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
I used synplicity on a Xilinx 40150. The versions I used were: 5.0.6, 5.0.7 and 5.0.8. Before getting to know synplicity, I used Synopsys's FPGA Compilor on a Xilinx 4036. All works were done on SUN workstations. Here is a summary of problems that I have seen: 1. Crashing the WS from 5.0.6 to 5.0.7 when browsing the design under technology view. 5.0.8 seemed to fix the problem although there exist a Solaris patch to fix the a video card bug. 2. Running out of memory on 5.0.7. There is a memory leakage problem that eats up all 1Gbytes of RAM in our Ultra. 5.0.8 fixed this problem. 3. Doing mapping forever on 5.0.6. The design ran for 20 hours and still not done! 5.0.8 fixed this problem. 4. Misleading CLB usage count. In the beginning, I used the stats reported in the .srr file to estimate how big the device would be. Big mistake, as a lot of CLBs are used as route through and the final size is 20% bigger (roughly). 5. Timing estimation is way off. At one time, we have hundreds of timing violation from par. The par figure reported was like 19 nanosecong worst than that estimated by synplicity. Overall synplicity's estimate is always too optimistic. These are the worst problem that I have. There are many more. On the plus side, the compile speed is very high compare to the FPGA compiler. On the 4036 design, it was like 180 seconds vs 3.5 hours. As well, the support is excellent. Every call is answered very quickly with work around that works. The tool has improved a lot since the beginning but is still being developed. At the end, we decided to stay with synplicity inspite of its many problems. Good luck, Hing-Fai Nortel Networks Dept 1D13, Mail Stop 162 Phone: 613-765-2097 P.O. Box 3511, Station C FAX : 613-763-1327 Ottawa, Canada, K1Y 4H7 Email: crm182c@nortelnetworks.comArticle: 14201
*************** DesignPRO Inc. ****************** Does your company have programs that are being delayed due to a shortage of skilled ASIC/FPGA Design Engineers? Maybe we can help! DesignPRO is an engineering design organization, located in Ottawa, Canada, specializing in telecom solutions in silicon. Our specialty is ASIC and FPGA development and we provide a full turnkey service using industry leading design tools that incorporate support for multiple vendors. We offer a full spectrum of product development services, including system design, chip and board specification, ASIC and FPGA design and verification, DSP coding, firmware and software development, and hardware and software integration testing. We have significant experience throughout the network, including: - Transport (ATM and PPP over SONET/SDH with UTOPIA interfaces - OC-1, OC-3, OC-12 & OC-48c); - Access devices (xDSL Serial ATM Mux (Utopia & PCI Interfaces), cable modems); and - Wireless technologies (GPS, CDMA, LMDS, PCS). We have an extensive library of reusable blocks that provide our customers with the additional "value added" of cutting time to market, while reducing the risks associated with new designs. For additional information about DesignPRO, please visit our website at www.designpr.com. We are currently completing the development of 5 FPGAs for a high speed access product and expect to have 12 design engineers available in the next several weeks. We are very interested to explore all opportunities and would welcome inquiries at (613) 596-5030, or by email at abunsick@designpr.com. Thank you for your consideration. Sincerely, Andrew Bunsick General Manager DesignPRO Inc.Article: 14202
You need to put the readback component in your design and connect it up. If you are doing schematic entry, the symbol is called readback. If you are using an HDL you need to instantiate the readback macro in the xilinx library. As far as the misbehavior of you design goes, first check to make sure the questionable logic was not trimmed out (look at the trim report-hopefully your design is hierarchical and you labeled your hierarchy blocks so you can make some sense of the report). Assuming that your simulation accurately portrays the real world signals, the next place to look is at the timing analysis. Make sure all paths meet the timing you need to make the design work. Normally I do a functional simulation (no timing) till I am satisfied the logic does what it is supposed to do. Then I run the place and route followed by a static timing analysis using the timing analyzer. If you used timing constraints, check to make sure all were met. Also check the report of paths without timing constraints to make sure you got good coverage. If all is in order, check the lists of registers under each time group to make sure you didn't fall victim to overlapping constraints (sorry, that is painful but there is currently no better way to do it). If you pass a complete functional sim, nothing critical is trimmed out and everything meets timing then the design should work. If not then it is time to start looking at power, ground, clock and signal integrity. Gerd Beil wrote: > Hi to everyone! > > I'm using Xilinx M1.4 (Alliance) on a workstation (HP) and the > Xchecker-cable for downloading the generated *.bit-file on a XC4010E. My > problem is that the downloaded design doesn't do what the simulation > makes me expecting it to do. Some parts of my design are working fine > what makes me sure that I did the implementation correctly. But some > parts don't work at all and the report files don't show any > warnings/errors that seem to be the cause for this problem. I know there > is a posibility for a hardware debugging but I don't know how to get > there. > When implementing the design the check-box "Enable Bitstream > Verification and In-Circuit Hardware Dedugging" is checked in the XC4000 > Configuraion Options and the Readback Clock is set to CCLK. But when I > call the Hardware Debugger it tells me that the "design does not have a > readback block connected". The the xchecker-program run from the command > line detects a cable "serial-readback". Does this just mean that the > cable supports readback-mode if the design does too? And what exactly is > the "Readback Block"? Do I have to add some functions to my design? Or > do I have just to change some settings for the implementation? What do I > have to do to run the Hardware Debugger??? > > Bye, > > Gerd -- -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/~randrakaArticle: 14203
In a dusty old storage closet here, I dug up an XACT XC-DS21 evaluation kit, complete with demo board and dongle. I installed XACT on an IBM box, 486 w/ 16M ram. Unfortunately, when I try to run the software, I get an error to the effect of "can't find the key." Xkey reports no valid keys. I checked the key on other machines and check my bios settings to make sure that it wasn't the parallel port failing, but no machine found any keys. Does anyone have any idea on where to go from here? Be Well, Chris -- Chris Kuethe: System Administrator - U of A Math Dept http://www.[math.]ualberta.ca/~ckuethe/ pager: 403.917.6448 office: CAB553, x1704 cell: 903.9475 wargames@edmc.net ckuethe@ualberta.ca ckuethe@math.ualberta.ca ckuethe@gecko.math.ualberta.caArticle: 14204
Based on how bad Orcad schematic entry still is after 7.2 generations, I wouldn't trust anything else they do. (I tried their first VHDL and it couldn't implement a counter within a testbench - they'd released garbage) Seems they could get all the schematic features working after all these years. It's obvious their designers don't draw schematics. Everytime I do schematics with Orcad I curse. They eventually work, but its a drag. Cryptic error messages, subtle changes from version to version. Tweaky performance. Xilinx tools have their problems too, but why deal with 2 sets of problems and interface problems to boot. Seems like trusting Orcad is asking for problems. Bruce Pawel Michocki wrote in message ... >Does anyone use Orcad Express Plus v9.0 ? >Now I use Orcad Express 7.20 with Xilinx Aliance 1.5, but Orcad won't >provide >support for virtex in Express (only in Express Plus). Is it worth to upgrade >to Plus >version or simply buy Foundation from Xilinx ? >Any other suggestion ? > >Regards, >Pawel Michocki, SIM POLAND > > >PS >Any satisfied Orcad user ? (not from USA) > >Article: 14205
This looks like something a few other "free" projects would be interested in using and perhaps in contributing towards: gEDA: GNU Electronic Design Automation http://www.geda.seul.org/ The Freedom CPU Project http://f-cpu.tux.org/ mailing list: http://www.egroups.com/list/f-cpu/ (wants to build early versions of their CPU out of FPGAs, so there is faster turn-around, trying out different ideas) FreeHDL http://www.freehdl.seul.org/ A project to develop a free, open source VHDL simulator for Linux. Perhaps you can re-use some work already done on a simulation and timing estimator: Magic - A VLSI Layout System http://www.research.digital.com/wrl/projects/magic/magic.html comes with "irsim", a circuit simulator and timing estimator. Freeware, Shareware, or Academic VLSI CAD/CAE Software. http://www.mrc.uidaho.edu/cgi-bin/w3-msql/vlsi/CADfree.html Alliance: A Complete CAD System for VLSI Design http://www-asim.lip6.fr/alliance/ Alliance source code is released under the GNU General Public License (GPL). I've heard rumors that someone created a crude tool that translated from ordinary C code to VHDL (or was it Verilog ?) for his thesis; the idea was to run certain compute-intensive subroutines *really* fast, blurring the distinction between hardware and software. So, now what ? Jamie Lokier <spamfilter.jan1999@tantalophile.demon.co.uk> enthused: > I am interested in exploring FPGA fitting & optimisation strategies, for > real FPGAs. If I develop anything interesting, I will be wanting to > produce a _free_ (GPL) FPGA synthesis tool. ... > Summary: > > Project goal: To develop a program which (1) translates logic circuits (as > netlists or equation lists) to FPGA bitmaps, subject to > timing constraints; (2) can be targetted at many > different FPGA and FPGA-like architectures; (3) analyses > bitmaps to recover timing information; (4) can do > simulation of compiled bitmaps, not necessarily quickly, > but accurately. > > _Not_ a goal: Higher level languages for describing circuits are not > part of the scope of this project. Of course, > interoperability with other tools, especially those with > open development models, is important. And being free > software, others will be able to use and modify the > synthesis tool as they require. ... > -- Have already written the equivalent of GCC for FPGAs? :-) ... > - Is there any demand for this sort of tool? ... > -- Jamie -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your OwnArticle: 14206
In article <783hsj$816$1@nnrp1.dejanews.com>, <d_cary@my-dejanews.com> wrote: ... >I've heard rumors that someone created a crude tool that translated from >ordinary C code to VHDL (or was it Verilog ?) for his thesis; the idea was to >run certain compute-intensive subroutines *really* fast, blurring the >distinction between hardware and software. The one I'm aware of is called "nlc" by Christian Iseli. I downloaded a copy of it way back when he announced it (version 0.8). It spits out Viewlogic netlists, so you'll need wir2xnf to munge them into xnf. I tried to compile it, but it needs a library called LEDA which I wasn't able to track down. I don't know what happeneded to the author or if there is a newer version, but the code I have seems to be GPL. I'd be glad to send out a copy or put it up for ftp, if anyone wants to give it a try. See ya, -ingo -- /* Ingo Cyliax, cyliax@derivation.com, Tel/Fax: 812-333-4854/4852 */Article: 14207
--------------F172D99685E6E6103DD408B9 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit VLSI Architect <circuit design and implementation research engineers (Members of Technical Staff)> LOCATION: Murray Hill, NJ SHORT-TERM EXPECTATIONS: FIRST 90 DAYS The succesful candidate will be assigned to 1) identify the circuit implement opportunities for optical, wireless, switching and access communicaitons systems, 2) identify the algorithm implementation opportunties for various communicaitons systems, 3) identify the design challenges for implementing Gb/s circuits, or 4) identify the opportunity for circuit realization cycle time reduction in communication systems MEDIUM-TERM EXPECTATIONS: 90 DAYS TO ONE YEAR The successful candidate will, based on the initial studies from the first 3 months, 1) propose feasible architectures for efficient circuit realization of various communications functions, 2) generate mathematical models of the various algorithms to compare peformance and cost of various implementation options 3) provide initial designs of impementing Gb/s circuits on advanced silicon processing technologies 4) suggest methodology for cycle time management The successful candidate will have completed the design projects, and is expected to continue to identify emerging opportunities of circuit realization for communication systems IMPORTANT TASKS THIS PERSON WILL PERFORM ON AN ONGOING BASIS. 1) Identify opportunities for realizing communications functions on integrated circuits 2) VLSI architecture and high speed circuit design for communication systems 3) Realization of signal processing algorithms (layer I) on integrated circuit 4) Realization of communications protocols (layer II and above) on integrated circuits 5) Cycle time management IMPORTANT TASKS THIS PERSON WILL PERFORM DAY ONE OF JOINING YOUR ORGANIZATION. Identifying and realizing communications functions on integrated circuits SKILLS THE SUCCESSFUL CANDIDATE MUST POSSESS IN ORDER TO ACCOMPLISH THE GOALS OF THE ORGANIZATION? Skills and Experience · Candidates should possess expert design skills in one or more of the following areas together with the knowledge of the technical challenges facing the industry and, most importantly, a creative ability to lead research projects that address these challenges. Design Skills · custom circuit design · placement, routing and synthesis · VHDL design · signal integrity management · circuit simulation, timing and verification · mixed-signal design · FPGA design · Design Experience · high speed digital circuits · application-specific integrated circuit · co-processor and accelerators · application-specific standard circuit product · digital signal processors · design of peripheral, bus, and I/O · microcontrollers and RISC processors · hardware and software co-design · analog circuits including filters, analog-to-digital converters, digital-to-analog converters, trans-impedance amplifiers, and laser drivers · VSLI Design For Testability (DFT) techniques · memory and DMA Architecture and System Design Experience · integrated circuit for communication systems, including optical network, wireless, switching, and access · realization of communication protocols (layer II and above) on integrated circuits · realization of signal processing algorithms (layer I) on integrated circuits Circuit Production Experience · cycle time management · design process analysis and optimization WHAT WE ARE N O T LOOKING FOR The candidate must not expect an elaborate development process surrounding the projects, nor to have complete detailed frozen specification before beginning design. The candidate must not expect or require a large team environment – the project teams are typically small and very focused. The candidate must not expect to concentrate only on circuit design – full stream involvement from opportunity identification, system definition, planning, architecture and development is expected. The person should be very hands-on and not expect a large staff of technicians or other support personnel to be assigned to projects. WHO WILL THIS POSITION REPORT TO? Research department head ORGANIZATIONAL STRUCTURE The projects will be in about 3-4 research departments, as part of the Bell Labs Research organization. The Research organization, which has approximately 1000 people, is responsible for creating new technology and systems in support of the Lucent business units. The design research engineers have been recruited steadily over the years but the present goal is to staff another 20-30 people very quickly. TRAVEL Depending on the specific function, the travel could range from very little to a couple of times a month. The travel could be both domestic and international. YEARS OF PREVIOUS EXPERIENCE 3-5 years or more of industrial experience would be a plus. EDUCATIONAL REQUIREMENTS PhD degress from EE or CS are strongly preferred. FUNCTIONAL EXPERIENCE The same as described above would be strongly preferred. (New PhD graduates are welcome as well.) INDUSTRY EXPERIENCE R&D organizations from telecommunication companies or integrated circuits vendors. WHAT DO YOU SEE AS SOME OF THE KEY "PLUSES" OF THE POSITION AND YOUR ORGANIZATION? The successful candidates will have the opportunity and freedom to define circuit design and implementation projects for making communications systems more efficient and more cost-effective. The working environment will be with many talented and highly motivated researchers at Bell Labs. COMPENSATION The candidate will receive a base salary, plus an annual bonus based on the financial performance of Lucent, plus an additional bonus based on the performance of the individual in the preceding year. DESIRED CANDIDATE TRAITS Exceptional technical skills, self confidence and the ability to work independently are essential. The candidate will work with other very strong technical contributors who expect outstanding performance from all of their colleagues. OTHER REQUIREMENTS Advancement in the Bell Labs environment typically requires taking a lead technical role in an innovative and successful project. Key to advancement is the ability to identify new technology opportunities, developing and clearly articulating a plan for capitalizing on the opportunity, securing management support, leading the effort, and delivering as proposed. MISC This organization has already 20-30 experienced researchers in the same area. The laboratory is called “Wireless Research Laboratory” but the projects are more than just for wireless. >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Visit Lucent on ejob @: http://www.ejob.com/lucent.htm http://www.ejob.com/lucent3.htm email: mailto:lucent@ejob.com --------------F172D99685E6E6103DD408B9 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: 8bit <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML> <B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>VLSI Architect <circuit design and implementation research engineers (Members of Technical Staff)></FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>LOCATION:</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Murray Hill, NJ</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>SHORT-TERM EXPECTATIONS: FIRST 90 DAYS</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The succesful candidate will be assigned to</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>1) identify the circuit implement opportunities for optical, wireless,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>switching and access communicaitons systems,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>2) identify the algorithm implementation opportunties for various</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>communicaitons systems,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>3) identify the design challenges for implementing Gb/s circuits, or</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>4) identify the opportunity for circuit realization cycle time reduction in</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>communication systems</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>MEDIUM-TERM EXPECTATIONS: 90 DAYS TO ONE YEAR</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The successful candidate will, based on the initial studies from the first 3</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>months,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>1) propose feasible architectures for efficient circuit realization of</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>various communications functions,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>2) generate mathematical models of the various algorithms to compare</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>peformance and cost of various implementation options</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>3) provide initial designs of impementing Gb/s circuits on advanced silicon</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>processing technologies</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>4) suggest methodology for cycle time management</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The successful candidate will have completed the design projects, and is</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>expected to continue to identify emerging opportunities of circuit</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>realization for communication systems</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>IMPORTANT TASKS THIS PERSON WILL PERFORM ON AN ONGOING BASIS.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>1) Identify opportunities for realizing communications functions on</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>integrated circuits</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>2) VLSI architecture and high speed circuit design for communication systems</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>3) Realization of signal processing algorithms (layer I) on integrated</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>circuit</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>4) Realization of communications protocols (layer II and above) on</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>integrated circuits</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>5) Cycle time management</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>IMPORTANT TASKS THIS PERSON WILL PERFORM DAY ONE OF JOINING YOUR ORGANIZATION.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Identifying and realizing communications functions on integrated circuits</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>SKILLS THE SUCCESSFUL CANDIDATE MUST POSSESS IN ORDER TO ACCOMPLISH THE GOALS OF THE ORGANIZATION?</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Skills and Experience</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· Candidates should possess expert design skills in one or more of the</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>following areas together with the knowledge of the technical challenges</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>facing the industry and, most importantly, a creative ability to lead</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>research projects that address these challenges.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1> Design Skills</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· custom circuit design · placement, routing and synthesis</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· VHDL design · signal integrity management</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· circuit simulation, timing and verification · mixed-signal design</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· FPGA design ·</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1> Design Experience</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· high speed digital circuits · application-specific integrated circuit</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· co-processor and accelerators · application-specific standard circuit</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>product</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· digital signal processors · design of peripheral, bus, and I/O</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· microcontrollers and RISC processors · hardware and software co-design</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· analog circuits including filters, analog-to-digital converters,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>digital-to-analog converters, trans-impedance amplifiers, and laser drivers</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· VSLI Design For Testability (DFT) techniques</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· memory and DMA</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1> Architecture and System Design Experience</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· integrated circuit for communication systems, including optical network,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>wireless, switching, and access · realization of communication protocols</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>(layer II and above) on integrated circuits</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· realization of signal processing algorithms (layer I) on integrated</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>circuits</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1> Circuit Production Experience</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>· cycle time management · design process analysis and optimization</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>WHAT WE ARE N O T LOOKING FOR</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The candidate must not expect an elaborate development process surrounding</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>the projects, nor to have complete detailed frozen specification before</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>beginning design. The candidate must not expect or require a large team</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>environment – the project teams are typically small and very focused. The</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>candidate must not expect to concentrate only on circuit design – full</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>stream involvement from opportunity identification, system definition,</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>planning, architecture and development is expected. The person should be</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>very hands-on and not expect a large staff of technicians or other support</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>personnel to be assigned to projects.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>WHO WILL THIS POSITION REPORT TO?</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Research department head</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>ORGANIZATIONAL STRUCTURE</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The projects will be in about 3-4 research departments, as part of the Bell</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Labs Research organization. The Research organization, which has</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>approximately 1000 people, is responsible for creating new technology and</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>systems in support of the Lucent business units.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The design research engineers have been recruited steadily over the years</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>but the present goal is to staff another 20-30 people very quickly.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>TRAVEL</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Depending on the specific function, the travel could range from very little</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>to a couple of times a month. The travel could be both domestic and</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>international.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>YEARS OF PREVIOUS EXPERIENCE</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>3-5 years or more of industrial experience would be a plus.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>EDUCATIONAL REQUIREMENTS</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>PhD degress from EE or CS are strongly preferred.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>FUNCTIONAL EXPERIENCE</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The same as described above would be strongly preferred. (New PhD graduates</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>are welcome as well.)</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>INDUSTRY EXPERIENCE</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>R&D organizations from telecommunication companies or integrated circuits</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>vendors.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>WHAT DO YOU SEE AS SOME OF THE KEY "PLUSES" OF THE POSITION AND YOUR</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>ORGANIZATION?</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The successful candidates will have the opportunity and freedom to define</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>circuit design and implementation projects for making communications systems</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>more efficient and more cost-effective. The working environment will be with</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>many talented and highly motivated researchers at Bell Labs.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>COMPENSATION</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>The candidate will receive a base salary, plus an annual bonus based on the</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>financial performance of Lucent, plus an additional bonus based on the</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>performance of the individual in the preceding year.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>DESIRED CANDIDATE TRAITS</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Exceptional technical skills, self confidence and the ability to work</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>independently are essential. The candidate will work with other very strong</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>technical contributors who expect outstanding performance from all of their</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>colleagues.</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1> </FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>OTHER REQUIREMENTS</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Advancement in the Bell Labs environment typically requires taking a lead</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>technical role in an innovative and successful project. Key to advancement</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>is the ability to identify new technology opportunities, developing and</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>clearly articulating a plan for capitalizing on the opportunity, securing</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>management support, leading the effort, and delivering as proposed.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>MISC</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>This organization has already 20-30 experienced researchers in the same</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>area. The laboratory is called “Wireless Research Laboratory” but the</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>projects are more than just for wireless.</FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>></FONT></FONT></B><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1></FONT></FONT></B> <P><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>Visit Lucent on ejob @:</FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1><A HREF="http://www.ejob.com/lucent.htm">http://www.ejob.com/lucent.htm</A></FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1><A HREF="http://www.ejob.com/lucent3.htm">http://www.ejob.com/lucent3.htm</A></FONT></FONT></B> <BR><B><FONT FACE="Arial,Helvetica"><FONT SIZE=-1>email: <A HREF="mailto:lucent@ejob.com">mailto:lucent@ejob.com</A></FONT></FONT></B></HTML> --------------F172D99685E6E6103DD408B9--Article: 14208
Well that is certainly true. >It's obvious their designers don't draw schematics. -- Peter. Return address is invalid to help stop junk mail. E-mail replies to zX80@digiYserve.com but remove the X and the Y. Please do NOT copy usenet posts to email - it is NOT necessary.Article: 14209
In article <76svdr$8t8$1@info-server.surrey.ac.uk>, nospam_ees1ht@ee.surrey.ac.uk (Hans) wrote: > > Actel recently published an article on DPLL's in an Actel FPGA. Unfortunately I > can remember were I've seen it. I remember that the author was Steff > Niewadomski (spelling?). > > Hans. > > Brett George wrote in message <367F428F.5641E93E@clarityeq.com>... > >Hi all, > > > >Has anyone tried implementing a PLL in an FPGA (eg. ALTERA), we are > >thinking of using it to reduce clock jitter from an externally generated > >clock source. > > Altera's 10KE devices with PLLs will offer the Clock lock clockboost options to reduce clock delay and skew, and to perform clock multiplication. More info on these features can be found on the Altera Web page at www.altera.com -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your OwnArticle: 14210
Hi to everyone ! We were trying to use JTAG Progammer from Xilinx tools. We were experimenting with Boundary Scan and it long time did not work. After a lot of headache we realised, that Boundary Scan Chain Debug program instead of 1 pulse of TCK put 3 pulses, instead of 5 pulses put 7 pulses and so, on (if you remove two aditional pulses on TCK it works !). I hope our experience could help to someone. Now I have question. Does anybody have any experience with using ReadBack property in JTAG interface ? There is not much information in the Xilinx docummentation about it. Thanks Milos Becvar ASICentrum, PragueArticle: 14211
> This looks like something a few other "free" projects would be interested in > using and perhaps in contributing towards: Thanks for your interest! The list of resources is very useful. If you find any more, please post them or send them my way. > I've heard rumors that someone created a crude tool that translated from > ordinary C code to VHDL (or was it Verilog ?) for his thesis; the idea was to > run certain compute-intensive subroutines *really* fast, blurring the > distinction between hardware and software. There are a few "C" to netlist translators around. These are the commercial ones I know of: - Handel-C (which I am affiliated with) http://www.embedded-solutions.ltd.uk/ProdApp/handelc.htm - C Level Design (formerly Compilogic) http://www.cleveldesign.com/ I have heard of a few academic (i.e. non-commercial) translators, but I don't know of any in particular. I don't think any of them translate the full C language properly. Handel-C has a nice approach to parallel processing, concurrency and timing. > So, now what ? Read other people's source, encourage interest, check legal position. Find out what I can use that already exists. Think real hard about retargettable fitting. -- JamieArticle: 14212
i've put some C code up at: http://www.riverside-machines.com/pub2/xilinx/configuration/configuration.htm i've also included some useful stuff that has has appeared on this newsgroup. feel free to mail me any useful docs/code/etc, or articles that i've missed from the group, and i'll add them. evanArticle: 14213
Has anyone ever tried to use the 'secondary global nets' for clock distribution in a Xilinx Virtex ? The Virtex only has 4 primary global clock buffers, but I have a design that uses more than 4 clocks, so I need to use secondary nets. So far, I have not been able to convince the Xilinx tool to use them. It routes the clocks using a mess of local routing, which give large delay and skew. Anyone know how to do this ? -- ----- Terry Fraser Hardware Designer Applied Microelectronics Inc Halifax, NS Phone: (902) 421-1250 ext 269 FAX: (902) 429-9983 mailto:fraser@appliedmicro.ns.ca http://www.appliedmicro.ns.caArticle: 14214
This week's Tech Note is from Vantis and discusses the design considerations when chosing between using CPLDs and FPGAs. It can be found on http://www.edtn.com/pld Murray Disman Editor EDTN PLD Design CenterArticle: 14215
After synthesizing my VHDL source code and creating the ".xnf" file I create a project in Foundation Design Manager 1.4. Everything works until I try to implement the project, at that point I receive an error that states "illegal command line for invoking the Flow Engine". I can't find anywhere what command line it is trying to run or any information of the source of error. Any ideas would be helpful. Thanks NickArticle: 14216
stuart2775@my-dejanews.com wrote: > > >Has anyone tried implementing a PLL in an FPGA (eg. > ALTERA), we are > > >thinking of using it to reduce clock jitter from an > externally generated > > >clock source. > > > Altera's 10KE devices with PLLs will offer the Clock lock > clockboost options > to reduce clock delay and skew, and to perform clock > multiplication. More > info on these features can be found on the Altera Web page > at www.altera.com > The Xilinx Virtex devices ( two of them have been shipping since Oct/Nov '98 ) each have four delay-locked loops that can perform the functions asked for or mentioned in this thread. For detailed information, click on http://www.xilinx.com/xapp/xapp132.pdf Peter Alfke, Xilinx ApplicationsArticle: 14217
I've only done functional simulation and it operates ok, it's when I download the bit file to the FPGA that I physically test an expensive wire. Also, is there anyway to make that IOB setting from the DOS prompt? the syntax for the "map" command shows settings to pack FF/latches into input, output, or both types of IOB's (map -pr i|o|b ...). Thanks Nick Rickman wrote: > Here is my guess as to what is happening. I think your circuit is being > synthesized correctly. But the FF is being pushed into the IOB of the > input port. The messages above indicate this. The warning may be from > the fact that the input and output IOB are now connected by a wire. > > If you go to the project manager and select "Implementation" and > "Implementation Option" you will get the options dialog box. On the line > that says "Implementation" click the "Edit Template" button and a new > dialog box will come up. Find the "Pack I/O Registers/Latches into IOBs > for:" control. Select "OFF". Now click OK and OK again. Now redo your > Implementation. and you should get a good result. > > When you say that the output follows the input without waiting for the > clock, are you running a simulation, or testing a chip? If a simulation, > is it a functional simulation or timing (after place and route)? > Rick CollinsArticle: 14218
Check out Xilinx Answer 2141 on the Xilinx website. The URL is: http://www.xilinx.com/techdocs/2141.htm Regards, -Hobson Frater Xilinx ApplicationsArticle: 14219
Dominic Reitman wrote in message <369F99A4.D859CB19@iupui.edu>... >I'm beginning to believe that the problem does not exist in the synthesis >but rather the implementation. I've been implementing the designs at the >dos prompt (not using XILINX Design manager) because the XILINX software >gives me an error. What error would that be? MAP (Foundation 1.5) is throwing access faults, even on designs that previously mapped. Xilinx's response was, "Can you install the tools on another machine?" Luckily, they gave me a site license rather than node-locked so I was able to install the tools on another NT box; voila, the design implemented - on the other box. Not exactly helpful. The response to that was, "You'll have to reinstall NT." Which I've accepted as the only answer. Actually, I blame Microsoft more than Xilinx. Running map from the NT command line causes the same crash. I pulled the command-line parameters from what the GUI spit out. -- andy ------------------------------------------ Andy Peters Sr. Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters@noao.edu "In the beginning, there was darkness. And it was without form, and void. And there was also me!" -- Bomb #20, John Carpenter's "Dark Star"Article: 14220
Ido Kleinman wrote in message <77lek9$5mo$1@news2.inter.net.il>... >I've been using Foundation 1.4 with it's Metamor synthesis for a while now, >and I've got a few working designs. >I recently moved to Foundation 1.5 and it's FPGA Express synthesis - In case it wasn't clear from Evan's post: Metamor is a VHDL'93-compliant tool, FPGA Express is VHDL'87-compliant (with a couple of '93 things thrown in). The problems you're having are a result of Synopsys still living in the Reagan era. >1. FPGAEXP won't accept: >if (LowerAddressBus > X"01F0") then >LowerAddressBus is just a std_logic_vector(15 downto 0). >it yells about type mismatch between left/right binary operand. The X seems >to be disturbing - it would only accept Binary notation without any prefix >to the " character so I have to write: The X"xxxx" construct is valid only for bit vectors in '87. (See Evan's comments.) SLVs are fine with it with '93. That's why Metamor was happy and FPGA Express was not. >I tried using based literals such as 16#01F0# and variations - and it won't >work. Any solution, or I will have to do all my comparisons in Binary? Make sure you've included ieee.std_logic_arith.all and ieee.std_logic_unsigned (or signed) and do a conversion: if (LowerAddressBus > CONV_STD_LOGIC_VECTOR(X"01F0", 16) then which of course is a Synopsys work-around. >2. My inhibit_buf attribute on my global Clk signal wasn't accepted! What's >the attribute for inhibiting insertion of IBUFs/OBUFs/etc on FPGAEXP? I >don't want my clock signal buffered with a standard IBUF, that why I usually >instanciate a BUFG/BUFGS for it and inhibit other buffers on the pad in the >code. FPGA Express is smart enough to recognize a clock as a clock (based on sensitivity lists and how the signals are used in processes), so you don't need to inhibit the IBUF on your clock signal. A BUFG is automagically instantiated for you. >3. After I made the annoying adjustments to my code, my previously working >simulation files yielded strange new results in which some output signals >were defined as "????" - what's this? I thought std_logic "only" has 9 >states... Sounds like a VHDL'87 vs '93 problem. What simulator are you using? Make sure you compile all of your sources as VHDL'87 now! >4. What about my two-dimensional arrays? I am not trying to synthesize >anything special - just some better-looking-VHDL multi-bit look-up tables... > FPGAEXP won't support it? >Forever? With Synopsys, who knows? Maybe when VHDL'2K comes out. Then they'll be at the '93 level. -- andy ------------------------------------------ Andy Peters Sr. Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters@noao.edu "In the beginning, there was darkness. And it was without form, and void. And there was also me!" -- Bomb #20, John Carpenter's "Dark Star"Article: 14221
On Wed, 20 Jan 1999 15:55:26 GMT, Terry Fraser <noone@dont.bother.com> wrote: >Has anyone ever tried to use the 'secondary global nets' for clock >distribution >in a Xilinx Virtex ? This was probably the first question I asked of Xilinx when 1.5 dropped through the door with Virtex. It appears to be rather a bone of contention, especially for most (if not all) of the Xilinx FAEs here in the UK. The "official" line is that the router is supposed to be smart and use these for wide fanout signals, but as yet there is no way of actually telling the tool to use them. My common example is a clock divider (by 100) which even on it's own gets crappy skew. In Spectrum I just BUFFER_SIG it with a BUFGLS for 4000XL and hey-presto, nice clock. In Virtex, I have not found anything to duplicate this. I couldn't even see ay sign of these 24 internal buffers in EPIC. Perhaps they are not true buffers? Mr Alfke, can you help? Cheers Stuart An employee of Saros Technology, The HDL Solutions Company: Renoir Model Technology Exemplar Logic TransEDA www.saros.co.uk (I sell these products, so paint me biased)Article: 14222
We have several job openings in Portland (Oregon), San Jose (California) Boulder (Colorado) and Phoenix (Arizona) We provide US work visa for qualified candidates. Please send us text only resume if you are interested. Software Engineer: Position #1, San Jose: Experience with one or more of the following is desired: Integrated Circuit or board design, HDL (VHDL, Verilog, or ABEL) design, Simulation and/or Synthesis tools, MFC, GUI, and graphics software development, and compiler design or GUI oriented text editors development. Pos #2: Development and enhancement to existing download and test tools to generate Java code. Development of suite of tools to execute Java executables and verification of Java executables on a variety of platforms. Job Requirements: BSEE or CS. Experience with Java, C, C++ (3-5 years). Digital logic design basics. Position #3, Arizona: UNIX, C++, EDA knowledge with 2 years exp. Position #4: Senior Test Engineer Job Responsibilities: Responsible for Test program development, augmentation and characterization, assisting with Yield Analysis/Enhancement, Statistical Data Collection and Interpretation, and the Transfer of Test Programs and associated Test Hardware into production. Job Requirements: BSEE or equivalent. 2-3 years experience with Versatest V11xx/V12xx/V33xx, strong programming skills>> "C" a must. Strong Memory Background, strong Logic Theory, and a desire to be challenged in a dynamic field of programmable logic. Applications Engineer 1. BSEE and 3 years minimum digital design experience. Working knowledge of FPGA and ASIC applications. At least 3 years of experience working with digital ICs and design tools. In-depth Hardware Descriptive Language (VHDL and/or Verilog) capability, strong synthesis and simulation tools experience, and good communications skills. Application Engineer 2. Provide expert support of customers using Java as an essential system solution. This involves understanding the operation and use of IEEE Std. 1149.1 as well as the capabilities of Xilinx's associated H/W and S/W. Job Requirements: BSEE or CS. Digital logic design incorporating design-for-test rules. In particular those related to IEEE Std. 1149.1 Good oral and written communication skills. Applications Engineer (Physical Design) 3. Perform Place and Route Activity for the HardWire designs. Assist in layout for HardWire Base array designs. Job Requirements: AA degree with 5-8 yrs. experience as a mask designer. Excellent knowledge of Cadence Gate Ensemble tool. Knowledge of Edge layout system desirable. Very good understanding of Dracula DRC/LVS flows and Ability to debug DRC/LVS results. Familiarity with Fracturing and mask view procedures. Layout Designer Custom IC Layout of the CPLD products, layout verification using dracula LVS and DRC tools automatic routing. Job Requirements: Hands-on experience with EDGE and OPUS layout verification tools (LVS, DRC). And requires 5+ years of related experience. CAD Engineer Responsible for CAD software development and support. Also responsible support of Cadence tools. BSEE. 3+ yrs. experience in CAD, IC design, or software development. Knowledge of IC design and good software development skills. Knowledge of Cadence tools. Programming in skill, c, UNIX shell, Perl. Sincerely: Gary Lang http://www.acdcon.com (949) 360-4155Article: 14223
Terry Fraser wrote: >The Virtex only has 4 primary global clock buffers, but I have >a design that uses more than 4 clocks, so I need to use secondary >nets. So far, I have not been able to convince the Xilinx tool to >use them. It routes the clocks using a mess of local routing, which >give large delay and skew. > >Anyone know how to do this ? The "secondary global nets" is a misleading term that has been used to describe how Virtex handles large fanout nets. A correct term would be backbone routing. The Virtex backbone routing is comprised of 12 horizontal longlines with specialized hex-lines in the top and bottom IOB rows and 12 vertical longlines with specialized hex-lines in the vertical columns with the hex-lines having the ability to connected directly to the CLB, IOB and RAM clock and control inputs. This provides for up-to 24 high-fanout, reduced-skew and reduced-delay nets with a maximum of 12 per column. Since these are flexible resources without a specific entry or exit point there isn't a library cell that represents them. The router has built-in algorithms to determine the need to use these routes and the recipe for building the cells. We have found two minor problems in routes to the IOB and BlockRAMs that will be fixed in the next M1.5i Service Pack release. Designers can improve the overall delay and skew of the backbone nets in several ways. 1) Place the source of the net in the middle of the top or bottom row. This provides for easier access to the horizontal longlines at the start of the backbone. This can be done with a simple location constraint in the design's UCF file. INST myclk_pad LOC = AL17 ; # Valid for BG432 package or INST myclkdv_reg LOC = CLB_R32C24.S1 ; # Valid for XCV300 2) Cluster the logic together to reduce overall delay. This can be done with a simple area location constraint in the design's UCF file. INST mymodule LOC = CLB_R1C1:CLB_R32C10 ; #Valid for XCV300 Routes that use these resources are quite easily determined in the EPIC viewer. When you have highlighted the net you will see a top or bottom backbone that feeds the vertical longlines that in turn feed the vertical hex-lines that then connect directly to the CLB, IOB or RAM inputs. You should not see any other horizontal routes in the core, unless you are also routing to LUT inputs. I would include a pretty picture here showing a typical net, but we don't want binaries cluttering up the newsgroup. Ed McGettigan -- Xilinx Inc. -- Ed --Article: 14224
> There are a few "C" to netlist translators around. These are the > commercial ones I know of: > > - Handel-C (which I am affiliated with) > http://www.embedded-solutions.ltd.uk/ProdApp/handelc.htm > - C Level Design (formerly Compilogic) > http://www.cleveldesign.com/ > > I have heard of a few academic (i.e. non-commercial) translators, but I > don't know of any in particular. Coincidentally, this is my research topic as well. Here are some notes I've gathered on other attempts: C to Netlist Compiler (nlc): Given a program written in a ``small subset of C++ with a few extensions,'' nlc compiles the program into an FPGA configuration. ftp://lslsun5.epfl.ch/pub/ CEDES: C++ as an HDL. Several C++ classes are provided to allow users to generate netlists programmatically, and synthesize them into an FPGA configuration. http://www.fbeedle.com/bookinfo/09-0.html COBRA-ABS: A. Duncan, D. Hendry, P. Gray, ``An Overview of the COBRA-ABS High Level Synthesis System for Mult-FPGA Systems,'' Proceedings: IEEE Symposium on FPGAs for Custom Computing Machines, Napa Valley, CA, to appear, April 1998. DISC: The Dynamic Instruction Set Compiler maps various function calls onto a partially reconfigurable FPGA. Precompiled functions accelerate normal execution. Data Parallel C Compiler: M. Gokhale, R. Minnich, ``FPGA Computing in a Data Parallel C,'' IEEE Workshop on FPGAs for Custom Computing Machines, Napa Valley, CA, pp. 94-101, April 1993. GigaOps X Language Compiler (xc): http://www.reconfig.com/giga/xc_ic.htm Handel C: Uses ``a subset of the C language extended with a few powerful extra features'' to describe a circuit. Sisal: A ``functional language for parallel numerical computation''. http://www.physics.nmt.edu/raymond/sisaltutorial/Tutorial.html Systolic Parallel C: ``C dialect with systolic and parallel extentions'' R. Zoz, http://www-mp.informatik.uni-mannheim.de/groups/mass_par_1/projects/spc.html Transmogrifier C: Takes in ``a restricted subset of the C programming language'' and generates a netlist. http://www.eecg.toronto.edu/EECG/RESEARCH/tmcc/tmcc/ > I don't think any of them translate the full C language properly. > Handel-C has a nice approach to parallel processing, concurrency and timing. So far as I know, none of these can take full ANSI-C code. My research attempts to support full ANSI-C by running the input through the freely available retargettable compiler, LCC. I then just use the output from LCC as the input to the core compiler. --Jim (http://www.cs.princeton.edu/software/lcc/) (http://www.ee.vt.edu/~ccm/)
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