IRC logs for #openrisc Wednesday, 2016-04-20

--- Log opened Wed Apr 20 00:00:29 2016
-!- SMD is now known as Guest672900:19
-!- blueCmd_ is now known as blueCmd02:59
wallentomithro: no03:29
wallentoI am not sure why03:29
_franck__mithro, wallento : I think we should run gdb regression tests to get some numbers then send patches upstream. I think jeremybennett can help to get some approval or at least support04:03
mithro_franck__: hrm?04:03
_franck__we should test gdb with its own test suite. Then if failure numbers are too high, fix things04:05
_franck__then send patches to get or1k upstream04:05
mithro_franck__: SGTM04:06
_franck__I think wallento was volunteer for that :)04:07
wallento_franck__: lol, not really ;)04:52
shornewallento: _franck__: I am thinking of rebuilding my toolchain, ill check it out when I get to gdb05:19
_franck__cool thanks05:19
shorneI was getting weird behavior debugging gdb with the remote mor1kx target, so I did want to test gdb to see if it actually is all sane. The regression tests sound good05:20
shornedebugging through gdb, not debugging gdb itself ;)05:20
olofkHow come no one has implemented or1k support?
_franck__do we have a date for orconf 2016 (or whatever you call it) ?11:47
olofk_franck_: 7-9 October in Bologna12:03
andrzejrshorne, last time I have seen such crashes they were caused by (my) memory controller errors.14:23
_franck_olofk: thanks15:43
Andy___hi there17:00
Andy___is this place alive?17:01
olofkAndy___: From time to time17:01
Andy___I was wondering how active the OpenRISC project still is17:02
olofkNot that much activity nowadays, but there are still new hw and sw features added and support for more hw targets17:09
Andy___thanks for the info olofk, is it because of other open architectures or loss of interest by the community?17:11
olofkRISC-V has been getting a lot of attention last year17:12
olofkAnd many people seem to enjoy building things from scratch instead of maintaining existing stuff, so there has been a lot of basic stuff rewritten for RISC-V17:13
olofkLike toolchains, OS support, a ton of new UARTs and SPI controllers etc17:14
Andy___could it be that the strenght of OpenCores itself went down and pushed developers away from its projects?17:15
Andy___also the more successful ones like OpenRISC?17:15
olofkOpenCores has done more to damage the community than help it the last years17:15
Andy___yes, I also felt that a bit17:16
Andy___do you know who is still maintaining the site?17:16
olofkYes, I do17:16
Andy___is there a way to get in contect with someone?17:17
Andy___* contact17:17
olofkBut we started FOSSi Foundation last year to have a independent body that could work to improve Open Source Silicon17:18
olofkAnd one of our main projects is a new site called librecores.org17:18
olofkWe have talked to the OpenCores owners, but we had different goals, so we will not pursue that any further17:19
Andy___I am browsing about librecores and fossi17:20
Andy___i didn't know they existed17:21
olofkWe are still trying to reach out to people17:22
olofkAnd there are several projects in the works, so if you are interested, make sure to check in regularly, or follow us on twitter, join our mailing lists or subscribe to our rss feed to get more info17:24
olofkOr ask here :)17:24
olofkOr in #fossi or #librecores :)17:25
mafmthere's also the sh4 folks, toolchain and other things seem to be in a good shape, I haven't followed the news from the group17:25
olofkmafm: You're talking about the j-core guys?17:25
Andy___I see you're in contect with people at CERN also!17:26
mafmthere's still a Debian port going on, which comes from many years, toolchain and everything should be working17:26
olofkI haven't really got what they're doing. On their page they say that they've talked to the OpenRISC people, but except for a mail that I sent about two years ago, no one has heard anything from them17:26
olofkAndy___: Yes indeed. Our yearly conference, orconf, was held at CERN last year, and we have been talking a lot to several of the people there over the years17:27
olofkmafm: A sh4 toolchain?17:27
mafmI think that sh4 patents are expiring, so if sh4 is still good design by todays standards (?! I wouldn't know), the ecosystem is there17:28
mafmRISC-V is more of an academic attempt to get things right by the things that are known by 201x, which happened to gain some industry interest17:28
olofkYeah, that is a good thing. A CPU is nothing without software and IP cores to connect it to17:28
mafmthat's the V in the name, I guess :D  (incidentally, I think that they equate SH4 with RISC-III or RISC-IV)17:29
olofkFrom what I have heard, RISC-V seem to be pretty well balanced, but I couldn't tell. Despite being on the OpenRISC project for six years, I don't know much about CPU design :)17:29
mafmI was somehow hoping that RISC-V did progress / settle a bit quicker, the toolchain and everything is not upstreamed17:30
olofkYeah, I'm actually surprised by that, given how extremely much interest they have had for some time17:30
mafmI have the impression that risc-v or sh4 is more for people wanting boards / computers with that arch, while openrisc attracts more people wanting to experiment with cpu designs17:31
olofkI'm a bit afraid that the novelty might wear off. Don't want to end up with another semi supported arch17:31
olofkmafm: Actually, I think RISC-V is for people interested in CPU designs. I counted 11(!) different implementations that were presented last year at orconf17:32
olofkEverything from supersmall to big superscalar OoO versions17:32
mafmI'm specially a bit frustrated with how long it's taking for them to upstream things and settle on the supervisor/priviledge ISA, it was supposed to be ready 1 year ago :-/17:33
mafmolofk: yeah, perhaps that's true at one level (e.g. research projects)... I guess that the industry interest and other individuals is more about "having a CPU that can be used in boards / IoT / huge datacentres"17:34
olofkYeah, that is one of the most common critiques people have. There is very little insight in the decision-making17:34
olofkI agree. Most people would probably be fine with a standard ARM-like RISC-V core with the right peripherals17:35
olofkMy interest is more in the ecosystem around the CPU than the CPU itself17:35
olofkBut now it's time to sleep. Good night17:37
mafmolofk: nite :)17:37
olofkOr good whatever part of day in your local timezone17:37
mafmbritish summer (hah!) time17:38
olofkI hate timezones. Worst idea ever17:38
Andy___good night mafm17:38
mafmAndy___: it's olofk the one going away17:38
Andy___good night olofk17:39
Andy___thanks for all the info!17:39
mafmI'm mostly interested in anything that can be used in a computer or a small device, free to implement and so on17:39
mafmI think that this is going to become more and more important in the future, with backdoors and whatnot17:40
mafmI'd be happy with openrisc, but probably nobody is producing silicon that can be purchased.  I don't have a good use for openrisc on FPGAs, other than learning about design and stuff17:41
mafmso I had some hopes about risc-v/lowrisc producing something, but it's moving at a glacial speed sometimes, and actually unknown if it's moving or not in others17:42
mafmand as olof said, everything quite secretive / hard to collaborate from outside17:42
Andy___goodbye everyone18:14
-!- SMD is now known as Guest7405923:39
--- Log closed Thu Apr 21 00:00:31 2016

Generated by 2.15.2 by Marius Gedminas - find it at!