--- Log opened Fri Oct 04 00:00:24 2013 | ||
hansfbaier | stekern: How much memory does fitting the sockit eat on your machine? | 03:33 |
---|---|---|
hansfbaier | stekern: never mind, seems to work now | 03:44 |
stekern | good, because I have no clue | 03:46 |
stekern | haven't checked, I've got 16GB of RAM in this machine | 03:47 |
stekern | poke53282: at least someone is responding to your patches, that's better than them being ignored | 03:59 |
hansfbaier | stekern: I installed in a VM with 3,5 GB. Wasn't enough. Swap storm :( My host has 8 | 04:03 |
hansfbaier | Ah wait it still seems to be alive :) | 04:04 |
hansfbaier | kind of .... :] | 04:05 |
stekern | it's just crazy that it needs that much memory | 04:05 |
hansfbaier | stekern: right now it eats about 3G | 04:05 |
hansfbaier | placement and routing | 04:05 |
stekern | hansfbaier: do you have any linux running on the board already? | 04:24 |
hansfbaier | stekern: Yes I the demo image | 04:25 |
hansfbaier | stekern: s/I// | 04:25 |
hansfbaier | stekern: nice toys.... LCD / LEDs etc | 04:25 |
hansfbaier | stekern: tested some of the software examples | 04:26 |
stekern | ok, when you get to the point that you've got an fpga image to load, there's a few things you need to know (that also should go into the README) to make it useful with the orpsoc image | 04:26 |
stekern | first, you will need to build your own u-boot spl (preloader) and put that to the sdcard | 04:27 |
stekern | because otherwise the uart connection inside the fpga will not work | 04:28 |
hansfbaier | stekern: Where do I get the spl? | 04:30 |
stekern | you could alter the registers manually in u-boot too I guess, but since the second uart of the hps isn't connected to anything on the outside, you'll always want it to be internally connected | 04:30 |
stekern | it's in altera/13.0sp1/embedded/host_tools/altera/preloader/ | 04:34 |
stekern | but they have some convoluted way of building that which involves using some gui, wait and I'll check for the instructions how to do it | 04:35 |
hansfbaier | stekern: How did you find out about all this?!?! | 04:41 |
hansfbaier | stekern: A question: In Qsys in some designs the AXI bus from the HPS seems to be directly connected to some avalon ports. Shouldn't be there some kind of AXI->avalon bridge, or are the busses nearly identical? | 04:47 |
stekern | hansfbaier: one question at the time!!! =) | 04:49 |
hansfbaier | stekern: ok, I'll google in the meantime :] | 04:49 |
hansfbaier | http://www.alterawiki.com/wiki/AXI_Support | 04:49 |
stekern | first, the spl building process was described in the workshop, that's was about as far as it went | 04:50 |
hansfbaier | spl? | 04:50 |
hansfbaier | preloader | 04:50 |
stekern | but I can't find the lab instructions anywhere on the net | 04:50 |
stekern | yes, u-boot spl | 04:50 |
stekern | is the preloader | 04:50 |
hansfbaier | There is a hardware and a software lab | 04:50 |
stekern | yes | 04:50 |
hansfbaier | I downloaded somewhere | 04:50 |
hansfbaier | Got the files | 04:50 |
stekern | do you have those notes | 04:50 |
stekern | great! | 04:50 |
hansfbaier | yes I think so. | 04:50 |
stekern | look in the sw lab | 04:50 |
hansfbaier | Yes, I figured out I really should do the labs | 04:51 |
hansfbaier | stekern: ok. | 04:51 |
stekern | ok, so onto the second question. qsys automatically inserts axi<->avalon bridges when you connect an axi port to an avalon port | 04:52 |
hansfbaier | stekern: Ah, ok. I'll do the labs and come back with questions when I'm done with the labs | 04:52 |
stekern | I think it's enough that you read the documents | 04:53 |
stekern | ...but feel free to do the labs ;) | 04:53 |
hansfbaier | stekern: The HW lab gives me compile errors Error (12006): Node instance "u0" instantiates undefined entity "soc_system" | 04:54 |
hansfbaier | Does this mean I have to run qsys first? | 04:54 |
stekern | the lab notes doesn't describe how to get the spl onto the right partition on the sdcard though | 04:54 |
hansfbaier | looks like that | 04:54 |
hansfbaier | what does spl stand for? | 04:54 |
stekern | yes, you have to run qsys and do all the steps in the lab (it configures the DDR3 SDRAM among other things) | 04:55 |
stekern | spl = second program loader | 04:56 |
hansfbaier | stekern: Ah, yes sorry, I started in the middle of the PDF, thought I was at the beginning because evince remembered the page | 04:57 |
hansfbaier | last opened | 04:57 |
stekern | but still about the axi<->avalon bridges, I took a bit of a short-cut based on that fact. Since I did the avalon<->wishbone bridges in orpsocv3, I connect the axi port of the hps to a "dummy" avalon bridge and then connect the wishbone bus to that. | 04:59 |
stekern | we shouldn't do that, we should have a straight axi<->wb bridge instead | 04:59 |
hansfbaier | stekern: Is it worth the effort? I think avalon is only a couple of gates away from wishbone (converting read and write into cyc and we etc. iirc) | 05:09 |
stekern | probably not for what it's connected to now, the "lightweight axi bus". controlling peripherals | 05:18 |
hansfbaier | Info: Quartus II 64-Bit Fitter was successful. 0 errors, 99 warnings | 05:18 |
hansfbaier | Info: Peak virtual memory: 4128 megabytes | 05:18 |
hansfbaier | Info: Elapsed time: 01:33:56 | 05:18 |
stekern | the SDRAM have avalon interfaces anyways | 05:19 |
hansfbaier | stekern: that is some turnaround time | 05:19 |
stekern | but, it's ugly to have to inferr that extra bridge, and it would be nice to have an axi bridge as well | 05:20 |
stekern | and converting avalon pipelined accesses or bursts is a bit more than a couple of gates | 05:20 |
stekern | (take a look here: https://github.com/openrisc/orpsoc-cores/blob/master/cores/wb_avalon_bridge/verilog/wb_to_avalon_bridge.v) | 05:21 |
stekern | Info: Elapsed time: 00:14:52 | 05:25 |
stekern | that's what it took to build orpsoc last on my machine | 05:26 |
stekern | s/build/run the fitter/ | 05:26 |
hansfbaier | stekern: I ran it in the 64bit vm, was so slow because of swapping and no hd cache available probably | 05:27 |
hansfbaier | it's done the .sof is there. | 05:28 |
hansfbaier | stekern: is there a testbench to the wb bridge somewhere? If not, how did you develop it / make sure it works? | 05:29 |
stekern | I'm not sure it works ;/ | 05:33 |
stekern | a more complete answer, I should do a testbench for it, the problem is that the avalon BFMs are some encrypted piece of crap, so I would need to run them through modelsim. I should do that, but it put me off at the moment | 05:35 |
stekern | so I've just tested them in hardware | 05:35 |
stekern | and they seem to work fine for everything I've thrown at them there, but that's of course no guarantee | 05:36 |
stekern | as a debug tool I used the diila core (as you can see, the avalon and wishbone buses are connected to that) | 05:37 |
hansfbaier | stekern: What does BFM stand for? (googled but could't find something sensible) I know its some kind of simulation model.... | 05:38 |
hansfbaier | Behavioural functional models? | 05:39 |
stekern | and I have a set of scripts and tools on the board to load a binary into memory, flip the resets on the or1k core, setup diila, run the or1kx program and dump the diila log into a text file, and send it via scp to my dev machine | 05:39 |
hansfbaier | stekern: that is really cool ... | 05:39 |
stekern | Bus Functional Model, I didn't know what it stood for a week ago neither ;) | 05:39 |
stekern | but we discussed it then, so I thought maybe you had seen the explanation | 05:40 |
stekern | I'll give you the tools and scripts.. but for some reason I haven't transferred them from the board it seems | 05:41 |
stekern | and I have the board packed down now | 05:41 |
hansfbaier | stekern: no hurry | 05:45 |
hansfbaier | stekern: I've got enough to learn.... | 05:45 |
hansfbaier | in the meantime | 05:45 |
stekern | yeah, alot of this stuff was new for me too, it takes a while to sink in | 05:46 |
_franck_ | juliusb: jeremybennett : I'll do a quick presentation of the current GDB / openocd status if it fits the planning | 09:02 |
juliusb | _franck_: that'd be great | 09:15 |
juliusb | we'll figure out where to fit it in.... | 09:15 |
juliusb | I'm wondering if anyone has any experience with running the DE0 nano programmer and OpenOCD in a VM running on top of Windows? | 09:16 |
juliusb | It should "just work", right? | 09:16 |
_franck_ | I did | 09:17 |
_franck_ | and it works | 09:17 |
juliusb | awesome, thanks _franck_ | 09:18 |
juliusb | _franck_: when are you planning to travel? | 09:18 |
_franck_ | I'll be at gatwick at 8 am | 09:20 |
_franck_ | saturday | 09:20 |
juliusb | ah, that's right, I remember now. Well, if anyone is in Cambridge tonight and would like a warm-up beer, email me! | 09:23 |
_franck_ | is the cambridge train station 1 hour walk from the university ? | 09:26 |
_franck_ | (that's what's google says) | 09:26 |
juliusb | from the Computer Labs, maybe | 09:31 |
juliusb | but the university buildings are distributed around town, the center of town is maybe a 20 minute walk from the station | 09:31 |
_franck_ | ok thanks | 09:38 |
olofk | stekern: Yeah, a proper axi<->wb bridge would be useful for a lot of things | 10:32 |
olofk | It's probably easiest to start out with AXI4-Lite. The complete AXI4 bus is a beast | 10:33 |
stekern | yes, I agree | 10:33 |
stekern | and currently it's AXI4-lite that I need | 10:34 |
stekern | ...I think the "lightweight hps2fpga interface" use axi4-lite at least | 10:36 |
stekern | I have the pleasure to work with a pentium machine from the mid-90s with 64MB of ram today | 10:37 |
olofk | I'm impressed | 10:37 |
hansfbaier | stekern: that would classify as an embedded machine today | 10:38 |
stekern | it's funny how old computers make that special sound that modern machine don't do | 10:38 |
olofk | stekern: When they boot? | 10:38 |
stekern | yes, the harddisk sound and the blippiing | 10:38 |
olofk | I miss that :( | 10:38 |
stekern | and loud fans | 10:38 |
olofk | I don't miss that :) | 10:39 |
stekern | it brings old memories back ;) | 10:39 |
olofk | You should run coremark on it and see how it fares against mor1kx on a de0_nano :) | 10:39 |
stekern | haha, true | 10:39 |
stekern | not sure how I'd get it onto it though, it only has floppy disk | 10:41 |
stekern | drive | 10:42 |
hansfbaier | stekern: I bougt a power macintosh 2000 for about $4000. It had a killer 768 MHz CPU with a HUGE heat sink on it. And it had GIGABIT ETHERNET. Don't know what for, but it had it. Was a big waste of money. I ended up running Debian on it. | 10:43 |
hansfbaier | and it was LOUD | 10:43 |
* hansfbaier stopped wasting money on apple products. PC still rules | 10:44 | |
stekern | I've always went for the hardware right under the price dip | 10:45 |
hansfbaier | stekern: me too but I wanted it for recording..... And Linux wasn't far enough back then | 10:46 |
stekern | ah, ok, fair enough | 10:46 |
stekern | I remember the first sound card I bought, a soundblaster pro 8 bit stereo! | 10:47 |
stekern | I gave around 200$ for it back then | 10:48 |
stekern | oops $ sign went on the wrong side... | 10:48 |
hansfbaier | stekern: It took the PC worls more than a decade to reach the power of the AMIGA | 10:49 |
hansfbaier | s/worls/world/ | 10:49 |
stekern | I could record about 5 seconds of audio before the harddisk was full | 10:49 |
stekern | mmm, I had an amiga too | 10:49 |
hansfbaier | stekern: my first PC harddisk was 80 MB with 4MB RAM. First Linux system. 386SX25 | 10:49 |
hansfbaier | stekern: ditched my AMIGA for it. Finally had a UNIX in my home | 10:50 |
hansfbaier | that was '95 | 10:50 |
stekern | I think my first linux install was on a 486 25MHz | 10:50 |
hansfbaier | Linux 1.0 | 10:50 |
stekern | also in '95 | 10:50 |
hansfbaier | stekern: gotta go and play music in the mall. Bye! | 10:51 |
stekern | play well! =) | 10:52 |
hansfbaier | Thanks | 10:53 |
hansfbaier | I'll try | 10:53 |
Powermaniac | ...Play music in the mall? | 10:54 |
Powermaniac | Too late anyway he is gone | 10:54 |
stekern | he plays with a band | 10:54 |
Powermaniac | Oh | 10:55 |
stekern | redhat 2.0 it was | 10:55 |
stekern | I remember the troubles I had to get the modem working | 10:55 |
stekern | and the catch 22 to not be able to check internet since the modem wasn't working | 10:56 |
Powermaniac | Ha | 10:58 |
Powermaniac | Whoever decided to have quardratic equation questions with fractions on Khan Academy is a bastard...xD | 10:59 |
Powermaniac | So you have all been preparing for the ORCONF right? | 11:00 |
stekern | hmm, you know you're a geek when your luggage contains more dev boards, computers and electronics than clothes... | 11:01 |
Powermaniac | Hahaaa | 11:01 |
Powermaniac | Are you traveling by car, bus, train or plane? | 11:01 |
Powermaniac | Maybe boat? | 11:01 |
stekern | plane | 11:01 |
Powermaniac | Have fun with the metal detectors!... | 11:02 |
Powermaniac | Unless that isn't an issue there, if so lucky you | 11:03 |
ams | it is not the metal detectors that are run... but the alone time you get with a nice officer. | 11:05 |
Powermaniac | ams: Oh even better...O_o | 11:05 |
Powermaniac | Can I ask you guys a question. Well the question is why do you guys seem to use Google+ over say Twitter? | 12:43 |
juliusb | Google+ and Twitter are fairly different things | 12:50 |
Powermaniac | True but I'm still curious why you use one over the other | 12:51 |
Powermaniac | Mainly wondering because I'm trying to make contacts with people in the different fields I'm interested in | 12:51 |
juliusb | You don't have to. You can use both. | 13:04 |
Powermaniac | Well that is what I've been doing but yeah... | 13:04 |
olof_ | It's bad that we recorded the talks last year. I would like to reuse most of my presentation, but don't want to get caught | 16:59 |
olof_ | Actually... I'll probably reuse most of it anyway. Lots of new users | 17:00 |
olof_ | So if you have seen it before you can use the time to take a nap | 17:00 |
olof_ | jeremybennett: I just realized that we haven't allocated time for a follow-up on the processes that we discussed last year | 17:27 |
knz | hi all | 21:48 |
knz | I was just wondering what the best way is to travel from the train station to the niversity with public transportation | 21:51 |
juliusb | knz: https://www.google.co.uk/maps/preview#!data=!4m31!3m30!1m5!1sCambridge%2C+Railway+Station+(Stop+A)%2C+Cambridge%2C+Cambridgeshire+CB1!2s0x47d87083b81c08ff%3A0xccde5e30c9f6db5!3m2!3d52.19449!4d0.13685!1m5!1sUniversity+of+Cambridge+Computer+Laboratory%2C+William+Gates+Building%2C+15+JJ+Thomson+Ave%2C+Cambridge+CB3+0FD%2C+United+Kingdom!2s0x47d8774a3f6e55cd%3A0xabf8227343e684c7!3m2!3d52.210891!4d0.091732!2e3!3m8!1m3!1d23839!2d0.1099993!3d52.2058215!3m2!1i540!2i468!4f13.1!6m4!1m3!1e4!2e2!3j1380974400!8m2!1e1!2b1!6m1!1e1&fid=0i1 | 23:08 |
juliusb | I'd just cab it :-/ | 23:09 |
knz | thx | 23:10 |
--- Log closed Sat Oct 05 00:00:26 2013 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!