| --- Log opened Wed Oct 08 00:00:28 2014 | ||
| poke53282 | Oh, I found, that someone submitted some patches for QEMU. http://www.mail-archive.com/qemu-devel@nongnu.org/msg260450.html | 03:33 |
|---|---|---|
| poke53282 | Valentin Manea | 03:34 |
| poke53282 | Most of the stuff based on my emulator. | 03:36 |
| poke53282 | Even the touchscreen. | 03:37 |
| poke53282 | I wonder if the devices are tested. | 03:45 |
| stekern | poke53282: yes, he posted to the openrisc mailing list about that before he posted them | 05:26 |
| poke53282 | Ok, I missed that. Oh maybe I reckignized it but then forgot. | 05:28 |
| poke53282 | It's Ok. The patches look good. | 05:28 |
| poke53282 | The touchscreen has nothing to with openrisc. You can guess how I have chosen it :) | 05:29 |
| stekern | well, neither does framebuffer, keyboard nor IDE ;) | 05:31 |
| poke53282 | At least they have "open" in its titles. | 05:32 |
| stekern | is the keyboard and ide based on some opencores cores? | 05:35 |
| stekern | I spent the morning coding a threaded starfield demo ;) | 05:36 |
| poke53282 | I guess so. Not sure. | 05:36 |
| stekern | it runs too slow on 1 core, but pretty smooth on 4 | 05:37 |
| poke53282 | we won't see a parallelized Mandelbrot zoom calculation? | 05:37 |
| stekern | I actually tried compiling xaos yesterday | 05:37 |
| stekern | but it was to cumbersome | 05:38 |
| stekern | it didn't understand cross-compiling at all | 05:38 |
| stekern | I could of course fix the one I have to be parallell | 05:39 |
| poke53282 | then just try to add CC=.... If the program is simple it will work. | 05:39 |
| poke53282 | Ok, I have booked a hotel room. Hotel Pension Theresia-Regina. I will arrive at the main station of Munich at around 3:30PM. | 05:41 |
| poke53282 | After a ten hour flight from Vancouver to Frankfurt and an 3 hour train ride. | 05:42 |
| stekern | xaos is not simple | 05:43 |
| stekern | it picks up the compiler correctly, but nothing else | 05:44 |
| stekern | did you get the rooms for 30€ there? | 05:45 |
| poke53282 | no, 50 Euro | 05:46 |
| poke53282 | I want my own WC and shower this time. | 05:46 |
| stekern | did they have rooms with that? | 05:47 |
| poke53282 | As an exception I don't care this time. | 05:47 |
| poke53282 | http://www.hoteltheresia.de/e_index.htm | 05:47 |
| poke53282 | rooms with lavatories | 05:47 |
| poke53282 | for some reason I don't pay 59 Euro, but 50 Euro. | 05:48 |
| stekern | why is the single room price more expensive than the double room price? | 05:48 |
| poke53282 | per person? | 05:48 |
| poke53282 | not sure | 05:48 |
| stekern | ah | 05:48 |
| poke53282 | I have stay awake after I arrive for at least 6 more hours. This is the second worst part of such a travel. | 05:50 |
| poke53282 | probably no free wifi. | 05:51 |
| stekern | is there even non-free wifi? | 05:51 |
| stekern | you have 13 hours of sleeping behind you then, no problem ;) | 05:52 |
| poke53282 | not in Germany. But I am get used to it in Vancouver. It is so easy to find a free wifi. | 05:52 |
| poke53282 | But there is something called "eduroam" wifi. When you are a student or work for a university, you get free access wordwide at each university. | 05:53 |
| stekern | it's the same in finland and sweden, hotels without free wifi is pretty much unheard of | 05:53 |
| poke53282 | When you are at a university check for this id. Most of the time you will find such a wifi id. | 05:53 |
| stekern | but I had free wifi in the small ~35 person hotel I stayed in last I was in germany | 05:54 |
| poke53282 | stekern: Germans are afraid. There is something called "Mitstoererhaftung" | 05:54 |
| poke53282 | "accomplice" maybe is best translation. | 05:55 |
| poke53282 | That means that you can partly make responsible what people do with your free wifi. | 05:55 |
| poke53282 | Yes, I had the feeling. In cheaper hotels you have a higher chance for free wifi. | 05:55 |
| poke53282 | It is stupid in Germany. It works everywhere else in the world. | 05:56 |
| poke53282 | Ok, wifi where you have to login is possible in Germany. | 05:58 |
| poke53282 | But really free wifi like at Starbucks you probably won't find. | 05:58 |
| stekern | yes, I was speaking about free as in no cost | 05:59 |
| poke53282 | this is rare but you can find it. | 06:00 |
| poke53282 | At the whole university in Canada where I am, everyone has free access. No login, you just have to say Ok to a disclaimer. | 06:01 |
| poke53282 | Finally also Frankfurt has free internet for one hour. | 06:02 |
| poke53282 | Frankfurt airport I mean. | 06:03 |
| poke53282 | I think this the third biggest airport in europe. | 06:03 |
| stekern | after submitting your e-mail iirc | 06:03 |
| poke53282 | Not sure, if they need your email. | 06:04 |
| stekern | tegel is at least the worst airport in the world | 06:09 |
| stekern | ...that I've been to | 06:10 |
| poke53282 | Yes, you are right. Tegel is terrible. And a shame for the main city of Germany. | 06:11 |
| poke53282 | They have build a new one. | 06:11 |
| poke53282 | But it is not open yet. | 06:12 |
| poke53282 | The question is if we and others should meet on Friday for a german beer and bavarian cuisine :) | 06:13 |
| poke53282 | http://en.wikipedia.org/wiki/Bavarian_cuisine | 06:13 |
| stekern | the best I've been to is singapore | 06:13 |
| stekern | yes, I think there are some plans for beer and salty food among us that arrive on friday | 06:14 |
| poke53282 | Vancouver airport is nice, but nothing special. Was never in Singapore. | 06:15 |
| stekern | another anecdote from my last visit in german, me and the collegue went to eat at some restaurant where there was no english menus and no english speaking staff. we managed to figure out something that sounded like bread from the starters menu and signed that we wanted two of those. the waitress looked suspicously at us and asked "zwei?" and we answered "yes, yes" | 06:21 |
| stekern | when the starters arrived, it turned out that one plate would have been enough for 4 people... | 06:22 |
| poke53282 | :) Yes, that can happen everywhere. Normally Germans understand english. We learn it in school for at least 6 years. But people forget. | 06:24 |
| poke53282 | bread -> Brot | 06:24 |
| poke53282 | I hope they packed the rest for you :) | 06:26 |
| stekern | yes, it was exactly what we thought it was, it was just that we didn't get the amount of it ;) | 06:27 |
| poke53282 | http://en.wikipedia.org/wiki/Bavarian_cuisine | 06:28 |
| poke53282 | http://www.muenchen.de/int/en/restaurants/bavarian-menu.html | 06:28 |
| poke53282 | First link was wrong | 06:28 |
| poke53282 | I would suggest that you make an offline copy of this list. | 06:29 |
| drakhogekko | hi | 06:46 |
| poke53282 | Hi | 06:47 |
| poke53282 | Ok, that's enough for today. Figured out, that my emulator runs faster on m pad with an ARM CPU than on my Core-i7. I don't have a clue why. That makes no sense. Something I am doing wrong or right. | 07:15 |
| olofk | haha. JavaScript has become insane | 07:20 |
| poke53282 | the optimization effort which is put into Javascript to optimize it is insane. | 07:21 |
| poke53282 | at least four companies with big teams working on it. | 07:22 |
| olofk | Yeah, that's what I really meant with insane. Has any other language gotten the same treatment? | 07:22 |
| olofk | What would happen if the same thing was done to python for example | 07:22 |
| poke53282 | Not that I know. | 07:23 |
| poke53282 | is python typed? | 07:24 |
| olofk | I think so, but I'm not entirely sure of the definition of typed | 07:25 |
| poke53282 | But it is interesting, that such a language like Javascript can indeed reach speeds which are not that far away from native compiled C code. | 07:25 |
| poke53282 | ("1" == 1) => true | 07:25 |
| poke53282 | no ints, only doubles for examples. But you are allowed to do logical operations with the lower 32 Bits. | 07:26 |
| poke53282 | But what is currently done with Javascript is really really insane. They should develop a new web language. | 07:27 |
| poke53282 | But maybe, that is the only feature, that is unique for each browser. | 07:28 |
| poke53282 | So cannot change how a website looks, and have to follow standards. | 07:29 |
| poke53282 | So, how to sell such a product. This is difficult. | 07:29 |
| poke53282 | When there is no difference between the browsers. | 07:29 |
| poke53282 | But at least with their Javascript engines, they can argue. | 07:30 |
| poke53282 | But why the hell is my PAD faster than my Intel CPU. Ok, I go to bed now. Hopefully, the answer will come in my dream. :) | 07:31 |
| olofk | They _have_ made new languages. Tons of them. Rust, Java applets, Dart, Flash... just that the browsers can't agree on which languages to support | 07:46 |
| olofk | Arrgghh! I'm going to kill whoever came up with that stupid FuseSoC crap! | 09:03 |
| olofk | oh wait....never mind | 09:03 |
| sb0 | olofk, just use misoc | 09:49 |
| sb0 | writing anything, and especially socs, in verilog is a bad idea | 09:49 |
| stekern | sb0: well, that's what fusesoc (as well as misoc) is out to prevent | 10:07 |
| olofk | sb0: misoc is so bloated ;) | 11:08 |
| -!- Netsplit *.net <-> *.split quits: aburgess, ams, tariq786, mithro, 16WAAR95O | 11:11 | |
| -!- Netsplit over, joins: aburgess, tariq786 | 11:11 | |
| sb0 | olofk, how so? | 15:10 |
| sb0 | you can run misoc/lm32 on lx4 | 15:19 |
| poke53282 | Slept, No dream. Checked again. Still the same result. | 17:33 |
| olofk | sb0: Yeah, I was just kidding | 18:57 |
| olofk | stekern: I'm running some simulations now on my SoC and it looks like the bottle neck is that I have to wait quite a long time to get data from the SDRAM. Does that sound reasonable? Can I do something to improve that? | 19:08 |
| olofk | Hmm.. can I tell gcc to align my buffer better? | 19:13 |
| olofk | Right now gcc has put it on 0x4a7c, which is a really bad match for the sdram controller | 19:14 |
| olofk | Just getting it to 0x4a80 would probably do wonders for performance | 19:14 |
| olofk | Looks like it's __attribute__((aligned(0x1000))) I was looking for | 19:18 |
| olofk | Unfortunately this seems to break the application | 19:19 |
| olofk | Aligning to 0x100 instead helped. (why?) | 19:20 |
| olofk | And why can I never learn how to measure distance between two markers in gtkwave? | 19:24 |
| olofk | Nah.. that didn't help very much actually | 19:28 |
| olofk | Is SDRAM really that slow, or am I holding it wrong? | 19:30 |
| olofk | stekern: Does the BURST_LENGTH parameter in wb_sdram_ctrl have to match the burst length from mor1kx? I set BURST_LENGTH (0) and it stopped working | 19:35 |
| olofk | aha.. the SDRAM was running at 50MHz in my simulations. Setting it to 100MHz helped quite a lot | 20:10 |
| olofk | Now I can saturate the I/O bandwidth. Nice | 20:11 |
| olofk | Compiling with -O2 to avoid some unnecessary memory accesses increased the bandwidth for the streamer even more | 20:37 |
| olofk | So all in all I went from ~8us to ~3us for streaming 64 words from memory | 20:38 |
| olofk | Has anyone done FFT in FPGA? Is that sensible thing to accelerate? | 20:44 |
| olofk | Looks like I found a nice one | 20:46 |
| olofk | Is doing FFT in software on OpenRISC a really really bad idea, or should I try that first? | 20:53 |
| juliusb | can't advise olofk | 20:58 |
| juliusb | wish I could, I never got into DSP as much as I thought I would | 20:59 |
| juliusb | hey so I'm wondering, who else from here is around in Munich on Friday night? | 20:59 |
| juliusb | There's dinner plans afoot right? | 20:59 |
| olofk | juliusb: Nice to see you around | 20:59 |
| juliusb | yeah I never am anymore :( | 21:00 |
| juliusb | stupid dayjob | 21:00 |
| olofk | Yes, I think that wallento has planned something | 21:00 |
| olofk | Yeah, same here. I wish I could hack on OpenRISC instead...oh wait... that's what I'm doing :) | 21:00 |
| juliusb | well it's the Paulaner Brauhaus in Munich at 19:30 but if you want to come (and you don't know if you are already) you should email olofk, wallento or myself so we can book | 21:00 |
| juliusb | but the more the merrier :) | 21:00 |
| olofk | Amen that | 21:01 |
| juliusb | poke53282: are you in Munich on Friday night? | 21:01 |
| olofk | Time to sleep now. Tomorrow I'll book the hotel | 21:04 |
| poke53282 | juliusb: Yes | 21:05 |
| poke53282 | Arrovomg around 3:30PM at the mein station | 21:05 |
| poke53282 | Arriving around 3:30PM at the main station. | 21:08 |
| juliusb | nice one | 21:08 |
| juliusb | I've AirBnB'd it. I should have thought about this in advance, and gotten a big apartment we could have shared :-/ Oh well | 21:08 |
| juliusb | Next year :) | 21:08 |
| poke53282 | :) Big apartment sounds good. | 21:09 |
| poke53282 | Well I arrive directly from Vancouver. | 21:11 |
| poke53282 | So Saturday would have been tough. | 21:12 |
| juliusb | ah you'll be in good shape for a night out, I know I always am after flying a long way east :) | 21:12 |
| --- Log closed Thu Oct 09 00:00:29 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!