| --- Log opened Tue Nov 04 00:00:13 2014 | ||
| poke53282 | stekern: It is indeed the snoop hit in http://git.openrisc.net/cgit.cgi/stefan/linux/tree/arch/openrisc/include/asm/spinlock.h?h=smp#n83 | 01:32 |
|---|---|---|
| poke53282 | At least for two cores and switching the core after every instruction. | 01:33 |
| poke53282 | You know, when I first saw this snoop hit in the doc, I thought that I don't need it. | 01:34 |
| poke53282 | And I wondered how this hit should behave if you access the 32 Bit word with an non 32 bit aligned 16 Bit write. | 01:35 |
| poke53282 | I thought this would be undefined and no one would ever implement it this way. Well, I was wrong :) | 01:36 |
| poke53282 | The first guy who comes implemented the locks using this behavior. | 01:38 |
| poke53282 | Do you use any other snoop hits like this? | 01:39 |
| mithro | anyone here going to Linux.conf.au? | 02:24 |
| stekern | poke53282: to be fair, I stole that particular piece of code from: http://git.openrisc.net/cgit.cgi/stefan/linux/tree/arch/arm/include/asm/spinlock.h?h=smp#n109 | 04:27 |
| stekern | mithro: I'm tempted, mostly considering how the weather start to look like here in Finland at the moment | 04:31 |
| poke53282 | stekern: A solution which is based solely on the atomic instructions would be preferable. So, increasing the size of the spinlock to 2 32 Bit words would do it for example. | 04:39 |
| poke53282 | Or implementing the restore with lwa and swa | 04:40 |
| poke53282 | If the arm implementation really works the same way then it is a shame, that this is not documented properly. | 04:42 |
| stekern | I don't understand the problem, I think it's just using a natural property of how things are supposed to work? (well, I understand that it's problematic for you when you write the emulator, but apart from that ;)) | 04:44 |
| poke53282 | If this behavior would be solely part of the l.swa instruction I would be fine. But at the moment it looks like that every write to the memory has to be checked against n variables, with n the number of cores. This is a huge waste of energy just to win one or two cycles in the atomicity operations. | 04:49 |
| poke53282 | By the way. 16 cores still don't work. So I still haven't figured out the main bug :) | 04:52 |
| stekern | well, to be honest, I don't know if other code make use of this property as well. But that particular piece of code or not, you'll have to check all writes from other codes, that's kind of in the definition of an atomic rmw. | 04:54 |
| stekern | s/codes/cores | 04:54 |
| poke53282 | I will think about it and check the speed penalty. Maybe I can get rid of this check with a few tricks. But first I need a stable emulation. | 05:04 |
| poke53282 | I don't like the definition, but it might be necessary for security. Who knows. What happens for a DMA access? | 05:06 |
| stekern | it's not necessarily that dma accesses cause snoop hits, but implementations might | 05:27 |
| stekern | but that's more related to cache coherency, not atomicity | 05:28 |
| -!- Netsplit *.net <-> *.split quits: zama | 06:19 | |
| -!- Netsplit over, joins: zama | 06:20 | |
| mor1kx | [mor1kx] bandvig pushed 1 new commit to withfpu: https://github.com/openrisc/mor1kx/commit/602c45fcabc5c691803e74491eb1da06c23f1605 | 07:55 |
| mor1kx | mor1kx/withfpu 602c45f Andrey Bacherov: Continue re-factoring for pipelined FPU: multiplier. Plus cosmetics for add/sub. | 07:55 |
| wallento1 | stekern: can I find libpthread or1k-sysdeps for uClibc anywhere? | 09:58 |
| stekern | wallento1: there's only support for the old Linux threads in the openrisc port of uClibc | 10:03 |
| wallento1 | okay, thanks. what is the problem with the new? | 10:04 |
| stekern | no other problem than that no-one done the work with NPTL in uClibc | 10:05 |
| wallento | okay, I will ask my student then | 10:05 |
| stekern | personally, I shifted interest towards musl (which have a good pthreads implementation) | 10:05 |
| wallento | yes, we might do the same, but the fiasco OS somehow depends on uClibc | 10:07 |
| wallento | need to have a closer look | 10:07 |
| stekern | blueCmd_ has done the necessary NPTL work in glibc, so if you have someone doing the work, peeking at glibc first might be a good idea (and ask advice from blueCmd_ might be another good idea ;)) | 10:07 |
| wallento | thanks! I found the nptl there, we will try to integrate this one | 10:12 |
| olofk | I also want students to do stuff for me | 11:22 |
| olofk | btw stekern, I'm almost done with wb_sdram_ctrl now, and it's currently 8-14% faster now | 11:22 |
| olofk | (with my benchmark) | 11:22 |
| olofk | Are there any real benchmarks that are affected by memory speeds? Would be interesting to see that | 11:23 |
| stekern | olofk: you can try the membench I wrote | 11:40 |
| stekern | it's very naive, but gives some indication | 11:40 |
| olofk | bare-metal or linux? | 11:41 |
| stekern | http://git.chokladfabriken.org/?p=membenchmark.git;a=summary | 11:42 |
| olofk | stekern: It's too bloated | 11:50 |
| stekern | I agree, the memory usage is way over the top | 11:55 |
| olofk | lol | 11:55 |
| olofk | stekern: https://github.com/olofk/wb_sdram_ctrl | 12:43 |
| olofk | Works, but need some cleanup | 12:43 |
| olofk | How do people name git tags to avoid getting them confused with branches? Just v1.0 ? | 12:47 |
| stekern | olofk: why didn't you build upon the latest version of wb_sdram_ctrl? | 13:18 |
| stekern | https://github.com/olofk/wb_sdram_ctrl/commit/a1c877503c70570eb069d52879d3251876452da9 is already there | 13:19 |
| stekern | (but by franck) | 13:19 |
| stekern | I'll need to get my de0 nano back and give it a try, I've lent it to a guy at work | 13:31 |
| olofk | stekern: What the hell... I wonder why I missed that | 15:45 |
| olofk | ah... now I remember. I wanted to merge the fixes that _franck_ did, so I checked out the version just before he forked it | 15:46 |
| olofk | Crap | 15:46 |
| olofk | Thinking of how to merge this makes me want to cry | 15:46 |
| olofk | git, I hate you so much | 15:46 |
| olofk | or actually... I'm not at all sure what I did | 15:52 |
| stekern | me neither, since you seem to have some of the new stuff manually merged in ;) | 15:54 |
| stekern | just do a 'git pull -r' and fix the conflicts | 15:54 |
| olofk | On which repo? :) | 15:54 |
| stekern | my repo | 15:54 |
| olofk | And then force push it to my github repo? | 15:55 |
| stekern | yes, that you can do when you verified that you didn't screw up your changes when you fixed the conflicts ;) | 15:56 |
| juliusb | I presume a few of you here recieved that email from Jerry Harthcock about the GPU thing | 16:03 |
| juliusb | it's worth Googling his name, too, if you have a moment | 16:03 |
| juliusb | an even more internet history :) | 16:03 |
| juliusb | err, even more *interesting* internet history | 16:03 |
| juliusb | This is the best one, though: http://www.escapistmagazine.com/forums/read/7.339774-Psychic-Apollo-16-Astronauts-Found-Alien-Life-on-the-Moon | 16:04 |
| juliusb | although his GPU thing looks real, at least | 16:11 |
| poke53282 | why he is talking about Android applications? | 16:20 |
| poke53282 | Well, I can try to compile Dalvik :) | 16:21 |
| poke53282 | But his steps make sense. | 16:31 |
| wallento | "Harthcock also wants to profit from importing alien technology by assembling a crack team of 10 professional psychics." | 16:32 |
| wallento | "Harthcock also wants to profit from importing alien technology by assembling a crack team of 10 professional OpenRISC developers." | 16:32 |
| wallento | All in | 16:32 |
| wallento | It actually sounds interesting | 16:32 |
| wallento | although I am sceptical about OpenRISC in there | 16:33 |
| poke53282 | The article is fun | 16:33 |
| wallento | I think the large cores in the GPGPU need to be more powerful | 16:33 |
| poke53282 | "Beginning in 1998, I have been formally trained to the advanced level (Stage-6) in Controlled Remote Viewing by ex-Defense Intelligence Agency military remote viewer" | 16:34 |
| wallento | nevertheless, it might be a chance to bring in stuff like out-of-order into mor1kx | 16:34 |
| wallento | or multi-issue | 16:34 |
| wallento | maybe the name is just some stealth mode name? | 16:35 |
| wallento | like herbert schlangemann | 16:35 |
| poke53282 | :) | 16:36 |
| poke53282 | Offtopic: Take a look at the video. http://www.heise.de/newsticker/meldung/Mit-VR-Brille-und-Motion-Capturing-zum-eigenen-Holodeck-2441906.html | 17:16 |
| poke53282 | I have a Oculus DK1 and know the potential. My next apartment must have an additional room for this ;) | 17:17 |
| wallento | :) | 17:25 |
| stekern | juliusb: how do you know that your google name searches hit the right guy? | 19:33 |
| stekern | I mean, I don't mind if you believe that I'm a 65 year old general in the Swedish military ;) | 19:35 |
| olofk | I'm a really old guitarist on google | 19:35 |
| stekern | olofk: but we look much younger in person! | 19:39 |
| olofk | And thinner :) | 19:39 |
| rschmidlin | Hi there | 19:39 |
| stekern | haha | 19:39 |
| stekern | hi rschmidlin | 19:39 |
| olofk | Can I just change remote "origin" to a different repo in .git/config to push it somewhere else? Or will that make the git gods angry? | 19:40 |
| stekern | olofk: at least we're still alive, in contrast to juliusb: http://www.findagrave.com/cgi-bin/fg.cgi?page=pv&GRid=83250106&PIpi=61400209 | 19:41 |
| olofk | Yeah, I also use findagrave.com a lot | 19:41 |
| olofk | I can never remember where I put all my graves, so it's really helpful | 19:42 |
| olofk | stekern: I think I managed to sort out the repo situation now and force-pushed to olofk/wb_sdram_ctrl | 19:43 |
| stekern | I think you can, but you can also do 'git remote remove origin; git remote add origin <other origin>' | 19:43 |
| olofk | rschmidlin: Did you get my shared document invitation? | 19:44 |
| rschmidlin | yes, I did. I’m trying to put down some ideas. | 19:45 |
| olofk | Perfect. I'm out of ideas myself | 19:45 |
| rschmidlin | anyone want to join to input ideas? | 19:54 |
| rschmidlin | I have come up with 4 main goals for now | 19:55 |
| olofk | 1. Start foundation | 19:56 |
| olofk | 2. Collect money | 19:56 |
| olofk | 3. ??? | 19:56 |
| olofk | 4. Profit | 19:56 |
| rschmidlin | I’m trying more to put the goals together first | 19:56 |
| rschmidlin | but straight to point 4 works perfectly well too | 19:56 |
| rschmidlin | https://docs.google.com/document/d/18RxSLq_2pjz4PN0N5V3L0Uj45hZV0SVyBLt3iy5m34M/edit?usp=sharing | 19:59 |
| rschmidlin | ok, I think that’s what I have without further brainstorming. | 20:20 |
| olofk | I'm back now. Will read what you have written down | 20:34 |
| rschmidlin | :) | 20:37 |
| olofk | rschmidlin: Did you see my messages on google docs? It looks like you are just disconnecting and reconnecting all the time | 20:45 |
| rschmidlin | oh | 20:45 |
| rschmidlin | let me see | 20:45 |
| olofk | stekern: Do I need IMMU and DMMU to run Linux? | 21:01 |
| olofk | rschmidlin: Yep. We're on a roll now :) | 21:02 |
| rschmidlin | woowoo | 21:03 |
| stekern | olofk: not sure if I should take that question seriously? =P | 21:10 |
| rschmidlin | it’s out | 21:10 |
| olofk | stekern: ah right. That was the whole point of uClinux | 21:11 |
| stekern | the answer is yes, unless you add no-mmu support to openrisc Linux | 21:12 |
| stekern | ...but I see that you didn't actually state that this was about openrisc in your question, so the answer is no. | 21:13 |
| olofk | Yes. It's for one of the many other processors that I'm heavily involved with :) | 21:14 |
| rschmidlin | See you guys, take care! | 21:16 |
| olofk | bye | 21:16 |
| --- Log closed Wed Nov 05 00:00:14 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!