IRC logs for #openrisc Sunday, 2013-09-01

--- Log opened Sun Sep 01 00:00:36 2013
stekernpoke53282: that's a nice board, I have one, olofk has one and _franck_ too03:01
stekernyou can easily "ignore" the "silicon heater" (as olofk puts it), and it has a nice set of peripherals and a big fpga03:02
stekernI'd go for that, the price is very decent too03:03
stekernjuliusb: (addicted to celebrity) yeah, I've heard that the OpenRISC groupies are the best, but I haven't seen any of them yet, so I just keep on pushing harder03:04
stekernknz: regarding LCD, I have a 4" LCD addon for de0 nano, but it has been discontinued, there's this replacement though:
stekernthis is the one I have:
knzstekern: thanks for the hint08:35
olofkThe Atlys board has a nice price too, but I'm so fed up with Xilinxs toolchain that I would recommend an Altera board11:18
stekernolofk: I agree, and the sockit goes for less money (the retail price is $249 now) too12:48
stekernto bad we don't have an orpsoc port for it yet, but I have said it's on my todo list to do an orpsocv3 port for, and it's close to the top of the list now12:52
stekernonly two items before it12:53
stekernwould be nice if we had de0 nano and sockit orpsocv3 ports by the conference12:54
stekernthe parallella boards are of course pretty nice and affordable too, just a little uncertain lead time on them12:54
olofkYeah, I'll probably order a DE0-nano to make an orpsocv3 port. Seems like the most common board14:27
knzI found this one has a balanced set of features:,719,1185&Prod=NEXYS414:27
olofkHopefully I can reuse much of the stuff that _franck_ is doing for the de114:27
knzfor the de0, is the multi-touch kit the only workable console option at the moment?14:28
olofkIsn't there a UART in the same cable as the blaster?14:37
knzthat would be convenient but I haven't seen it in the specs (will re-check)14:38
stekernno, theres only jtag in that, nios ii designs on that board use the "jtag-uart"15:20
PowermaniacAmadiro: O_o keep finding people from #learnprogramming and ##programming everywhere16:49
PowermaniacOh and hi everyone16:49
PowermaniacA person replied for once, yes!16:50
PowermaniacOkay knz do you have a lot of experience with OpenRISC/OpenCores?16:51
PowermaniacI ask because I'm completely new to this and hoping someone can point me in the right direction.16:52
mschulzeHi Powermaniac!16:52
PowermaniacHey mschulze!16:52
PowermaniacMschulze maybe you can help?16:53
mschulzeI read your questions on the opencores channel. But I don't understand exactly what you are planning to do.16:53
PowermaniacWell I sort of want to make an open hardware pc16:54
mschulzeDo you want to by a complete board that is abled to do all that things?16:54
PowermaniacWell a complete board or parts and put it together myself16:54
mschulzeOr are you happy with an FPGA board?16:54
PowermaniacThat's the thing from what I understand you can get the OpenRISC 1200 processor which is open source. But you can't get a FGPA board that is open source to put it on....16:56
PowermaniacAnd the OpenRisc can also supposedly run Linux...16:56
mschulzeI managed to run linux on a DE0-nano board from terasic and I think SD card and VGA is possible.16:57
stekernthere are open source fpga boards, as in, the pcb layout is open source16:57
PowermaniacStekern, isn't there a chip on a FPGA board that does all the work though?16:57
mschulzeThe Linux port is in the official sources since 3.116:57
stekernall the things you mentioned in #opencores are possible (although we don't have a debian distribution) on an FPGA board16:58
stekernwell, the fpga, naturally16:58
stekernthere aren't really any open source fpgas, and the tools for them are all closed source16:58
mschulzeKeyboard and Mouse should be easy with the USB core, for ethernet you will need a ethernet PHY16:59
stekerngood old ps/2 works too16:59
stekernbut an fpga board will probably not give the performance you are imagining17:00
PowermaniacYeah I was reading there are supposedly no fully open source FPGA boards....17:00
mschulzeThe "chip that does all the work" is the FPGA containing the OpenRISC cpu ;-)17:00
stekern(that's of course just a guess on my part)17:00
PowermaniacWell at maximum I might have it sit as a cjdns node...for a meshnet17:01
PowermaniacBut I need a Linux distribution to be able to setup the system17:01
mschulzeI think there was something about "up to ~150 MHz" for Cyclone III FPGAs17:01
stekernyou can run Linux without a distribution17:02
PowermaniacThe cjdns system I mean17:02
PowermaniacThe Papilio FPGA boards claims to be open source17:04
PowermaniacBut it contains a Xilinx chip17:04
mschulzeWell, I don't know the cjdns system. Are you aiming for high server load?17:04
mschulzeAre you looking for an open source FPGA?17:04
PowermaniacMschulze not really, as currently in Australia I'm one of the only people on Hyperboria the rest are scattered elsewhere.17:05
PowermaniacMschulze essentially I want an open source computer17:05
PowermaniacWell open hardware not just software17:05
PowermaniacAs I sort of also wish to learn from it all17:06
PowermaniacAnd get to understand how all of it works, so eventually I can try to improve it, if my abilities allow17:07
mschulzeI think then you'll need to setup a complete chip design... FPGA chips are closed hardware as far as I know.17:08
PowermaniacSo I have a feeling I don't really understand what I'm talking about yet17:09
PowermaniacAnd I'm wondering can you put say the OpenRISC onto an Arduino type setup?17:09
PowermaniacOr some other setup as Arduino isn't actually entirely open source either17:10
mschulzeYou could have an open specification for a board and for connection for shields. But the FPGA silicon desing would still be closed source.17:12
mschulzeThe OpenRISC processor with all of it's peripherals are open source, you compile it and put the design into the FPGA17:14
PowermaniacSo no way around having an FPGA or not hmm17:14
PowermaniacSo okay where can I actually buy an OpenRISC and peripherals?17:15
mschulzeYou could use an ASIC, but the silicon process and the fitting tools still are closed source.17:16
PowermaniacAhh okay17:17
mschulzeYou don't by an OpenRISC. It's an open hardware design.17:18
mschulzeDownload the toolchain and the OpenRISC sources, you'll need to do some programming, and then you compile it and run it on an FPGA board.17:19
PowermaniacWait so it's not an actual chip?17:19
mschulzeThat's the big picture.17:19
PowermaniacOkay I've really confused myself here17:19
mschulzeNo, it's not an actual chip :-)17:20
PowermaniacWait does that mean there areen't any design sheets that give you schematics to manufacturer one?17:20
mschulzeThe FPGA is the chip. You load the OpenRISC design into the FPGA and it works as OpenRISC.17:20
PowermaniacLike is there enough information to manufacturer a chip from that information?17:21
mschulzeOr you give the design to an ASIC manufracturer that will make you some hundred thousands and you have your "OpenRISC chip". But that's a bit expensive17:22
PowermaniacOkay so now I'm wondering do any actual processor chips exist that are open source, as now I'm kind of puzzled.17:22
mschulzeThe ASIC production still isn't a standard silicon process but rather a mask programming of the ASIC chip.17:27
PowermaniacSo you can't go to say an ASIC manufacturer and say I want a chip made like this and give them the schematics etc.?17:32
PowermaniacWait what is the OpenSPARC then?17:34
mschulzeI don't know the OpenSPARC, sorry.17:39
stekernyou can go to an ASIC manufacturer with the sources to openrisc and tell them to make a chip out of it (overly simplified, but true)17:40
PowermaniacStekern, I'm assuming the catch is it will cost me an arm and a leg right?17:40
hnomschulze, OpenSPARC is the SPARC platform. Sun Microsystems released it under a open license before Oracle bought them.17:40
stekernPowermaniac: I don't think your arms and legs are worth enough to do it ;)17:41
PowermaniacSeems like I'm up shit creek without a paddle...17:42
stekernbut to get an open source FPGA, you'd need to do an ASIC of that too, so you're back to square one17:42
hnoOpenSPARC is a little bigger design than or1200.17:42
stekernbut once in a while people come in with the same "problem", for you, what is the value of the actual chip making process being open?17:43
mschulzeI would say: go, get an FPGA board of your choice that is already supported by the ORPSoCv2 and give it a try. You will learn a lot about open source hardware then.17:43
stekernI'm not saying there wouldn't be value in it, but the key here is for *you*17:44
mschulzeDoing a silicon desing by yourself is a really big thing.17:45
PowermaniacWell I wanted to have the option to read about how it works, what's in it etc. And then one day (eventually Uni will happen) I might be able to improve thee designs17:45
stekernyou still can do that17:47
stekernthe CPU design is open source17:49
PowermaniacHuh? Are you suggesting FPGA or ASIC or?17:49
stekernwe already established that you doing an ASIC wasn't feasible, so FPGA17:50
PowermaniacOriginally this also seemed like a cheap venture to have some fun with before getting into it further. Now I have a feeling it won't be so cheap...17:50
stekernFPGA boards aren't very expensive17:50
PowermaniacWas looking at the Olimex boards, and RaspberryPi and Pandaboard etc.17:51
stekernyes, but those do *not* feature open source cpu designs ;)17:51
PowermaniacAlthough Olimex claims otherwise17:52
stekernno, I believe they claim the PCB designs are open, which they probably are17:53
stekernI mean, if you're going to be religious about it then you'd need to open source the process of making plastic too...17:54
PowermaniacCan't you technically find how anyone makes plastic easily on the net though?17:56
PowermaniacRecommend me a FPGA board please17:57
PowermaniacAs I think that is the closest I'm ever going to get without spending a fortune.17:58
mschulzeTry the DE0-nano, get it from farnell but not directly from Terasic.17:58
mschulzeYou will have a lot of fun with it :-)17:58
mschulzeIt has 32MB of ram but you will need to build a additional board for SD card and ethernet phy.17:59
mschulzeIf you got some more money to spend, then look for the DE2-11518:00
PowermaniacOh okay18:00
mschulzeThat has everything you are looking for, including ethernet and SD card slot.18:01
hnoThe SoCKit discussed earlier might also be interesting. But likely needs a bit of effort to get orpsoc running on it when it becomes available.18:09
PowermaniacUno: ?18:13
PowermaniacHno: SoCKit?18:13
mschulzeThat board needs to do a journey from taiwan into my hands...18:23
mschulzeThe terasic site makes me beleaving it's available...18:30
mschulzeOh no, preorder... :-(18:31
PowermaniacWell I'm finally going to leave you all alove18:32
PowermaniacThanks all for the help18:32
hnomschulze, Arrow says it is in stock.
mschulzeI'm thinking of a linux running on the arm core loading an linux image onto a multicore openrisc processor in the fpga side... :-D my brain is running hot now. That board looks awesome.18:40
mschulzeIt's linux all the way down.18:41
hnoyou probably want some daughterboards as well to make some interesting I/O.  Those  HSMC  connectors is not exactly something you plug wires to..18:46
stekernmschulze: I have had similar thoughts, using the arm cores to cross compile stuff for the openrisc core etc18:47
stekernI have one of those on my table here, just waiting for me to show it some love...18:48
mschulzeif we could get quartus running on the arm core we could dynamically adopt the fpga design :-D18:48
mschulzestekern will you be there at the ORCONF this year?18:49
stekernyes, that's the "showstopper" to get a full development machine on the devboard18:49
hnoyou could probaly use the ARM as your desktop even.. with only a little help from the FPGA side to provide a framebuffer.18:49
stekerneven cooler would of course be to have the altera tools running on openrisc... ;)18:50
stekernmschulze: I'll be there, yes18:50
hnoif you accept to temporarily loose your monitor output each time you reprogram the fpga.18:50
mschulzethen could you please bring that board with you? i really would like to take a look on it, because we are looking for a arm+fpga board for lab work at my university18:52
stekernsure, I'll bring it, I'm also hoping I have a orpsocv3 port for it by then18:53
stekernor at least running openrisc on it in some way18:53
mschulzeYou should be abled to see my smile...18:54
mschulzeOne of my goals is an dual core architecture on my DE0-nano, one core running linux and the other core running some real time stuff, both together bringing my quadcopter into the air.18:57
mschulzeBut I fear that the board cannot hold two OpenRISC cores.18:57
hansfbaiermschulze: It might work19:15
hansfbaierI run linux on an EP4CE1019:16
hansfbaierthat has half the LUTs of the DE0_nano19:16
mschulzeYes, but I'll need some space for the SD card controller and a connection to a WLAN card.19:17
hansfbaiermschulze: Yes the sd controller eats quite a bit of space. Do you run the sd controller from orpsoc? I get errors When I use it19:19
hansfbaiermschulze: How did you get it working?19:19
hansfbaiermschulze: I use sd with SPI. Very economical. But terribly slow19:20
mschulzeErr... I still have to get it working...19:20
mschulzeAt the moment I am doing a lot of work on another level, you will get more mails on it at the mailing lists this week.19:21
mschulzeHow do you manage the copyright statements in your code files? There are some essential things that we don't learn on university...19:43
mschulzeDo I have to change the year each year in each file?19:45
stekernmschulze: that's a good question, I usually just change it on files I change20:21
olofkI think that's the standard procedure. Update the copyright year when you do changes to a file20:30
mschulzeOkay, thanks. Thats a bit funny, we learn how to make software, but we don't learn how to publish it correctly in the real world...20:41
mor1kx[mor1kx] skristiansson pushed 2 new commits to master:
mor1kxmor1kx/master fc65949 Stefan Kristiansson: mor1kx-defines: whitespace cleanup20:43
mor1kxmor1kx/master b5ca2ea Stefan Kristiansson: mor1kx v1.020:43
stekernolofk: ^20:43
mschulzemor1kx just joyned to tell us that there are new commits?20:46
stekernyes, but that wasn't the significance of it20:47
olofkstekern: Now that deserves a few beers and a tweet!20:48
olofkAnd gives me a reason to add support for getting tagged releases in orpsocv320:49
mschulzebecause we have version 1.0?20:49
stekernI've already enjoyed a beer, perhaps I'll have another one just for the occasion20:49
stekernmschulze: yes, we are notoriously bad at making releases and olofk have been bugging us (and I have been bugging juliusb) that we should make a mor1kx release20:50
mschulzeI thought this is about free software and now you tell me that it is payed in beer :-) Congrats!20:50
olofkmschulze: It's not alcohol free software, if you thought that :)20:52
stekernso, we decided to not make so much fuss about it and just put the v1.0 tag on what we have now20:52
stekernjuliusb: since I got to do the tagging, will you do the honours of a small ml announcement?20:53
stekernthere's your chance to steal back some of the celebrity (and perhaps the groupies) =)20:54
olofkAh crap. I forgot to send out on the ml that orpsocv3 was released20:54
mschulzeorpsocv3 is released? but v2 is not totally obsolete now, is it?20:55
olofkmschulze: Not at all. orpsocv3 s not on feature parity with orpsocv2 yet, and there aren't any board ports at all right now20:56
olofkBut it works well enough to slap a release tag on it20:56
mschulzeOkay, that means I have to migrate my DSL to v3 very soon. It's getting overwhelming now.20:59
olofkmschulze: What does your DSL do? I probably haven't understood really what's it about21:00
mschulzeIt can generate the verilog files for the wishbone bus and toplevel entity and the tcl scripts for quartus at the moment.21:02
olofkThat sounds like an almost perfect match for orpsocv3. Top-level generation and things like TCL scripts are two tasks that I have left for other tools to take care of21:03
olofkWhat do you mean with verilog files for the wishbone bus?21:03
olofkLike arbiters and things like that?21:03
mschulzeYes, the arbiters.21:04
olofkI'm almost done with my interconnect generator. We should definitely talk about how to integrate these two efforts at orconf21:04
mschulzeI don't like to make changes in four or more files just for adding a single IP core21:05
stekernI'm very interested in the top-level generation21:05
olofkPeople generally are interested in automating top-level files, so I felt that I didn't want to make another half-assed attempt to do it myself. Looks like I was right :)21:06
mschulzeAnd it is said that it will be possible to make a graphical editor for Eclipse. Now think of an integrated toolchain and we are done...21:06
stekernfrom a user perspective, that's perfect21:06
stekernbut I assume that it's possible to drive it from command line as well?21:06
mschulzeyes for the generation part.21:07
mschulzethat war one reason why I have chosen Xtext because I know that people need to automate thair builds21:08
mschulze(need some IRC whiteout, sorry for bad spelling)21:10
stekernyou probably won't beat me in that department, so no worries21:10
stekernbut yeah, automated makefile based builds are what makes us nerds happy ;)21:12
mschulzeyes I do. and I'm geting nervous, you will have a talk from me on ORCONF21:12
_franck_mschulze: add a debugger to that and you'll have a perfect environnement21:13
_franck_you can easily integrate eclipse with openocd21:13
mschulze_frank_: that is my goal for the next years.21:13
stekernyeah, I really like the thought of that kind of unified IDE21:14
_franck_we should also try to compile our toolchain under cygwin21:14
stekernI'm perhaps not seeing myself using the guiy things, but it's a huge leap for users21:14
mschulzeI learned the Nios 2 processor at university but was not happy with that expensive thing...21:15
_franck_if we have the nios2 EDS equivalent for openrisc we'll get a lot more interest21:15
stekernI agree21:15
olofkMe too. We need more eye candy to make people interested21:16
mschulzeYou cannot teach SoPC basics with complete textmode tools21:17
_franck_I've always thought of a OpenQSYS :)21:17
olofkOne of the reasons that I scrapped my initial makefile-based orpsocv3 was that it would be a pain to do some GUI interface for that21:18
olofkNow that was a redundant interface21:18
mschulzeIntegrating a good makefile build into eclipse is not that hard.21:19
mschulzeIn my oppinion it could lead to one IDE where you can do everything from hardware design over the C/C++ project up to linux configuration.21:21
olofkBut for orpsoc it probably would be hard as it's not only about building stuff21:22
olofkorpsocv3 is quite differet from orpsocv2 in that regard21:22
mschulzeOkay, I will do what I can to look at v3 so that I can get an idea how I can integrate into in future21:23
mschulzeI have to leave now but sure be back in the next days. I'll have my final exam this week and than heading for a first release on github.21:26
mschulzeGood night :-)21:28
stekerntoo slow ;)21:32
_franck_openrisc and jtag_vpi are now pushed to gerrit (openocd), just have to wait (for a long time....) now21:33
olofk_franck_: That's great. Do you have a URL where we can see the progress?21:34
stekernI have the fun task of cleaning up my store buffer work in front of me21:34
stekernthe backside of my messy development model21:35
stekernso that's the second on my list before sockit on orpscov3, the first was to do the mor1kx release tag21:37
_franck_stekern: great to here you'll do the job for us on sockit ;)21:37
stekernthat the Linux spi driver rework just swooped in and took the first place for a while was all hansfbaiers fault =)21:39
olofk_franck_: Thanks.21:39
stekern_franck_: is that tracking patches on the ml or how do you "push to gerrit"?21:40
_franck_I "git push review" to gerrit21:41
stekernah, I see21:41
stekernI'm completely lost when it comes to gerrit ;)21:42
olofkstekern: Yeah, that's an old git trick. We git-pros use it all the time21:42
stekernhaha, you take all the chances you can, don't you ;)21:43
olofkI'm trying so hard :(21:43
_franck_me too to be honest, I just followed this:
stekernbut I believe that is completely unrelated to git actually21:43
_franck_I just simple commands and when it's too complicate to handle, I just pull somewhere else and redo my changes :)21:44
stekernyou just push it to a server that is linked to gerrit21:45
--- Log closed Mon Sep 02 00:00:38 2013

Generated by 2.15.2 by Marius Gedminas - find it at!