work. Any ideas? Thanks, Allan Redenbaugh Raytheon Systems
Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
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