Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch.fpga Subject: Finally, an FPGA tool chain for Linux (Altera Quartus II) X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 10 May 2001 11:27:34 -0700 Message-ID: Lines: 25 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 10 May 2001 11:40:32 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!grolier!dispose.news.demon.net!demon!newsfeed.mathworks.com!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch comp.arch.fpga:6276 Altera has announced a port of Quartus II (including MAX+PLUS II) to Linux! http://www.altera.com/corporate/press_box/releases/pr-linux_quartus.html http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.050701/211270104 At last, one of the programmable logic vendors gets it. They say "Linux has enjoyed dramatic success over the last several years as a platform for a variety of EDA point tools, such as simulation, because of the low cost per compute cycle." An interesting contrast from Xilinx' claims that there is no customer demand for Linux. The availability is a bit confusing; the press release says "Altera [...] today announced plans to port the Quartus (R) II design software to Linux-based environments". Later it says "The Quartus II version 1.0 software is available today", by which they presumably mean a non-Linux version? On 29-Jun-2000, a Cypress employee told me that they were working on Linux support for a future release of their Warp software, which they expected to have developed within a year. It will be interesting to see whether they follow through. Do any other programmable logic vendors support Linux? ###### Message-ID: <3AFAE933.3419276C@prowokulta.org> Date: Thu, 10 May 2001 21:17:07 +0200 From: Kolja Sulimma Reply-To: kolja@prowokulta.org X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.23.52.1 X-Trace: 10 May 2001 21:17:06 +0100, 213.23.52.1 Lines: 53 X-Complaints-To: abuse@arcor-ip.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!news-fra1.dfn.de!newsfeed.arcor-ip.de!news.arcor-ip.de!213.23.52.1 Xref: chonsp.franklin.ch comp.arch.fpga:6246 I am not familiar with the Altera tools, but I am wondering why Xilinx has no ambitions to port their software. It should be quite easy. Coregen for example should work under linux right away. The backend tools probably do not use more than stdio and should run within a couple of hours. The GUI frontend has been copmletely rewritten recently, and with the use of the right GUI toolkit (Java?) multi platform support should have been not much of a problem. FGPA Express would maybe be more complicated, but thats what XST is for, isn't it? On the other hand: Large parts of tools have not even been ported to Windows 95 yet: No long file names, and the backend tools run in a 16-Bit environment. Also: Did anybody try to run Foundation in a NT multi user environment? Write permission for everybody to a couple of files and directories, and tool preferences are kept globaly, not in a user directory, as it is common with almost every UNIX tool. Kolja Sulimma At least Eric Smith wrote: > Altera has announced a port of Quartus II (including MAX+PLUS II) > to Linux! > > http://www.altera.com/corporate/press_box/releases/pr-linux_quartus.html > http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.050701/211270104 > > At last, one of the programmable logic vendors gets it. They say "Linux > has enjoyed dramatic success over the last several years as a platform > for a variety of EDA point tools, such as simulation, because of the low > cost per compute cycle." An interesting contrast from Xilinx' claims > that there is no customer demand for Linux. > > The availability is a bit confusing; the press release says "Altera > [...] today announced plans to port the Quartus (R) II design software > to Linux-based environments". Later it says "The Quartus II version 1.0 > software is available today", by which they presumably mean a non-Linux > version? > > On 29-Jun-2000, a Cypress employee told me that they were working on > Linux support for a future release of their Warp software, which they > expected to have developed within a year. It will be interesting to see > whether they follow through. > > Do any other programmable logic vendors support Linux? ###### Message-ID: <3AFAF39C.2923316D@algor.co.uk> From: Rick Filipkiewicz X-Mailer: Mozilla 4.75 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Organization: Algorithmics Ltd. Cache-Post-Path: mudchute.algor.co.uk!root@oval.algor.co.uk X-Cache: nntpcache 2.4.0b2 (see http://www.nntpcache.org/) Lines: 39 Date: Thu, 10 May 2001 21:01:32 +0100 NNTP-Posting-Host: 62.254.210.251 X-Complaints-To: abuse@ntlworld.com X-Trace: news6-win.server.ntlworld.com 989524896 62.254.210.251 (Thu, 10 May 2001 21:01:36 BST) NNTP-Posting-Date: Thu, 10 May 2001 21:01:36 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!colt.net!newspeer.clara.net!news.clara.net!nntp.news.xara.net!xara.net!gxn.net!news5-gui.server.ntli.net!ntli.net!news6-win.server.ntlworld.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6225 Eric Smith wrote: > > Altera has announced a port of Quartus II (including MAX+PLUS II) > to Linux! > > http://www.altera.com/corporate/press_box/releases/pr-linux_quartus.html > http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.050701/211270104 > > At last, one of the programmable logic vendors gets it. They say "Linux > has enjoyed dramatic success over the last several years as a platform > for a variety of EDA point tools, such as simulation, because of the low > cost per compute cycle." An interesting contrast from Xilinx' claims > that there is no customer demand for Linux. > > I wonder if Xilinx's attitude stems from these 2 (my opinion only) equivalences: (1) Linux user <=> serious user (2) GUI user <=> !serious user From which we can derive (3) Linux user <=> !GUI user Given that VMWare provides a way of using the command line tools under Linux then (3) => Xilinx doesn't have to do a Linux port. Question: Do Quartus & Warp tools run from the command line as well as the Xilinx ones do ? are the command line i/f's as well documented ? ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFAF39C.2923316D@algor.co.uk> X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 10 May 2001 17:00:03 -0700 Message-ID: Lines: 26 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 10 May 2001 17:13:04 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.stanford.edu!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch comp.arch.fpga:6268 Rick Filipkiewicz writes: > Given that VMWare provides a way of using the command line tools under > Linux then > (3) => Xilinx doesn't have to do a Linux port. I use VMware. But running Windows software under VMware is really annoying compared to running native Linux software. VMware supports more than just the command line tools. Since it runs real Windows, it runs the GUI stuff too. However, I'd much prefer to completely remove Windows from my system. Right now the Xilinx software is the ONLY reason I have Windows installed. I haven't looked at Altera's FPGAs in a few years, but if their parts will do what I want, having native tools for Linux may well make me switch. The Xilinx Windows GUI wrapper is nice in some ways, but I'd trade it for command-line-only tools under Linux without any hesitation whatsoever. But if Xilinx ever does decide to support Linux, there are Windows GUI compatability libraries they could use to port the GUI. My personal preference would be to get the command line tools (only) in a first Linux release, rather than holding up a Linux release until the GUI is ported. However, given how hostile to Linux Xilinx seems to be, I'm not holding my breath. ###### Message-ID: <3AFB9333.40EE4A0D@prowokulta.org> Date: Fri, 11 May 2001 09:22:27 +0200 From: Kolja Sulimma Reply-To: kolja@prowokulta.org X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFAF39C.2923316D@algor.co.uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.23.52.101 X-Trace: 11 May 2001 09:22:28 +0100, 213.23.52.101 Lines: 26 X-Complaints-To: abuse@arcor-ip.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!uni-erlangen.de!newsfeed.germany.net!newsfeed.arcor-ip.de!news.arcor-ip.de!213.23.52.101 Xref: chonsp.franklin.ch comp.arch.fpga:6249 > Given that VMWare provides a way of using the command line tools under > Linux then > (3) => Xilinx doesn't have to do a Linux port. But I still have to pay and support for VMWare and Windows running inside VMWare. If you look at the processes running after starting a foundation compile, then you can see that the backend tools seem to be 16 Bit only, so you should be able to run them within DOSEMU. You could also get the freeware version of Alliance (include in SuSE Linux for example). With the help of JBits it should be possible to add support for Xilinx chips to it. But hey, wait, you can do all your designs with JBits. You will use a couple of month with respect to time-to-market, but everything will be hand optimized and fine tuned to get very good results. Kolja Sulimma ###### From: "Wolfgang Loewer" Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Fri, 11 May 2001 09:40:33 +0200 Organization: El Camino GmbH Lines: 55 Message-ID: <9dg569$9sv$00$1@news.t-online.com> References: <3AFAF39C.2923316D@algor.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: news.t-online.com 989566986 00 10143 9zk6SfeTS-HZVf 010511 07:43:06 X-Complaints-To: abuse@t-online.com X-Sender: 310055745130-0001@t-dialin.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2314.1300 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!uni-erlangen.de!news-nue1.dfn.de!newsfeed.r-kom.de!newsfeed00.sul.t-online.de!newsmm00.sul.t-online.com!t-online.de!news.t-online.com!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6256 Quartus II runs from the command line and actually this interface is more powerful than the GUI. It supports TCL and there are several pre-written scripts available from the Altera web site and some more, sophisticated and easy to use ones from FAEs. All Quartus specific TCL commands are well documented.With this inteface you can access the device and timing database and do powerful manipulations like floorplanning or setting timing constraints. The syntax is: quartus_cmd -f For big designs, especially when integrated with other synthesis or simulation tools we prefer to use TCL rather than the GUI. Rick Filipkiewicz wrote in message news:3AFAF39C.2923316D@algor.co.uk... > > > Eric Smith wrote: > > > > Altera has announced a port of Quartus II (including MAX+PLUS II) > > to Linux! > > > > http://www.altera.com/corporate/press_box/releases/pr-linux_quartus.html > > http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.050701/211270104 > > > > At last, one of the programmable logic vendors gets it. They say "Linux > > has enjoyed dramatic success over the last several years as a platform > > for a variety of EDA point tools, such as simulation, because of the low > > cost per compute cycle." An interesting contrast from Xilinx' claims > > that there is no customer demand for Linux. > > > > > > I wonder if Xilinx's attitude stems from these 2 (my opinion only) > equivalences: > > (1) Linux user <=> serious user > > (2) GUI user <=> !serious user > > > From which we can derive > > (3) Linux user <=> !GUI user > > Given that VMWare provides a way of using the command line tools under > Linux then > (3) => Xilinx doesn't have to do a Linux port. > > > Question: Do Quartus & Warp tools run from the command line as well as > the Xilinx ones do ? are the command line i/f's as well documented ? > > ###### From: martin capitanio Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Fri, 11 May 2001 22:23:57 +0200 Organization: J.W.Goethe Universitaet, Frankfurt am Main, Germany Lines: 33 Message-ID: <3AFC4A5D.13A2A58D@capitanio.org> References: <3AFAE933.3419276C@prowokulta.org> NNTP-Posting-Host: nafp2-144.rz.uni-frankfurt.de Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: styx.rbi.informatik.uni-frankfurt.de 989612442 24025 141.2.23.144 (11 May 2001 20:20:42 GMT) X-Complaints-To: usenet@informatik.uni-frankfurt.de NNTP-Posting-Date: 11 May 2001 20:20:42 GMT X-Mailer: Mozilla 4.7 [de]C-CCK-MCD QXW0322n (Win98; I) X-Accept-Language: de,en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!news0.de.colt.net!news-fra1.dfn.de!news.uni-frankfurt.de!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6301 Kolja wrote: > > I am not familiar with the Altera tools, but I am wondering why Xilinx has no > ambitions to port their software. > It should be quite easy. Hmm, maybe good news: seems that xilinx programmers have already taken notice about the existence of linux: ... #endif /***** end SYSWin32 ***********************************************************/ /***** LINUX ******************************************************************/ #ifdef SYSLinux #endif /***** end SYSLinux ***********************************************************/ (An excerpt from the JDrive code, "the world's first IEEE Std 1532 Programming Engine") Perhaps somebody should point out to xilinx managers that "the world's first Open Source FPGA/CPLD Development Suite for Linux" sounds also very sexy and automatically implies "the world's most user friendly FPGA/CPLD Chips". Martin ###### From: Rick Collins Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Fri, 11 May 2001 19:17:44 -0400 Organization: Arius, Inc Lines: 36 Message-ID: <3AFC7318.8F3BF20F@yahoo.com> References: <3AFAF39C.2923316D@algor.co.uk> <3AFB9333.40EE4A0D@prowokulta.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: UmFuZG9tSVaRTTAK4S2cCxHuUP/A6XymEHcTTa5M5ecIl74iO0xXNpnHv+hPJ4nF X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 11 May 2001 23:17:45 GMT X-Mailer: Mozilla 4.73 [en] (Win98; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6292 Kolja Sulimma wrote: > But hey, wait, you can do all your designs with JBits. You will use a couple of > month with respect to > time-to-market, but everything will be hand optimized and fine tuned to get very > good results. > > Kolja Sulimma This is an interesting spin on things. Most companies I have worked for would gladly trade off FPGA utilization to gain a couple months to market. The project I am currently working on costs the company $500,000 in lost sales every month we delay product ship. They focus on making their designs portable between different vendors so that they can use that as a lever to get pricing even lower than a vendor's "best" price. Kinda like the time I was car shopping and the dealer said that was his "best" possible price. When I said I was leaving, he dropped the price a few more hundred, (back in the days when you could get a car for $7000). So they actually waste silicon in the sense of not using much of the "special" features so that they are not locked into a single vendor. So size optimization is not important to some companies since they use a business model that maximizes profits in a different way. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design URL http://www.arius.com 4 King Ave 301-682-7772 Voice Frederick, MD 21701-3110 301-682-7666 FAX ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 11 May 2001 19:07:46 -0700 Message-ID: Lines: 20 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 11 May 2001 19:20:58 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch comp.arch.fpga:6313 martin capitanio writes: > Perhaps somebody should point out to xilinx managers that > "the world's first Open Source FPGA/CPLD Development Suite for Linux" > sounds also very sexy and automatically implies "the world's most > user friendly FPGA/CPLD Chips". It would, but I harbor no fantasies of Xilinx releasing any tools as open source. Even JBits is not open source. I'll be reasonably happy if they just release Linux versions as plain old closed-source. At least JBits will work on Linux, since it's written in Java, although the damn Linux installer that Xilinx used for the latest version won't run properly on my system with Red Hat Linux 7.0 and Sun's latest JDK. AFAIK, all the JBits installer does is extract files from an archive, so it seems like it should be even easier for them to simply supply a .tar.gz file. Sigh. ###### Message-ID: <3AFD0B19.96B8F8BF@prowokulta.org> Date: Sat, 12 May 2001 12:06:17 +0200 From: Kolja Sulimma Reply-To: kolja@prowokulta.org X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.23.52.79 X-Trace: 12 May 2001 12:06:19 +0100, 213.23.52.79 Lines: 11 X-Complaints-To: abuse@arcor-ip.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!newsfeed.arcor-ip.de!news.arcor-ip.de!213.23.52.79 Xref: chonsp.franklin.ch comp.arch.fpga:6306 > At least JBits will work on Linux, since it's written in Java, although > the damn Linux installer that Xilinx used for the latest version won't > run properly on my system with Red Hat Linux 7.0 and Sun's latest JDK. > AFAIK, all the JBits installer does is extract files from an archive, > so it seems like it should be even easier for them to simply supply a > .tar.gz file. Or a JAR file.... Kolja Sulimma ###### Message-ID: <3AFD0C72.3E191093@prowokulta.org> Date: Sat, 12 May 2001 12:12:02 +0200 From: Kolja Sulimma Reply-To: kolja@prowokulta.org X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.23.52.79 X-Trace: 12 May 2001 12:12:04 +0100, 213.23.52.79 Lines: 22 X-Complaints-To: abuse@arcor-ip.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!newsfeed.arcor-ip.de!news.arcor-ip.de!213.23.52.79 Xref: chonsp.franklin.ch comp.arch.fpga:6304 > martin capitanio writes: > > Perhaps somebody should point out to xilinx managers that > > "the world's first Open Source FPGA/CPLD Development Suite for Linux" > > sounds also very sexy and automatically implies "the world's most > > user friendly FPGA/CPLD Chips". > > It would, but I harbor no fantasies of Xilinx releasing any tools as > open source. Even JBits is not open source. I can see that JBits contains trade secrets. I know that Steve Guccione has very serious research interests in JBits generators, but I believe that to most of Xilinx JBits is only a way how they can make academia happy without publishing details about the bitstream. But this is not true for many other tools. How much is there to gain for Xilinx from keeping the hardware debugger source code secret? But there is something to gain from making it open source: Peter Alfke send me a code snippet a while ago and I found a bug within 30 seconds. Kolja Sulimma ###### From: martin capitanio Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sat, 12 May 2001 13:02:48 +0200 Organization: J.W.Goethe Universitaet, Frankfurt am Main, Germany Lines: 15 Message-ID: <9dj51k$t5u$1@styx.rbi.informatik.uni-frankfurt.de> References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> NNTP-Posting-Host: nafp2-040.rz.uni-frankfurt.de Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8Bit X-Trace: styx.rbi.informatik.uni-frankfurt.de 989665140 29886 141.2.23.40 (12 May 2001 10:59:00 GMT) X-Complaints-To: usenet@informatik.uni-frankfurt.de NNTP-Posting-Date: 12 May 2001 10:59:00 GMT User-Agent: KNode/0.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!news-fra1.dfn.de!news.uni-frankfurt.de!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6303 Eric Smith wrote: > martin capitanio writes: > > Perhaps somebody should point out to xilinx managers that > > "the world's first Open Source FPGA/CPLD Development Suite for Linux" > > sounds also very sexy and automatically implies "the world's most > > user friendly FPGA/CPLD Chips". > > It would, but I harbor no fantasies of Xilinx releasing any tools as > open source. Even JBits is not open source. Look at JDrive. It's possibly unusable, but OS (with the power of GPL have they yet some problems). Imho it is the first xilinx's big step in right direction (world domination in managers dictionary;-) Martin ###### From: martin capitanio Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sat, 12 May 2001 13:02:48 +0200 Organization: J.W.Goethe Universitaet, Frankfurt am Main, Germany Lines: 15 Message-ID: <9dj51k$t5u$1@styx.rbi.informatik.uni-frankfurt.de> References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> NNTP-Posting-Host: nafp2-040.rz.uni-frankfurt.de Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8Bit X-Trace: styx.rbi.informatik.uni-frankfurt.de 989665140 29886 141.2.23.40 (12 May 2001 10:59:00 GMT) X-Complaints-To: usenet@informatik.uni-frankfurt.de NNTP-Posting-Date: 12 May 2001 10:59:00 GMT User-Agent: KNode/0.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!news-fra1.dfn.de!news.uni-frankfurt.de!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6303 Eric Smith wrote: > martin capitanio writes: > > Perhaps somebody should point out to xilinx managers that > > "the world's first Open Source FPGA/CPLD Development Suite for Linux" > > sounds also very sexy and automatically implies "the world's most > > user friendly FPGA/CPLD Chips". > > It would, but I harbor no fantasies of Xilinx releasing any tools as > open source. Even JBits is not open source. Look at JDrive. It's possibly unusable, but OS (with the power of GPL have they yet some problems). Imho it is the first xilinx's big step in right direction (world domination in managers dictionary;-) Martin ###### Message-ID: <3AFD4548.60B8B923@home.com> From: Phil Hays Organization: "Real email is phil-hays aATt homeDOTcom" X-Mailer: Mozilla 4.7 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 32 Date: Sat, 12 May 2001 14:16:35 GMT NNTP-Posting-Host: 65.4.188.174 X-Complaints-To: abuse@home.net X-Trace: news1.sttls1.wa.home.com 989676995 65.4.188.174 (Sat, 12 May 2001 07:16:35 PDT) NNTP-Posting-Date: Sat, 12 May 2001 07:16:35 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newshub2.home.com!news.home.com!news1.sttls1.wa.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6320 Eric Smith wrote: > At last, one of the programmable logic vendors gets it. They say "Linux > has enjoyed dramatic success over the last several years as a platform > for a variety of EDA point tools, such as simulation, because of the low > cost per compute cycle." An interesting contrast from Xilinx' claims > that there is no customer demand for Linux. Given a choice between a port to Linux and fixing the "Map puts logic that must go into opposite corners of the die to make timing into one CLB, so PAR can't place it" bug, I would hope that Xilinx never ever ports to Linux. Linux is just another another OS. If Xilinx really wants to port to another OS, I'd rather see Tops20. (-;) (OK, hardware availability is a problem.) Or even VMS. Really, the OS is a tiny part of the design process, as least as I see it. Most of my time doing design is spend reading documentation, running the wetware, drawing cryptic stuff on the whiteboard, filling pages of the notebook, writing documentation, doing ^x(^s^s^y^x) sorts of things in EMACS, doing do compile.do and do sim.do stuff, hitting the Run button, floorplanning one of several different ways and running the make file that makes a bit file. What does the OS really have to do with this process? Ok, once this year the OS did get involved, a process got to zombie status and failed to die until I restarted the OS. Cost me a few minutes of time. The last time I was using UNIX, however, similar stuff was a weekly event. Back when I was using Windows 3.1 this was a daily event. OS is a non issue. The MAP issue, on the other hand, bites. -- Phil Hays ###### Message-ID: <3AFD4B7D.4F012BFF@prowokulta.org> Date: Sat, 12 May 2001 16:41:01 +0200 From: Kolja Sulimma Reply-To: kolja@prowokulta.org X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.23.52.25 X-Trace: 12 May 2001 16:41:05 +0100, 213.23.52.25 Lines: 39 X-Complaints-To: abuse@arcor-ip.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!newsfeed.arcor-ip.de!news.arcor-ip.de!213.23.52.25 Xref: chonsp.franklin.ch comp.arch.fpga:6307 > What does the OS really have to do with this process? A really do not care much about what OS I am working with, but I would really prefer to work in one OS. We are currently running Foundation in a windows NT multi user enviroment. Everybody here is anoyed because the preferences are shared among all users. I of course need some software that will not run under NT, because it uses some wierd driver to talk to the parallel port. So I need a VMWare to be able to securely wun WIN95 Than we have this cadence installation around that will not run with anything but solaris. Bought an extra machine for that. The windows X-Servers suck despite their price, so I boot my PC into Linux when I want to work with the solaris machine. And hey, I am doing more and more embedded stuff, so maybe I need to install QNX as a target platform, too. The cost of maintaining five operating systems is immense. This is not the fault of any of the software vendors, but each of them has the power to solve this problem for me, so I keep asking for it. > OS is a non issue. The MAP issue, on the other hand, bites. I keep asking for that too. My favorite: Read the XC4K datasheet. It says that you can put three arbitrary independant functions (no shared inputs) of 4, 4 and 3 inputs respectively into a single CLB if one or two of these functions are fed to registers. Ever tried this? Well, you can't, unless you use the FPGA-Editor. This is a documented bug from XACT 6.1 that has never been fixed. (I last tried it with Foundation 2.1i) Kolja Sulimma ###### From: Rick Collins Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sat, 12 May 2001 11:09:56 -0400 Organization: Arius, Inc Lines: 52 Message-ID: <3AFD5244.5A05B18E@yahoo.com> References: <3AFD4548.60B8B923@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: UmFuZG9tSVb+yYQP5ouD60t5UOH4VVkNst/9VNp9mC0vPsT41oDfRpexkyNQocXx X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 May 2001 15:10:15 GMT X-Mailer: Mozilla 4.73 [en] (Win98; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6290 Phil Hays wrote: ...snip... > Really, the OS is a tiny part of the design process, as least as I see it. Most > of my time doing design is spend reading documentation, running the wetware, > drawing cryptic stuff on the whiteboard, filling pages of the notebook, writing > documentation, doing ^x(^s^s^y^x) sorts of things in EMACS, doing do compile.do > and do sim.do stuff, hitting the Run button, floorplanning one of several > different ways and running the make file that makes a bit file. What does the > OS really have to do with this process? > > Ok, once this year the OS did get involved, a process got to zombie status and > failed to die until I restarted the OS. Cost me a few minutes of time. The > last time I was using UNIX, however, similar stuff was a weekly event. Back > when I was using Windows 3.1 this was a daily event. > > OS is a non issue. The MAP issue, on the other hand, bites. > > -- > Phil Hays I agree that an OS should have a very small effect on a design process. However, when the quality of an OS is so low that it impacts the usability of a tool, that is making a statement about the OS. I am currently using FPGA design tools under NT and I find that I need to boot the machine once every two or three days just to keep things running correctly. Many people claim that the real problems are in the various apps, which may be true. But I have never seen a Unix system that was taken down on a regular basis just to clean up problems. Just as the proof of the pudding is in the eating, the measure of an OS is in the running (or rebooting). Then again some of the tools are nothing to brag about. I find that Modelsim crashes once every 4 or 5 reloads of a design. Fortuately this is not a huge impact. If it crashed during a simulation rather than when I have ended one run and am starting another, it would be unusable. Still it costs me time and money when I have to restart and setup my simulation again perhaps 20 times a day. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design URL http://www.arius.com 4 King Ave 301-682-7772 Voice Frederick, MD 21701-3110 301-682-7666 FAX ###### From: Duane Clark Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sat, 12 May 2001 09:17:35 -0700 Organization: http://extra.newsguy.com Lines: 40 Message-ID: <3AFD621F.53E63AED@akamail.com> References: <3AFAF39C.2923316D@algor.co.uk> NNTP-Posting-Host: p-836.newsdawg.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12-20 i686) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!news2 Xref: chonsp.franklin.ch comp.arch.fpga:6298 Eric Smith wrote: > > Rick Filipkiewicz writes: > > Given that VMWare provides a way of using the command line tools under > > Linux then > > (3) => Xilinx doesn't have to do a Linux port. > > I use VMware. But running Windows software under VMware is really > annoying compared to running native Linux software. > > VMware supports more than just the command line tools. Since it runs > real Windows, it runs the GUI stuff too. > > However, I'd much prefer to completely remove Windows from my system. > Right now the Xilinx software is the ONLY reason I have Windows installed. > I haven't looked at Altera's FPGAs in a few years, but if their parts will > do what I want, having native tools for Linux may well make me switch. > > The Xilinx Windows GUI wrapper is nice in some ways, but I'd trade it > for command-line-only tools under Linux without any hesitation whatsoever. So (the obligatory question) just out of curiosity, why aren't you running them under wine? All of the Xilinx command line tools work perfectly this way, and have been doing so for more than a year. The one wine bug I am aware of that cropped up, about a half year ago, was fixed within days by the wine developers. Now that is service that you likely won't get from any CAE vendor. I have VMware, but I never run the Xilinx tools under it, because they are so much easier and faster under wine. I can type "ngdbuild -p xcv300e-6-bg432 -sd ../ngofiles top.ngo", and it behaves exactly as if I were under Solaris. I generally use scripts, but my scripts execute the commands exactly as specified in the Xilinx docs. Makefiles work too, and I would assume that TCL would work, though I have not tried it. Yes, getting Xilinx setup under wine is a little more effort than just inserting a CD and clicking setup. But having gone through that a year ago, it was well worth the effort to me. Duane ###### Message-ID: <3AFD6213.6D372313@home.com> From: Phil Hays Organization: "Real email is phil-hays aATt homeDOTcom" X-Mailer: Mozilla 4.7 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <3AFD4B7D.4F012BFF@prowokulta.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 30 Date: Sat, 12 May 2001 16:19:26 GMT NNTP-Posting-Host: 65.4.188.174 X-Complaints-To: abuse@home.net X-Trace: news1.sttls1.wa.home.com 989684366 65.4.188.174 (Sat, 12 May 2001 09:19:26 PDT) NNTP-Posting-Date: Sat, 12 May 2001 09:19:26 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!wn1feed!wn2feed!worldnet.att.net!24.0.0.38!newshub2.rdc1.sfba.home.com!news.home.com!news1.sttls1.wa.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6319 Kolja Sulimma wrote: > We are currently running Foundation in a windows NT multi user enviroment. Everybody > here is anoyed because the > preferences are shared among all users. Is the the part of Foundation that does simulation, aka Aldec aka SusieCAD? I used Aldec stuff under a WindowsNT 3.51 multi-user system about 5 years ago. There is (err.. or was..) a global registry entry set by the Aldec install that told the Aldec tools where to load the preferences from. If you set this as a user registry entry to your personal copy of the (directory or file? I can't recall.) that it was originally pointed to, then your preferences became your own. Had to do both for it to work, otherwise the Aldec stuff wouldn't start. Please note: this was two versions of NT and several versions of Aldec ago, and was before Aldec was OEM'd by Xilinx. Registry edits must be done with great care. Attempt this at your own risk. > Read the XC4K datasheet. It says that you can put three arbitrary independant > functions (no shared inputs) of 4, 4 and 3 inputs respectively into a single CLB if > one or two of these functions are fed to registers. > Ever tried this? > Well, you can't, unless you use the FPGA-Editor. I got a list as well. The map bug is just the current head of the list. -- Phil Hays ###### Message-ID: <3AFD71F1.4EEB022F@home.com> From: Phil Hays Organization: "Real email is phil-hays aATt homeDOTcom" X-Mailer: Mozilla 4.7 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <3AFD5244.5A05B18E@yahoo.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 39 Date: Sat, 12 May 2001 17:27:08 GMT NNTP-Posting-Host: 65.4.188.174 X-Complaints-To: abuse@home.net X-Trace: news1.sttls1.wa.home.com 989688428 65.4.188.174 (Sat, 12 May 2001 10:27:08 PDT) NNTP-Posting-Date: Sat, 12 May 2001 10:27:08 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newshub2.home.com!news.home.com!news1.sttls1.wa.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6321 Rick Collins wrote: > I am currently using FPGA design tools under NT and I find that I need > to boot the machine once every two or three days just to keep things > running correctly. Many people claim that the real problems are in the > various apps, which may be true. But I have never seen a Unix system > that was taken down on a regular basis just to clean up problems. Just > as the proof of the pudding is in the eating, the measure of an OS is in > the running (or rebooting). As I mentioned, my experience with NT 4 and Windows 2000 is rather better. I don't know why, different applications perhaps? Or different hardware perhaps (Dell or Micron vs ?)? I've seen frequent crashes out of a "Joes Clone Shop" machine, and a similar Dell machine had many fewer crashes. Windows 3.1 and follow-ups (Win95, Win98 and WinME) is a different story. A decade ago, I was using Viewlogic Xilinx under Unix, and Viewmumble would coredump about once a day. Between once a week and once a month the OS would follow, or about as often, would create a zombie process that "kill -9", the sure kill, couldn't kill. Attempting to keep using the system with this zombie would cause a system coredump, so we just rebooted. Oh, and the coredumps from Viewmumble would be in a different place every time, so we needed to grep for and delete these to maintain non-zero disk space. Oh, and the license manager would crash, and either we had to hunt down someone with the root password to restart it, or just tell everyone to save and log off so we could reboot. Time between forced reboots was about a week. I'd rather have been using VMS or Tops20. But maybe it was the applications fault? Or the hardware's fault? > Then again some of the tools are nothing to brag about. I find that > Modelsim crashes once every 4 or 5 reloads of a design. Do you reload or restart the simulation? -- Phil Hays ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 12 May 2001 22:37:58 +0200 Organization: My own Private Self Lines: 40 Message-ID: <6u66f6fgm1.fsf@chonsp.franklin.ch> References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> <3AFD0B19.96B8F8BF@prowokulta.org> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 989699879 608 10.0.3.2 (12 May 2001 20:37:59 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 12 May 2001 20:37:59 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6322 Kolja Sulimma writes: > > At least JBits will work on Linux, since it's written in Java, although > > the damn Linux installer that Xilinx used for the latest version won't > > run properly on my system with Red Hat Linux 7.0 and Sun's latest JDK. Will also not run on Slackware 7.0 and JDK 1.2.2.006. This seems to be universal. > > AFAIK, all the JBits installer does is extract files from an archive, Apart from one thing: assuring you have the password. So it seems to be one of these braindead distribution control things. Fits into the same mould as having to order an URL by mail instead of simply downloading it from an public known web site. > > so it seems like it should be even easier for them to simply supply a > > .tar.gz file. I have an theory: some "GUIs'R'Us" designer felt that an .tar.gz is not in line with offering "quality", so it had to have an GUI. At lowest possible cost, so they used this installer. > Or a JAR file.... Actually after the crash I tried to debug it, and looked at the left overs. It seems to unpack an JAR and a few other files out of the executable, then that JAR is unpacked to a load of files with dumb names (numbers, IIRC, this was a month ago), it then renames them, most likely with data from an totally unreadable binary file. Most likely the password is a decryption key for that file. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 12 May 2001 23:01:28 +0200 Organization: My own Private Self Lines: 48 Message-ID: <6u3daaffiv.fsf@chonsp.franklin.ch> References: <3AFD4548.60B8B923@home.com> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 989701288 661 10.0.3.2 (12 May 2001 21:01:28 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 12 May 2001 21:01:28 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6323 Phil Hays writes: > rather see Tops20. (-;) (OK, hardware availability is a problem.) Hardware availability is a problem? Methinks, you do not know, that you can still get TOPS-20 v7.xx together with fitting hardware TOAD-1 [1] (an KL-10 clone with 30bit extended addressing) from XKL (http://www.xkl.com/). It even uses FPGAs for its processor [2]. [1] http://www.xkl.com/xkl/toad-1.gif [2] http://www.io.com/~guccione/HW_list.html From [2]: FPGA Devices: 2 Xilinx XC4010E-3 Contact: Product Information XKL Systems Corp. 8420 154th Avenue NE Redmond, WA 98052 np-info@nospam.xkl.com Notes: This is the CPU for the TOAD-1, a 36-bit computer instruction-set compatible with the Digital Equipment Corporation PDP-10 family of computers (best known as the DECsystem-10 and DECSYSTEM-20, defined by the respective operating systems Tops-10 and Tops-20). The FPGAs are used for all arithmetic, including floating-point and byte-selection operations. (Note: The PDP-10 architecture is word-oriented, with byte size and position within the 36-bit word defined by a pointer word which also contains the word address.) Also FYI, my hobby FPGA project (target XC2S200, tool JBits on Linux, hey!, back on topic!) is an freeware (public domain, not GPL) PDP-10 clone. Initially aimed at KI-10 compatible with TOPS-10, but extension to KL-10 and TOPS-20 possible. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 12 May 2001 23:06:14 +0200 Organization: My own Private Self Lines: 29 Message-ID: <6uy9s2e0qh.fsf@chonsp.franklin.ch> References: <3AFAE933.3419276C@prowokulta.org> <3AFC4A5D.13A2A58D@capitanio.org> <9dj51k$t5u$1@styx.rbi.informatik.uni-frankfurt.de> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 989701574 661 10.0.3.2 (12 May 2001 21:06:14 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 12 May 2001 21:06:14 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6324 martin capitanio writes: > Eric Smith wrote: > > > It would, but I harbor no fantasies of Xilinx releasing any tools as > > open source. Even JBits is not open source. > Look at JDrive. What is that? Where is there info on it? > It's possibly unusable, but OS One can allways investigate it. > (with the power of GPL have they yet some problems). Quite a few people have a problem with PGL. I am a Linux user since over 5 years, founding member of a Linux user group. And I dislike the GPL and do not put my own software projects under it. RMS (inventor of the GPL) calls it an "legal hack", I call it law abuse. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 12 May 2001 16:29:46 -0700 Message-ID: Lines: 6 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 12 May 2001 16:43:08 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news.he.net!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch comp.arch.fpga:6341 Neil Franklin writes: > Methinks, you do not know, that you can still get TOPS-20 v7.xx > together with fitting hardware TOAD-1 [1] (an KL-10 clone with 30bit > extended addressing) from XKL (http://www.xkl.com/). Bzzzt! No longer commercially available. ###### From: Rick Collins Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sat, 12 May 2001 23:01:36 -0400 Organization: Arius, Inc Lines: 31 Message-ID: <3AFDF90F.5762178E@yahoo.com> References: <3AFD4548.60B8B923@home.com> <3AFD5244.5A05B18E@yahoo.com> <3AFD71F1.4EEB022F@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: UmFuZG9tSVbU85CIEAaFofTnpCp3e7TS6X31s5tOR2T5H6lqW9bL/BtSjeZfs9rj X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 13 May 2001 03:02:27 GMT X-Mailer: Mozilla 4.73 [en] (Win98; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6325 Phil Hays wrote: > > Rick Collins wrote: > > Then again some of the tools are nothing to brag about. I find that > > Modelsim crashes once every 4 or 5 reloads of a design. > > Do you reload or restart the simulation? > > -- > Phil Hays So far it only crashes when I am reloading a simulation. I often work in a mode where I find several small bugs one at a time. I fix the problem in my HDL and reload and restart the simulation. This does not require that I exit the tool. But once in several reloads, it will get to the end of complilation and instead of prompting me to input a command, it crashes. Then I have to run Modelsim again and reload the simulation. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design URL http://www.arius.com 4 King Ave 301-682-7772 Voice Frederick, MD 21701-3110 301-682-7666 FAX ###### Message-ID: <3AFE0330.B6026FA4@home.com> From: Phil Hays Organization: "Real email is phil-hays aATt homeDOTcom" X-Mailer: Mozilla 4.7 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 7 Date: Sun, 13 May 2001 03:46:50 GMT NNTP-Posting-Host: 65.4.188.174 X-Complaints-To: abuse@home.net X-Trace: news1.sttls1.wa.home.com 989725610 65.4.188.174 (Sat, 12 May 2001 20:46:50 PDT) NNTP-Posting-Date: Sat, 12 May 2001 20:46:50 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newshub2.home.com!news.home.com!news1.sttls1.wa.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6343 Neil Franklin wrote: Sorry, but XKL is no longer selling the TOAD. -- Phil Hays ###### From: hamish@cloud.net.au Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Newsgroups: comp.arch.fpga References: <3AFAE933.3419276C@prowokulta.org> User-Agent: tin/1.5.8-20010221 ("Blue Water") (UNIX) (Linux/2.4.3 (i586)) Lines: 11 Message-ID: Date: Sun, 13 May 2001 07:10:15 GMT NNTP-Posting-Host: 203.164.64.6 X-Complaints-To: abuse@optushome.com.au X-Trace: news1.eburwd1.vic.optushome.com.au 989737815 203.164.64.6 (Sun, 13 May 2001 17:10:15 EST) NNTP-Posting-Date: Sun, 13 May 2001 17:10:15 EST Organization: @Home Network Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!enews.sgi.com!harbinger.cc.monash.edu.au!newshub1.rdc1.nsw.optushome.com.au!news1.eburwd1.vic.optushome.com.au.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6328 Kolja Sulimma wrote: > On the other hand: Large parts of tools have not even been ported to Windows > 95 yet: > No long file names, and the backend tools run in a 16-Bit environment. Oh? The backend tools work fine with long filenames on NT and 2000. I don't use the front end tools (design manager, etc); Alliance only. Hamish -- Hamish Moffatt VK3SB ###### Message-ID: <3AFE90AE.D0BA307E@algor.co.uk> From: Rick Filipkiewicz X-Mailer: Mozilla 4.75 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <3AFD5244.5A05B18E@yahoo.com> <3AFD71F1.4EEB022F@home.com> <3AFDF90F.5762178E@yahoo.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Organization: Algorithmics Ltd. Cache-Post-Path: mudchute.algor.co.uk!root@oval.algor.co.uk X-Cache: nntpcache 2.4.0b2 (see http://www.nntpcache.org/) Lines: 48 Date: Sun, 13 May 2001 14:48:30 +0100 NNTP-Posting-Host: 62.254.210.251 X-Complaints-To: abuse@ntlworld.com X-Trace: news6-win.server.ntlworld.com 989761714 62.254.210.251 (Sun, 13 May 2001 14:48:34 BST) NNTP-Posting-Date: Sun, 13 May 2001 14:48:34 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!fr.clara.net!heighliner.fr.clara.net!grolier!btnet-peer0!btnet-peer!btnet!nntp.news.xara.net!xara.net!gxn.net!server6.netnews.ja.net!server4.netnews.ja.net!news5-gui.server.ntli.net!ntli.net!news6-win.server.ntlworld.com.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6327 Rick Collins wrote: > > Phil Hays wrote: > > > > Rick Collins wrote: > > > Then again some of the tools are nothing to brag about. I find that > > > Modelsim crashes once every 4 or 5 reloads of a design. > > > > Do you reload or restart the simulation? > > > > -- > > Phil Hays > > So far it only crashes when I am reloading a simulation. I often work in > a mode where I find several small bugs one at a time. I fix the problem > in my HDL and reload and restart the simulation. This does not require > that I exit the tool. But once in several reloads, it will get to the > end of complilation and instead of prompting me to input a command, it > crashes. Then I have to run Modelsim again and reload the simulation. > > -- > > Rick "rickman" Collins > > I've been using ModelSim PE on NT4.0SP4/6 for 2 1/2 years now & have had hardly any problems. The only time it seems to crash is when having failed to load a design because of e.g. a missing model I fix the error & then try to re-load the sim. The biggest problem for the PE version was a memory leak that didn't get fixed till 5.3c or 5.4c & we started with 4.7e. However that being said I run simulation in what might be a different way. I pre-compile all my Verilog source(s) into a set of libraries and then get the simulator to pull in the models from these using `-L' flags. So if I'm changing just a few bits of RTL source or maybe a testbench model or 2 then running the ``reload'' command only has to update these & not recompile the whole lot. I *do* have 1 or 2 gripes but in the whole I think its the best value for money of any EDA tool out there. I only wish I could presuade my colleagues to forgo their annual holidays so I could spend the dosh upgrading to the SE/EE edition under Linux. ###### From: "Dave Feustel" Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sun, 13 May 2001 09:13:40 -0500 Organization: MindSpring Enterprises Lines: 41 Message-ID: <9dm4pp$tk5$1@slb7.atl.mindspring.net> References: <3AFD4548.60B8B923@home.com> <3AFD5244.5A05B18E@yahoo.com> <3AFD71F1.4EEB022F@home.com> <3AFDF90F.5762178E@yahoo.com> <3AFE90AE.D0BA307E@algor.co.uk> NNTP-Posting-Host: 3f.11.ef.02 X-Server-Date: 13 May 2001 14:13:13 GMT X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.dplanet.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news.mindspring.net!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6338 Rick, I'm running Windows 2000. Where can I find info on ModelSim PE? How much does it cost? Does it work with Virtex-E FPGAs? Thanks, Dave Feustel Fort Wayne, Indiana -------------------------------- "Rick Filipkiewicz" wrote in message news:3AFE90AE.D0BA307E@algor.co.uk... [snipped] > I've been using ModelSim PE on NT4.0SP4/6 for 2 1/2 years now & have had > hardly any problems. > The only time it seems to crash is when having failed to load a design > because of e.g. a missing model I fix the error & then try to re-load > the sim. The biggest problem for the PE version was a memory leak that > didn't get fixed till 5.3c or 5.4c & we started with 4.7e. > > However that being said I run simulation in what might be a different > way. I pre-compile all my Verilog source(s) into a set of libraries and > then get the simulator to pull in the models from these using `-L' > flags. > > So if I'm changing just a few bits of RTL source or maybe a testbench > model or 2 then running the ``reload'' command only has to update these > & not recompile the whole lot. > > I *do* have 1 or 2 gripes but in the whole I think its the best value > for money of any EDA tool out there. I only wish I could presuade my > colleagues to forgo their annual holidays so I could spend the dosh > upgrading to the SE/EE edition under Linux. ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 13 May 2001 22:26:40 +0200 Organization: My own Private Self Lines: 18 Message-ID: <6uk83lknb3.fsf@chonsp.franklin.ch> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 989785600 513 10.0.3.2 (13 May 2001 20:26:40 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 13 May 2001 20:26:40 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6346 Eric Smith writes: > Neil Franklin writes: > > Methinks, you do not know, that you can still get TOPS-20 v7.xx > > together with fitting hardware TOAD-1 [1] (an KL-10 clone with 30bit > > extended addressing) from XKL (http://www.xkl.com/). > > Bzzzt! No longer commercially available. Not any more? Bummer. All the more reason for me to get on with my clone then. :-) -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### From: gah@ugcs.caltech.edu (glen herrmannsfeldt) Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 16 May 2001 00:29:23 GMT Organization: California Institute of Technology, Pasadena Lines: 23 Message-ID: <9dshl3$f5s@gap.cco.caltech.edu> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> NNTP-Posting-Host: seniti.ugcs.caltech.edu X-Newsreader: NN version 6.5.0 #1 (NOV) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!canoe.uoregon.edu!logbridge.uoregon.edu!nntp-server.caltech.edu!gah Xref: chonsp.franklin.ch comp.arch.fpga:6416 Neil Franklin writes: >Eric Smith writes: >> Neil Franklin writes: >> > Methinks, you do not know, that you can still get TOPS-20 v7.xx >> > together with fitting hardware TOAD-1 [1] (an KL-10 clone with 30bit >> > extended addressing) from XKL (http://www.xkl.com/). >> >> Bzzzt! No longer commercially available. >Not any more? Bummer. >All the more reason for me to get on with my clone then. :-) There is now a software emulator that runs under most unix systems. The remaining bugs are rapidly being worked out. I forget if they have TOPS-20 running yet. Maybe only TOPS-10. See the pdp10 newsgroup. -- glen ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 16 May 2001 21:05:51 +0200 Organization: My own Private Self Lines: 34 Message-ID: <6ubsotqflc.fsf@chonsp.franklin.ch> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 990039951 2203 10.0.3.2 (16 May 2001 19:05:51 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 16 May 2001 19:05:51 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6440 gah@ugcs.caltech.edu (glen herrmannsfeldt) writes: > Neil Franklin writes: > > >Eric Smith writes: > > >> Bzzzt! No longer commercially available. > > >Not any more? Bummer. > > >All the more reason for me to get on with my clone then. :-) > > There is now a software emulator that runs under most unix > systems. The remaining bugs are rapidly being worked out. Actually multiple of them: TS-10, E-10 and simh2.6 But emulators are not quite the same thing as real hardware. :-) > I forget if they have TOPS-20 running yet. Maybe only TOPS-10. AFAIK theye are all on TOPS-10 7.03, even 7.04 is crashing all of them. > See the pdp10 newsgroup. alt.sys.pdp10, which I read. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### From: Rick Collins Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: Sun, 20 May 2001 02:26:34 -0400 Organization: Arius, Inc Lines: 39 Message-ID: <3B07639A.7EB56AC6@yahoo.com> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: UmFuZG9tSVYJbQdzyKxeDjJ4+Puz5OGmG4qfAGRo9k9c5oqJ/Z8Yf+/Nm94AUUdX X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 20 May 2001 06:26:37 GMT X-Mailer: Mozilla 4.73 [en] (Win98; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6515 Neil Franklin wrote: > > gah@ugcs.caltech.edu (glen herrmannsfeldt) writes: > > > Neil Franklin writes: > > > > >Eric Smith writes: > > > > >> Bzzzt! No longer commercially available. > > > > >Not any more? Bummer. > > > > >All the more reason for me to get on with my clone then. :-) > > > > There is now a software emulator that runs under most unix > > systems. The remaining bugs are rapidly being worked out. > > Actually multiple of them: TS-10, E-10 and simh2.6 > > But emulators are not quite the same thing as real hardware. :-) Depends on the hardware. I would be willing to bet that an emulator on a current desktop will run faster than the original machine from when, the late 70's, early 80's? Even if you clone the hardware in an FPGA, I bet an emulator can keep up unless you do a lot of opimizations such as pipelining, etc which the original machine likely used sparingly. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design URL http://www.arius.com 4 King Ave 301-682-7772 Voice Frederick, MD 21701-3110 301-682-7666 FAX ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 20 May 2001 16:53:18 +0200 Organization: My own Private Self Lines: 56 Message-ID: <6u8zjsvzq9.fsf@chonsp.franklin.ch> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> <3B07639A.7EB56AC6@yahoo.com> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 990370398 611 10.0.3.2 (20 May 2001 14:53:18 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 20 May 2001 14:53:18 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6527 Rick Collins writes: > Neil Franklin wrote: > > > > gah@ugcs.caltech.edu (glen herrmannsfeldt) writes: > > > > > There is now a software emulator that runs under most unix > > > systems. The remaining bugs are rapidly being worked out. > > > > Actually multiple of them: TS-10, E-10 and simh2.6 > > > > But emulators are not quite the same thing as real hardware. :-) > > Depends on the hardware. I would be willing to bet that an emulator on a > current desktop will run faster than the original machine from when, the > late 70's, early 80's? Speed is not the problem. TS-10 emulator on its authors PIII-800 runs roughly twice the speed of an KL-10 (the fastest PDP-10 model). Any faster will be detrimental to the experience. It is more the "look and feel" thing. A program in an window on an PC desktop just does not look or feel like on an VT05 connected to real hardware. Just not primitive enough for that gut feeling. Not to mention the missing blinkenlights (drawing them in an GUI is not exactly the same thing). Light bulbs (no LEDs those days!) just are different to look at. > Even if you clone the hardware in an FPGA, I bet > an emulator can keep up unless you do a lot of opimizations such as My present simple straight forward FPGA design (one clock per memory or register file access plus a bit of waste) is already with 24MHz at the same speed level as TS-10. > pipelining, etc which the original machine likely used sparingly. AFAIK, pipelined instruction fetch was used from the KI-10 onwards, nothing more. A fully pipelined FPGA (think 486 level pipelining) should be over 10 times original speed. Assuming that is the aim. Actually an extended-addressing (30bit instead of 18bit) processor, running at 486 speed, with some form of graphical output, could be interesting for someone wanting to use an PDP-10 OS for daily work, just for the insanity of it. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### Message-ID: <3B077AC8.A6260184@jetnet.ab.ca> From: Ben Franchuk X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18 i586) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> <3B07639A.7EB56AC6@yahoo.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 31 Date: Sun, 20 May 2001 02:05:28 -0600 NNTP-Posting-Host: 207.153.6.62 X-Trace: newsfeed.slurp.net 990372224 207.153.6.62 (Sun, 20 May 2001 10:23:44 CDT) NNTP-Posting-Date: Sun, 20 May 2001 10:23:44 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.slurp.net!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6528 Rick Collins wrote: Depends on the hardware. I would be willing to bet that an emulator on a > current desktop will run faster than the original machine from when, the > late 70's, early 80's? Even if you clone the hardware in an FPGA, I bet > an emulator can keep up unless you do a lot of opimizations such as > pipelining, etc which the original machine likely used sparingly. It also depends on the the size of the emulator too! I am designing a small cpu with a 12 bit byte size. With the FPGA hardware I have the best speed I can get is a 4 Mhz 6809 style memory cycle ( 16 Mhz clock). The FPGA will clock up to 6 Mhz (24Mhz) in this design.A simple loop benchmark of 1000,000 times is 5.75 seconds with the 250 ns memory cycle. On a creaky old P150 running a emulator under linux the bench mark runs at 5 seconds real time. Since the emulator is small I can assume the program fits in the external cache. If the external cache is 25 ns ( 10x faster than my cpu) I can assume the emulator will always run faster than my cpu. FPGA speeds are about the same 74ALSxx or custom chips in the early 80's. ( Read before the 386 :-) ). With what little FPGA designs I have done (Altera/Quicklogic) I have found both chips to be about the same speed. Personally I like the quicklogic FPGA over the Altera FPGA. Since I found getting a PROM burnt for my FPGA board is very pricey if I do PCB for my computer I might pick the Quicklogic chip over the Altera one. -- "We do not inherit our time on this planet from our parents... We borrow it from our children." "Luna family of Octal Computers" http://www.jetnet.ab.ca/users/bfranchuk ###### Message-ID: <3B078A0E.9F193BB1@jetnet.ab.ca> From: Ben Franchuk X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18 i586) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> <3B07639A.7EB56AC6@yahoo.com> <6u8zjsvzq9.fsf@chonsp.franklin.ch> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 48 Date: Sun, 20 May 2001 03:10:38 -0600 NNTP-Posting-Host: 207.153.6.62 X-Trace: newsfeed.slurp.net 990376133 207.153.6.62 (Sun, 20 May 2001 11:28:53 CDT) NNTP-Posting-Date: Sun, 20 May 2001 11:28:53 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.slurp.net!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6529 Neil Franklin wrote: > > Speed is not the problem. TS-10 emulator on its authors PIII-800 runs > roughly twice the speed of an KL-10 (the fastest PDP-10 model). Any > faster will be detrimental to the experience. Look if speed is a problem I have a P150. I am sure for a small fee I will take your PIII and give you my computer so the emulator will run at the right speed. Why a even have a old "Archive" tape cartridge ( 150 meg ) that can give real tape I/O. > Not to mention the missing blinkenlights (drawing them in an GUI is > not exactly the same thing). Light bulbs (no LEDs those days!) just > are different to look at. You can still get the bulbs! > My present simple straight forward FPGA design (one clock per memory > or register file access plus a bit of waste) is already with 24MHz at > the same speed level as TS-10. I still think small anti-fuse FPGA's are the way to go, as for the BIG computer look - it needs to go into a rack! -. With the smaller FPGA's (84 pin plcc's) you can put the cpu on two boards (custom bit slice) and have the front panel logic. > A fully pipelined FPGA (think 486 level pipelining) should be over 10 > times original speed. Assuming that is the aim. The PDP-10 other than the floating point looks to memory bound rather than CPU bound. Pipelining may not help much. > Actually an extended-addressing (30bit instead of 18bit) processor, > running at 486 speed, with some form of graphical output, could be > interesting for someone wanting to use an PDP-10 OS for daily work, > just for the insanity of it. From what little I have seen of the PDP-10 the addressing is still only 18 bit segments (256kw). You would have a worse mess than intel's 8088. I think PDP-10 flavored cpu could be designed with 30+ bit addressing, but the you would have to write you own software! Ben. -- "We do not inherit our time on this planet from our parents... We borrow it from our children." "Luna family of Octal Computers" http://www.jetnet.ab.ca/users/bfranchuk ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) Date: 20 May 2001 23:48:53 +0200 Organization: My own Private Self Lines: 95 Message-ID: <6uzoc7vghm.fsf@chonsp.franklin.ch> References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> <3B07639A.7EB56AC6@yahoo.com> <6u8zjsvzq9.fsf@chonsp.franklin.ch> <3B078A0E.9F193BB1@jetnet.ab.ca> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 990395333 951 10.0.3.2 (20 May 2001 21:48:53 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 20 May 2001 21:48:53 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch comp.arch.fpga:6535 Ben Franchuk writes: > Neil Franklin wrote: > > > > Speed is not the problem. TS-10 emulator on its authors PIII-800 runs > > roughly twice the speed of an KL-10 (the fastest PDP-10 model). Any > > faster will be detrimental to the experience. > > Look if speed is a problem I have a P150. I am sure for a small fee > I will take your PIII and give you my computer so the emulator > will run at the right speed. Ehm, that is the TS-10 authors (Timothy Stark) PIII-800. I have an AMD K6/2-350. It manages to compile my PDP-10 design in its present state in less than 1 minute (many thanks to JBits being fast). BoardScope tends to page a bit heavy while startup because of only 64M RAM. > > Not to mention the missing blinkenlights (drawing them in an GUI is > > not exactly the same thing). Light bulbs (no LEDs those days!) just > > are different to look at. > > You can still get the bulbs! But not drive them from any emulator I know of. And changing that would be a _major_ speed killer. Developing an FPGA is more fun. > > My present simple straight forward FPGA design (one clock per memory > > or register file access plus a bit of waste) is already with 24MHz at > > the same speed level as TS-10. > > I still think small anti-fuse FPGA's are the way to go, Try an competing design! We have 3 emulators. So far there are only 2 FPGA projects (my one-hot state machine and Heinz Wolters microcoded, both using SRAM based FPGAs). Erm, you say here anti-fuse and in your post just before that you use Altera and Quicklogic. All Altera I know are SRAM based and Quicklogic is AFAIK[1] Flash. What chip series are you using? [1] when I was selecting architecture I failled to find an data sheet on their web site, only sales presentations. Dito Gatefield. Not to mention not even getting into Actels site (needs Macromedia Flash plugin) and Lattices site (wants tons of privacy invading questions answered). Altera, Atmel, Lucent and Xilinx gave me their data sheets without problems. From them I selected Xilinx Virtex/Spartan-II. > as for the BIG computer look - it needs to go into a rack! -. I have no space to put one of them. Single room 4.5x3.5m and so. > > A fully pipelined FPGA (think 486 level pipelining) should be over 10 > > times original speed. Assuming that is the aim. > > The PDP-10 other than the floating point looks to memory bound rather than > CPU bound. Pipelining may not help much. I also suspect caching to be worth more, particularle separate I and D caches. OTOH writing accumulators may be doable pipelined. Add to that parallel read and pipelining instruction fetch. > > Actually an extended-addressing (30bit instead of 18bit) processor, > > running at 486 speed, with some form of graphical output, could be > > interesting for someone wanting to use an PDP-10 OS for daily work, > > just for the insanity of it. > > From what little I have seen of the PDP-10 the addressing is > still only 18 bit segments (256kw). P-166, KA-10 and KI-10 models were 18bit address. KL-10 was designed for 30bit but with only somewhere 24..26 implemented. KS-10 was 18, XKL/Toad-1 implemented the full 30bit. Presently I am aiming for 18bit, but I am keeping 30bit open as an option. > mess than intel's 8088. I think PDP-10 flavored cpu could be designed > with 30+ bit addressing, but the you would have to write you own software! TOPS-10 7.someversion ran the kernel in 30bit, TOPS-20 was fully 30bit right into the user level. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### Message-ID: <3B0955B7.AFA89C0E@akamail.com> From: Duane Clark X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12-20 i686) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Also Synplicity) References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 13 Date: Mon, 21 May 2001 10:51:51 -0700 NNTP-Posting-Host: 63.206.136.150 X-Complaints-To: abuse@pacbell.net X-Trace: news.pacbell.net 990467512 63.206.136.150 (Mon, 21 May 2001 10:51:52 PDT) NNTP-Posting-Date: Mon, 21 May 2001 10:51:52 PDT Organization: SBC Internet Services Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!sn-xit-03!supernews.com!cyclone-sf.pbi.net!206.13.28.33!news.pacbell.net.POSTED!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6550 Eric Smith wrote: > > Altera has announced a port of Quartus II (including MAX+PLUS II) > to Linux! > ... > Do any other programmable logic vendors support Linux? Don't know, but here is good news from Synplicity: http://www.synplicity.com/about/pressreleases/SYB-104final.html It looks like Linux is really starting to get some respect. Duane ###### Message-ID: <3B07E5BC.ABAA5834@jetnet.ab.ca> From: Ben Franchuk X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18 i586) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Finally, an FPGA tool chain for Linux (Altera Quartus II) References: <3AFD4548.60B8B923@home.com> <6u3daaffiv.fsf@chonsp.franklin.ch> <6uk83lknb3.fsf@chonsp.franklin.ch> <9dshl3$f5s@gap.cco.caltech.edu> <6ubsotqflc.fsf@chonsp.franklin.ch> <3B07639A.7EB56AC6@yahoo.com> <6u8zjsvzq9.fsf@chonsp.franklin.ch> <3B078A0E.9F193BB1@jetnet.ab.ca> <6uzoc7vghm.fsf@chonsp.franklin.ch> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 69 Date: Sun, 20 May 2001 09:41:49 -0600 NNTP-Posting-Host: 207.153.6.62 X-Trace: newsfeed.slurp.net 990399599 207.153.6.62 (Sun, 20 May 2001 17:59:59 CDT) NNTP-Posting-Date: Sun, 20 May 2001 17:59:59 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-koe1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.slurp.net!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:6552 Neil Franklin wrote: > Ehm, that is the TS-10 authors (Timothy Stark) PIII-800. My offer applies to anybody, who wants a slower computer. > I have an AMD K6/2-350. It manages to compile my PDP-10 design in its > present state in less than 1 minute (many thanks to JBits being fast). > BoardScope tends to page a bit heavy while startup because of only 64M > RAM. That is one thing I don't like about ram based FPGA's and small CPU's. Our 16 bit machine access 64kb of memory but the FPGA rom is 512 KB. > But not drive them from any emulator I know of. And changing that > would be a _major_ speed killer. Developing an FPGA is more fun. That is a simple timer interrupt say every 1/2 a second.But your right the FPGA is more fun. > Try an competing design! We have 3 emulators. So far there are only 2 > FPGA projects (my one-hot state machine and Heinz Wolters microcoded, > both using SRAM based FPGAs). Only after I get the 500 lbs of paper documents. The 10 was a DO ALL processor with all the features. > Erm, you say here anti-fuse and in your post just before that you use > Altera and Quicklogic. All Altera I know are SRAM based and Quicklogic > is AFAIK[1] Flash. What chip series are you using? The Quicklogic data is on CD rom only. ( Free mind you ). I am using a Altera 10K10 ( 84 plcc ) chip but I also have designed for the Quicklogic PASIC 2 series ( Anti-fuse ) QL2007 chip ( 480 logic blocks ). The quick logic block is some simple gates feeding the input of a 4 input multiplexer block followed by a Flip/Flop. One nice feature of Altera is they have a large library of TTL macros useful for converting old TTL designs. > [1] when I was selecting architecture I failled to find an data sheet > on their web site, only sales presentations. Dito Gatefield. Not to > mention not even getting into Actels site (needs Macromedia Flash > plugin) and Lattices site (wants tons of privacy invading questions > answered). Altera, Atmel, Lucent and Xilinx gave me their data sheets > without problems. From them I selected Xilinx Virtex/Spartan-II. At the time I got my FPGA board the Altera chip was the only one that would fit my design and still be a usable chip and have free software. I refuse to use chips that need FANCY PCB manufacturing. > I have no space to put one of them. Single room 4.5x3.5m and so. OK a very thin RACK. ( smile ) > I also suspect caching to be worth more, particularle separate I and D > caches. OTOH writing accumulators may be doable pipelined. Add to > that parallel read and pipelining instruction fetch. Mind you for a PDP-10 FPGA running at the same speed as a real machine the cache module could be a dummy module as main memory would be at cache speeds already. Ben. PS. I still have finish my FPGA 24 bit cpu before I corner the 36 bit market. -- "We do not inherit our time on this planet from our parents... We borrow it from our children." "Luna family of Octal Computers" http://www.jetnet.ab.ca/users/bfranchuk