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
sundberg jeffrey r wrote: > need to imprelement a cruise control/engine... >... what is the best target device platform to implement the design on Hello Jeff, The auto industry has been doing that job in little single chip microcontrollers for many years. Their environment is far harsher than yours. The Microchip PIC and National COP-8 family of processors and tools are available from DigiKey. Motorola's 68HC05 family and the various flavors of 8051 from many manufacturers may also be suitable. If your needs are more sophisticated, I'd check your local Hitachi rep to see if they still have their $99 SH-1 evaluation board. This gives you a very capable RISC microcontroller with 128KBytes of external RAM and some external EPROM, complete with GNU development toolchain, for $99. Regards, ScottArticle: 5301
I'm looking at Lucent ORCA2 series & Xilinx 4000E series parts; I also looked at Flex10k from Altera. For a previous design I selected a F10k50 because the memory requirements fit nicely (number of EABs available, distribution of memory, etc.) Unfortunately (?), even though this design requires a lot less memory, it is split up into several small chunks which may (statistically speaking) need to be accessed concurrently. The total size is estimated at about 700 - 800 bytes, which by my calculations should easily fit in an ORCA2C10A or better, or a Xilinx 4010E or better. Of course, how much extra logic is required (5k? 10k? 15k gates?) may push for a larger part. I have experience with the X4010, using older DOS Viewlogic and Xact 4 & 5, so I have a feel for the process. I haven't used ORCA except to do a trial evaluation (for the 20 kbit design), and I found that design impossible to route inOR2C40A: but then, the embedded RAM was about half of the chip. This design is expected to migrate to an ASIC, so HDLs are desired (probably VHDL with Synopsys, which both vendors support.) I'm interested in people's [recent] experiences with these two products and the development environment. Thanks, Jason T. Wright P.S. Is it worth (possible) doing a design for the larger parts on the PC (under NT)? I recall long place & route times for the 4010 on a 486 PC, and don't want to go through that again unnecessarily.Article: 5302
I have a problem with a 16L8 PLD, and I can't find comp...pld, so I thought I'd try here - if anyone knows a better source, please say so. I need to copy a 16L8 to repair some obsolete equipment; I don't have a programmer, but can lay my hands on one with some difficulty. What I don't know is if it's possible to directly copy a PLD using a programmer, or if it's possible to derive the PLD logic and thus program another PLD from scratch. Any thoughts welcome. Will cwr@crash.cts.comArticle: 5303
On 5 Feb 1997 21:13:07 GMT, cwr@cts.com (Will Rose) wrote: >I have a problem with a 16L8 PLD, and I can't find comp...pld, >so I thought I'd try here - if anyone knows a better source, please >say so. > >I need to copy a 16L8 to repair some obsolete equipment; I don't >have a programmer, but can lay my hands on one with some difficulty. >What I don't know is if it's possible to directly copy a PLD using >a programmer, or if it's possible to derive the PLD logic and thus >program another PLD from scratch. If the two devices have the same jedec file, it's a no-brainer. Just copy them (changing family/pinout codes as necessary); however, if the manufacturer's parts aren't compatible enough (though I expect all 16L8s will be), then it is a more challenging problem (such as 16L8 to 16V8). Jason T. Wright > >Any thoughts welcome. > >Will >cwr@crash.cts.com >Article: 5304
Nicholas, Yes it certainly is possible. Just follow the instructions in this Solution Record that can be found on our Web site: http://www.xilinx.com/techdocs/789.htm For more solutions, take a look at our Solution Search Engine: http://www.xilinx.com/support/searchtd.htm Hope this information helps you out! -- David Dye Software Applications Engineer Xilinx, Inc., San Jose, CA ph> (800) 255-7778 email> david.dye@xilinx.com Nicholas C. Weaver wrote: > > We have Workview office and the XACT tools, but unfortunately > I don't have the documentation handy. Is it possible to do back > annotation from a placed & routed Xilinx design so that the workview > simulator (speedwave) gives correct timing information? And how is > this done? > > -- > Nicholas C. Weaver Ash C++ durbatuluk, ash C++ gimbatul, > nweaver@cs.berkeley.edu ash C++ thrakatuluk agh burzum-ishi krimpatul! > http://www.cs.berkeley.edu/~nweaver/ > It is a tale, told by an idiot, full of sound and fury, .signifying nothing.Article: 5305
Wasn't there a proposal years ago which used memory/CPU hybrid chips to break DES [exhaustive search] within a day for around $100k? But the machine was scalable. Wonder what happened to that... MIKE... Steve Casselman wrote: > > The guy who broke the 40-bit code made the LA times. > It took 250 workstations 3 1/2 hours. If he had to search > to the last combination it would have taken 10 hr. I > figured that a 50MHz engine would have taken 6 hrs. > The 48-bit key search (which is next) should take > 256 times longer. > > Steve CasselmanArticle: 5306
On Wed, 05 Feb 1997 18:31:43 GMT, Jason.Wright@ebu.ericsson.com (Jason T. Wright) wrote: >I'm looking at Lucent ORCA2 series & Xilinx 4000E series parts; I also >looked at Flex10k from Altera. For a previous design I selected a >F10k50 because the memory requirements fit nicely (number of EABs >available, distribution of memory, etc.) > >Unfortunately (?), even though this design requires a lot less memory, >it is split up into several small chunks which may (statistically >speaking) need to be accessed concurrently. The total size is >estimated at about 700 - 800 bytes, which by my calculations should >easily fit in an ORCA2C10A or better, or a Xilinx 4010E or better. Of >course, how much extra logic is required (5k? 10k? 15k gates?) may >push for a larger part. > >I have experience with the X4010, using older DOS Viewlogic and Xact 4 >& 5, so I have a feel for the process. I haven't used ORCA except to >do a trial evaluation (for the 20 kbit design), and I found that >design impossible to route inOR2C40A: but then, the embedded RAM was >about half of the chip. > >This design is expected to migrate to an ASIC, so HDLs are desired >(probably VHDL with Synopsys, which both vendors support.) > >I'm interested in people's [recent] experiences with these two >products and the development environment. > >Thanks, > >Jason T. Wright > >P.S. Is it worth (possible) doing a design for the larger parts on >the PC (under NT)? I recall long place & route times for the 4010 on >a 486 PC, and don't want to go through that again unnecessarily. > > Most of Lucents sales comes from captive AT&T accounts. ORCA is not a major player in the general marketplace. Xilinx offers a version of their foundation sw package which offers vhdl abel or schematic entry. The new routing sw that xilinx is releasing now for workstations and in april for windows NT will significantly improve your route times. the new 4Kex and 4kxl devices have double the routing resources of the 4ke. the next three parts available from xilinx will be the 4062(now), 4085(2q97) and 40125(3q97). I found the press release on their web site. (www.xilinx.com) Mike Seither Xilinx, Inc. (408) 879-6557 mike.seither@xilinx.com FOR IMMEDIATE RELEASE XILINX ROADMAP FOR HIGH DENSITY FPGAS CALLS FOR 0.25 MICRON DEVICES DURING 1997 SAN JOSE, Calif., February 3, 1997--Xilinx, Inc., (NASDAQ:XLNX) today unveiled a product roadmap that points to delivery this year of high density field programmable gate array (FPGA) devices manufactured using advanced deep submicron process technologies. In January Xilinx began shipping samples of the company's first FPGA devices using 0.35 micron process technology: the XC4062XL, XC4036XL and XC4013XL parts. During the second quarter of 1997 Xilinx expects to have a family of three-layer metal 0.35 micron devices. The 10-member XC4000XL family will comprise devices ranging in density from 5,000 to 85,000 gates, all operating at 3.3 volts with 5-volt tolerant I/Os. In a statement of direction, the company also said its expects to ship this year products that will be manufactured with more advanced five-layer metal 0.25 micron technology. Samples are expected during the summer of 1997 of a 0.25 micron FPGA device offering 36,000 logic gates, with a 125,000-gate device to follow shortly thereafter. The devices will operate at 2.5 volts at the core voltage level, while the I/Os will operate at 3.3 volts and 2.5 volts, and tolerate 5-volt signals. If technical advancements with CMOS process technology continue at their current pace, the company said it could be possible to manufacture FPGA devices with two million gates by the year 2001. Xilinx released a white paper, "The Future of FPGAs," in conjunction with the company's statement of direction for high density programmable logic products. Among other topics, the white paper provides an analysis of the coming voltage migration that will confront digital designers. For more than 20 years, the white paper points out, 5 volts has been the standard supply voltage for most electronic components, including FPGAs. Design engineers could expect predictable gains in price and performance without having to change designs to accommodate smaller and faster versions of an FPGA as the devices were manufactured with successively finer processes, moving from 3.0 micron to 2.0 micron, then 0.8, 0.6 and finally 0.5 micron. But future generations of FPGAs, and all other semiconductors manufactured with 0.35, 0.25 and 0.18 micron processes, will require incremental moves to lower voltages. This will force the issue of ever-changing supply voltages into customers' design and product life cycle assumptions. Xilinx said that its product strategy calls for devices that will have voltage compatibility with the previous two generations of products. The new 3.3-volt Xilinx XC4000XL products which are now sampling, and the 2.5-volt products slated for shipment later this year, can co-exist in 5-volt environments. Going forward, Xilinx plans to set forth a new methodology, also detailed in the white paper, for measuring the logic density of FPGAs. The new density metric is based on logic cells that typically consist of a 4-input look-up table and one flip flop. The company said this method will make it possible for customers to make meaningful comparisons of competing products. The goal is to end the confusion caused by programmable logic vendors who include on-chip RAM to arrive at inflated gate densities. Xilinx said the new XC4062XL device, when measured using logic cells, is currently the industry's largest FPGA device. The Xilinx XC4085XL device, the flagship product in the company's new 3.3-volt family planned for shipment this summer, is expected to offer 40 percent more density than the XC4062XL device, while the denser 2.5-volt device scheduled for sampling later this year has been designed to offer twice as much density as the XC4062XL device. In addition, Xilinx said its HardWire program will support the company's new, higher density devices for customers who require a migration path to a mask programmable solution in order to achieve the lowest possible cost. The Xilinx roadmap includes fully compatible 3.3-volt HardWire devices in 1997 and 2.5-volt HardWire devices in 1998. HardWire devices are guaranteed to be fully logic- and pin-compatible with their FPGA counterparts at ASIC-like prices, and they eliminate the need for customers to write test vectors or perform timing simulations as part of the conversion process. Xilinx also plans to announce the general availability of a new software platform that supports very high density designs and features ASIC-like design flows and hooks to synthesis. The software, now in the hands of selected users, is scheduled for general availability later in 1997.Article: 5307
On 3 Feb 1997 09:09:40 GMT, pac1@waikato.ac.nz (pac1) wrote: >I'm trying to decide if it is worth buying Xilinx Foundation BASE as >oppose to the standard XACT Step package (because its much cheaper). > >Can anyone tell me if it is any good, what are its limitations - I've >looked on Xilinx home page and can't find a good description of what it >can or can't do! > >Thanks Peter. Peter, There is also a foundation eval kit available. Call one of the xilinx distributors, hamilton hallmark, insight or marshall. the limitation is up to 5000 gates on the base package plus only 3 months support vs a year for the standard. The base also includes for now an unlimited site license except for abel. It's a great package for $600.Article: 5308
On Sat, 18 Jan 1997 19:01:36 -0500, David Charles Hirschfield <dch+@andrew.cmu.edu> wrote: >I'm currently working on a project that requires continuous programming >and reprogramming of a Xilinx XC4000 FPGA board. > >Due to some strange setup requirements, we are not going to be able to >directly use the xchecker application and cable to program the board. > >Does anyone have any information regarding the technical details of >programming Xilinx boards? > >Any help would be greatly appreciated, >-David Hirschfield > >+-===========================================================================-+ >| --== e-mail ==-- | >| _/_/_/ _/_/ _/ _/ _/_/_/ dch+@andrew.cmu.edu | >| _/ _/ _/ _/ _/ _/ _/ or | >| _/ _/ _/_/_/_/ _/ _/ _/_/ cddch@paleo.giss.nasa.gov | >| _/ _/ _/ _/ _/_/ _/ | >| _/_/_/ _/ _/ _/ _/_/_/ --== WWW ==-- | >| http://www.contrib.andrew.cmu.edu/~dch/ | >+-===========================================================================-+ > try the www.xilinx.com. You can e-mail a question to them from the web page.Article: 5309
Hi, I am looking for PMC ATM PHY products' alternatives, specially for 155M and 622M. Is anyone who knows this?Article: 5310
Hi, I am looking for PMC ATM PHY products' alternatives, specially for 155M and 622M. Is anyone who knows this?Article: 5311
In article <32f96df6.89153336@news.teleport.com>, kenk@teleport.com says... > Most of Lucents sales comes from captive AT&T accounts. ORCA is not a > major player in the general marketplace. Agreed that most of Lucent's FPGAs are sold into direct (big) accounts. Their pricing and distribution scheme "drove" them to this condition. However, in the last year Lucent has taken steps to broaden distribution of their parts, and attract more of the "little fish". Most of this effort is in the form of part pricing to end users and distributors. I don't see any reason why Lucent is any less of an FPGA vendor than Xilinx or Altera, for any size project or company. On the subject of "whose xxx is bigger": If you're going to use *lots* of gates (list of huge Xilinx FPGAs deleted), and use them at a high enough bandwidth, then power can become a limiting factor before you are bound by gates, routing channels, etc. Last time I looked, the Lucent 2cA series had a speed/power advantage of X 4KE, but that may have changed. Both the L and the X company have made much-needed improvements in the thermal characteristics of their packages, but last time I looked the Lucent 2C40 with its cavity-down packaging had a definite advantage in the field. **************************************************************** Bob Elkind mailto:eteam@aracnet.com 7118 SW Lee Road part-time fax number:503.357.9001 Gaston, OR 97119 cell:503.709.1985 home:503.359.4903 ****** Video processing, R&D, ASIC, FPGA design consulting *****Article: 5312
The published work has all been theoretical study (including pricing of the required hardware fabrication). The only entities interested in public construction of such a machine would be non-profit political groups such as the Electronic Privacy Information Center, and theoretical studies serve their purpose just as well. Those with a _real_ use for exhaustive search are served only if the fact that they have such capability is kept private. Do not expect to read about success on Usenet. Larry Kilgallen In article <32F9509A.4C2F@earthlink.net>, Michael Koch <mkoch@earthlink.net> writes: > Wasn't there a proposal years ago which used memory/CPU hybrid chips to > break DES [exhaustive search] within a day for around $100k? But the > machine was scalable. Wonder what happened to that... > > MIKE... > > Steve Casselman wrote: >> >> The guy who broke the 40-bit code made the LA times. >> It took 250 workstations 3 1/2 hours. If he had to search >> to the last combination it would have taken 10 hr. I >> figured that a 50MHz engine would have taken 6 hrs. >> The 48-bit key search (which is next) should take >> 256 times longer. >> >> Steve CasselmanArticle: 5313
I have troubles using the XILINX Synopsys Interface. After doing the "compile -map_effort med" the following error occours: Loading target library 'xprim_4005-5' Loading target library 'xprim_4000-5' Loading target library 'xgen_4000' Loading target library 'xio_4000-5' Loading target library 'xfpga_4000-5' Loading design 'pld1' Information: Design 'pld1' has no optimization constraints set. (OPT-108) Beginning FPGA optimization --------------------------- Beginning Resource Allocation (area only) ----------------------------- Allocating blocks in 'pld1' Allocating blocks in 'pld1' Error: The entity 'inc_dec_ub' depends on the package 'std_logic_arith' which has been analyzed more recently. Please re-analyze the source file for 'inc_dec_ub' and try again. (LBR-2 8) Information: Compile terminated abnormally. (OPT-100) Current design is 'pld1'. 0 Source files are: -------------------------- file pld1_ent.vhd library IEEE; use IEEE.STD_LOGIC_1164.ALL; use IEEE.STD_LOGIC_UNSIGNED."+"; entity pld1 is port( MAINCLK:in STD_LOGIC; RESET:in STD_LOGIC; D: inout STD_LOGIC_VECTOR(7 DOWNTO 0) ); end pld1; ------------------------- file pld1_arc.vhd architecture A_PLD1 of pld1 is begin P1: process(MAINCLK,RESET) begin if RESET='1' then D <= "00000000"; else if MAINCLK'EVENT and MAINCLK='1' then D <= D + "00000001"; end if; end if; end process; end A_PLD1; ---------------------------- Any hints? Has anyone has had the same troubles and solved the problem? If you need the scripts and the setup files please tell me. Ciao, GerhardArticle: 5314
In article <fliptronE4xrqG.n0@netcom.com>, fliptron@netcom.com says... > I am currently using products from Aries, for both the WW socket and a > ZIF that plugs into it. Aries is on the web at http://www.arieselec.com/ > For the WW socket, their partnumber would be (I think) 224-PGM-18-18005-20 > or -21 or -30 or -31. -2x is two level wrap, -3x is three level wrap. > -x0 is Gold collet, tin shell, -x1 is Gold collet, gold shell. > > You can order these products through CalSwitch 1-800-225-7924 ... > A suitable ZIF socket is probably 361-PRS-19-001-1-0 which is a 19 by 19 > array, and will require some modification to match up the WW socket. > Check with CalSwitch. ... > Philip > In article <32F2441F.3A7E@prism.gatech.edu> Scott McIntosh <gtd750a@prism.gatech.edu> writes: > >Hello, > > I'm using a Xilinx 4013E PG223 chip and currently the rest > >of the hardware is to be mounted with wire wrap sockets. Problem > >so far is I'm unable to find an 18x18,223 wire wrap socket. Are > >these just not available? Any other suggestions? > > > >Thanks, > > Scott McIntosh > > gtd750a@prism.gatech.edu **************************************************************** Bob Elkind mailto:eteam@aracnet.com 7118 SW Lee Road part-time fax number:503.357.9001 Gaston, OR 97119 cell:503.709.1985 home:503.359.4903 ****** Video processing, R&D, ASIC, FPGA design consulting *****Article: 5315
In theory, you should be able to put the 16L8 into a PLD programmer, read it (assuming the security fuse is not blown) and then put in a 16V8 and just program it with the same data. When the 16V8 was originally designed, it was done so it is fuse-map compatible with the PAL devices. Mot PC-attached programmers can produce a jedec file from a device. The same jedec file which programmed the 16L8 should produce a 16V8 which functions identically. Now, some programmers (e.g. Data I/O Chiplab) are fussy about accepting jedec files which are shorter (in the # of "fuses") than the device being programmed, and since the ex-16L8 jedec file will define fewer fuses than the 16V8 has, you may need to load the jedec file into a not-so-fussy programmer and re-output it for a 16V8. The fact that one 16V8 (e.g. AMD PALCE16V8, no user electronic signature, so 64 fewer bits) can have fewer fuses than another 16V8 (e.g. Lattice) makes life even more fun... Peter. Return address is invalid to help stop junk mail. E-mail replies to z80@digiserve.com.Article: 5316
Hi, Normally, any programmer can read a PLD as well as write one. However, PLDs have a security fuse, intended exactly to prevent copying them. If the security fuse is blown, you can't read the PLD - you get an empty fuse-array. I suggest you try and hope that whoever had programmed the original PLD wasn't paranoid... Regards Assaf SarfatiArticle: 5317
In article <32f96df6.89153336@news.teleport.com>, Ken Krolikoski <kenk@teleport.com> wrote: >Most of Lucents sales comes from captive AT&T accounts. ORCA is not a >major player in the general marketplace. Ken, I assume that you work for Thorson Pacific, a Xilinx sales representative. As such perhaps you provide some numbers (market share) to support your claims. Eric -- coffin@teleport.COM Public Access User -- Not affiliated with Teleport Public Access UNIX and Internet at (503) 220-1016 (2400-28800, N81)Article: 5318
In article <32f8ce58.179159657@cnn.exu>, Jason.Wright@ebu.ericsson.com (Jason T. Wright) wrote: > > I have experience with the X4010, using older DOS Viewlogic and Xact 4 > & 5, so I have a feel for the process. I haven't used ORCA except to > do a trial evaluation (for the 20 kbit design), and I found that > design impossible to route inOR2C40A: but then, the embedded RAM was > about half of the chip. > > Jason, if you have experience with XC4010, you may be interested in the functional improvements made in the XC4000E family ( and also carried forward into the XC4000EX and XC4000XL families ). The 16 x1, or 16 x 2, or 32 x 1 RAMs can now be configured for synchronous writing by a low-skew Global clock. The signals for address, data-in, and write-enable only have to meet a set-up time requirement with respect to that clock. The CLB can also implement a dual-port memory with synchronous write. These two features make it much easier to design with, and increase the performance significantly. Gone are the days of requiring a double-frequency clock, or implementing a glitch-maker, or worrying about the necessary overlap of address and data over write enable. Now you can pretend that the RAM is just a flip-flop or register. I wrote an app note on using these features to implement fast, efficient and fully asynchronous FIFOs, which was a pretty tough task with the old XC4000 structure. These FIFOs now now run with overlapping asynchronous read and write, up to 60 MHz on the present silicon, even faster in the upcoming faster parts. And the control logic keeps track of FULL and EMPTY through some neat logic tricks. XC4000EX has additional routing resources and faster clocks. XC4000XL is the same design, but optimized for 3.3 V. Technology is really moving ahead, much more rapidly than in the recent past. Of course I am enthusiastic. I work for the company. Peter Alfke, Xilinx ApplicationsArticle: 5319
Michael Koch <mkoch@earthlink.net> writes: > > Wasn't there a proposal years ago which used memory/CPU hybrid chips to > break DES [exhaustive search] within a day for around $100k? But the > machine was scalable. Wonder what happened to that... > I actually know a friend at the Univ. of Michigan who designed a chip for such a system based off of the proposal. I forget what he projected the price for the system to be, but I think it was more than $100k for the performance you quote. I'm pretty sure it used roughly 20 stages in .8um cmos to get to 200mhz per chip. Pretty cool project for a grad level VLSI design class.Article: 5320
Ed Vogel <epv@pcsi.cirrus.com> writes: > > Steve Casselman wrote: > > Well....:) I just read a simular paper where the design was > reconfigurable switches and 2-bit processors. You want to prototype a > FPGA in an FPGA. Sounds very feasible for a small device. > > Thanks Steve, > that sounds like a very reasonable approach. Is this perhaps > a BYU "nanoprocessor" paper? No. :) -- Brad L. Hutchings VOICE: (801) 378-2667 Dept. of Elec. & Computer Eng. FAX: (801) 378-6586 459 Clyde Building EMAIL: hutch@ee.byu.edu Brigham Young University Provo, Utah 84602 http://www.ee.byu.edu/faculty/hutch/hutch.htmlArticle: 5321
I am using Xilinx's Xact Step software 5.2. I ran xmake in a dos session and it gives me the error that it can't find the .wir files created by workview office from Viewlogic. I can see the files and they are in the same directory where I ran xmake. Any ideas? Email me an suggestions. Thanks. -Allan axliu@cory.eecs.berkeley.eduArticle: 5322
On 6 Feb 1997 21:10:22 GMT, axliu@cecil.EECS.Berkeley.EDU (ALLAN LIU) wrote: >I am using Xilinx's Xact Step software 5.2. I ran xmake in a dos session and it gives me the error that it can't find the .wir files created by workview office from Viewlogic. I can see the files and they are in the same directory where I ran xmake. Any ideas? Email me an suggestions. Thanks. > > > >-Allan >axliu@cory.eecs.berkeley.edu My first guess is that you are running xmake from within the .\wir subdirectory--i.e., they should NOT be in the same directory as you run xmake. I believe that their are properties in a file (don't remember which one) that might point to where the files are; Xilinx tech support will be able to answer that. The general tree structure (as I remember it): home\ this is where your xnf, abl, etc. files are; you run xmake, ppr, etc., here home\sch this is where Viewlogic stores the schematic files, format $name.sheet# home\wir this is where Viewlogic stores the wirelist files; same format as for the schematics. There are backup directories, and I don't remember what all else. Jason T. WrightArticle: 5323
I'm assuming that some of your files exceed the DOS naming convention of (8.3). The Xilinx core tools are still DOS base and therefore are limited to the 8.3 rule. ALLAN LIU wrote: > > I am using Xilinx's Xact Step software 5.2. I ran xmake in a dos session and it gives me the error that it can't find the .wir files created by workview office from Viewlogic. I can see the files and they are in the same directory where I ran xmake. Any ideas? Email me an suggestions. Thanks. > > -Allan > axliu@cory.eecs.berkeley.edu -- / 7\'7 Paulo Dutra (paulo@xilinx.com) \ \ ` Xilinx hotline@xilinx.com / / 2100 Logic Drive (800) 255-7778 \_\/.\ San Jose, California 95124-3450 USA (408) 879-6797Article: 5324
Larry Kilgallen wrote: > > The published work has all been theoretical study (including pricing > of the required hardware fabrication). Well, that's why I said proposal... > Those with a _real_ use for exhaustive search are served only if > the fact that they have such capability is kept private. Do not > expect to read about success on Usenet. Yup... so the crime/public/covert/military intelligence agencies around the world would be interested/or have them. I have no data whatsoever that columbian cocaine dealers encrypt their pick-up schedules w/ DES, nor if forgein intel/milit. uses DES for anything. Not sure if the fact that Unix boxes use it is compelling enough to merit multi-million dollar expenditures.... opps... hold it... we're talking government here... make that a yes! Laters, MIKE... > > Larry Kilgallen > > In article <32F9509A.4C2F@earthlink.net>, Michael Koch <mkoch@earthlink.net> writes: > > Wasn't there a proposal years ago which used memory/CPU hybrid chips to > > break DES [exhaustive search] within a day for around $100k? But the > > machine was scalable. Wonder what happened to that... > > > > MIKE...
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