| --- Log opened Tue Dec 16 00:00:13 2014 | ||
| stekern | I've been pondering about a legal issue for a while, we see corps (i.e. MIPS/ARM) hunting down implementations that clone their ISAs (which still have valid patents) | 05:52 |
|---|---|---|
| stekern | but, afaik, the same thing is not happening to emulators/simulators | 05:53 |
| stekern | is that because patents can't expand to software? | 05:53 |
| stekern | and if so, where does the line go? could instruction translation a'la transmeta be interpreted as emulation? | 05:54 |
| stekern | ...to make clear, it's a purely hypothetical question, I have no visions of putting it in practice ;) | 06:05 |
| poke53282 | I think one important point is, if you use the patent commercially or not. I am working in resaearch and we don't have a problem to use patented stuff, like algorithms. | 07:19 |
| poke53282 | But I simply think, that emulators don't really matter here because they are usually slow and have limited applications. To fight against the community on this level is also a bad idea. | 07:21 |
| poke53282 | commercial hardware implementations need a license I would say. | 07:22 |
| poke53282 | no matter if it is emulated or not. | 07:23 |
| stekern | yes, but that's was my question, are emulators allowed to exist because of the reasons you said, or because they can't 'fight them' on a legal basis | 07:31 |
| stekern | I remember that sony at least was upset when the first ps1 emulator emerged, since the console was still relevant when it was released, and cheap x86 hardware could be used to obtain same performance as the console | 07:34 |
| stekern | and the interesting follow-up discussion would be, when is hardware hardware? | 07:36 |
| bentley` | emulators of what? | 07:37 |
| bentley` | in the us, emulation is pretty clearly legal | 07:37 |
| bentley` | see sony v. bleem, etc | 07:37 |
| stekern | in this case, cpu ISA/architecture emulation | 07:38 |
| stekern | yes, sony v. bleem was the case I was referring to | 07:38 |
| bentley` | sony v connectix is sort of related (but not actually about emulation) | 07:39 |
| bentley` | same for sega v accolade... | 07:39 |
| stekern | right, so imagine on-chip software emulation of a patented ISA, would that be considered a patent infringement? | 07:43 |
| bentley` | hm | 07:43 |
| bentley` | most lawsuits that were ruled in favor of emulation were copyright related, not patent related... | 07:45 |
| bentley` | i would ask an ip lawyer :P | 07:46 |
| stekern | yeah, I'm not expecting legal advise here ;) | 07:46 |
| bentley` | i can't think of any patent related suits | 07:46 |
| bentley` | copyright and trademarks are both basically fair game if it's for interoperability purposes | 07:47 |
| bentley` | (the trademarks part comes from sega v accolade) | 07:47 |
| stekern | mmm | 07:47 |
| bentley` | scalia made an interesting dissent in the aereo case | 07:50 |
| bentley` | basically pointing out that the supreme court nearly made the vcr illegal (5-4 against) | 07:50 |
| bentley` | and that we should be cautious about labeling things as ip infringement that would cause damage | 07:51 |
| olofk | Is column width always equal to the number of address pins on a DDR2 IC? | 09:56 |
| ysionneau | I have a DDR1 datasheet here where this is not the case | 10:17 |
| ysionneau | but the row width is equal to the number of address pins in this case | 10:18 |
| olofk | ah ok. I mean row then | 10:18 |
| ysionneau | for the rows I would think so, but I'm not sure, I only worked with sdram and ddr1 so far | 10:23 |
| olofk | I'm so looking forward to when serial connections to memories become more common, like Hybrid Memory Cube. I'm sick and tired of this DDR junk | 11:05 |
| olofk | stekern: Did you have a mor1kx qsys wrapper somewhere? | 11:07 |
| LoneTech | olofk: sadly, that's being vindictively restricted using a patent | 11:07 |
| stekern | olofk: uhmmm, yes, I should have somewhere | 11:08 |
| LoneTech | iirc | 11:08 |
| stekern | http://oompa.chokladfabriken.org/openrisc/mor1kx_hw.tcl | 11:09 |
| stekern | not entirely up-to-date though | 11:10 |
| olofk | LoneTech: That's sad to hear. I was hoping for a revolution in memory connections here :( | 11:11 |
| olofk | stekern: A colleague of mine who used to work on the LEON stuff was home yesterday and played around with OpenRISC. | 11:14 |
| olofk | https://github.com/eiselekd/hw/tree/master/cyclone/bemicro_or1k | 11:14 |
| olofk | I will convince him to FuseSoCize it. Bemicro SDK is a cool little board. I wanted to get one | 11:14 |
| stekern | I have the industrial ethernet version of that here on my desk | 11:16 |
| stekern | ...unfortunately it's loaded with an fpga image I don't have, so I don't want to tamper with it | 11:17 |
| olofk | ah cool | 11:21 |
| stekern | hmm, is he using the avm interface? | 11:22 |
| stekern | maybe a warning about that might be in place... it's not thoroughly tested | 11:22 |
| olofk | stekern: Haven't looked, but that would make sense since he's doing it Altera style | 11:23 |
| olofk | Yeah. I'll tell him :) | 11:23 |
| olofk | Arrow sells the boards for $74, so I assume that the work spent on the unknown FPGA image is a bit more than that :) | 11:23 |
| olofk | So I wouldn't touch that one either | 11:24 |
| stekern | I appreciate if he want to be adventurous and report problems, but a safer bet is probably to use the wb2avalon cores | 11:24 |
| olofk | I'll tell him that | 11:25 |
| olofk | I was considering using that too, but I got a bit scared by some of the comments | 11:25 |
| olofk | How would I know which burst modes to use? | 11:25 |
| stekern | ...the only reason they are there is because I fatfingered a push and pushed the wrong branch | 11:25 |
| olofk | You could remove them for the 4.0 series | 11:25 |
| stekern | (the avm interface I mean) | 11:25 |
| olofk | I assumed that | 11:26 |
| stekern | yeah, I have considered doing that | 11:26 |
| olofk | I'm a bit annoyed by the dangling pins warnings I always get from the AVM IF | 11:28 |
| stekern | olofk: what that comment is saying is "you don't know what burst modes to use, so we don't use burst but pipelined transfers instead" | 11:38 |
| stekern | or at least that's what it is supposed to say ;) | 11:38 |
| olofk | ah ok. I was a bit stressed out when I read it, so I backed off. It was my plan B anyway | 11:39 |
| olofk | My home-made DDR2 Controller works now, so I didn't have to use the bridge + ALTMEMPHY | 11:40 |
| olofk | At least it works in simulations with a different part than what's on the board. | 11:40 |
| olofk | The first board burned up before I got a chance to try it on HW | 11:41 |
| stekern | smoketest success? | 11:44 |
| olofk | Definitely | 11:44 |
| -!- Netsplit *.net <-> *.split quits: jeremybennett, ed-jones, sb0, radii, dalias, aou | 14:30 | |
| -!- Netsplit over, joins: dalias | 14:30 | |
| -!- Netsplit over, joins: aou | 14:30 | |
| olofk | ok, so it looks like my Altera Blaster is broken. Would it be possible to use the DE0 NAno as a JTAG adapter? | 21:10 |
| olofk | The idea would be then to just route the pins from the internal JTAG adapter to some I/O pins and hook that up to my board | 21:11 |
| --- Log closed Wed Dec 17 00:00:14 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!