| --- Log opened Sat Apr 20 00:00:23 2013 | ||
| -!- Netsplit *.net <-> *.split quits: simoncoo1 | 03:15 | |
| -!- Netsplit over, joins: simoncoo1 | 03:20 | |
| JohnSelf | Hi everyone | 05:23 |
|---|---|---|
| -!- Netsplit *.net <-> *.split quits: blueCmd | 05:25 | |
| JohnSelf | I have a question | 05:27 |
| JohnSelf | is anyone here to answer? | 05:27 |
| mor1kx | [mor1kx] skristiansson pushed 3 new commits to master: https://github.com/openrisc/mor1kx/compare/217882c2ef42...c8f720188503 | 05:30 |
| mor1kx | mor1kx/master 98d426d Stefan Kristiansson: cappuccino/lsu: move length and extension decoding to decode... | 05:30 |
| mor1kx | mor1kx/master d851189 Stefan Kristiansson: cappuccino: remove ctrl_opc_ins... | 05:30 |
| mor1kx | mor1kx/master c8f7201 Stefan Kristiansson: (pronto)espresso/cpu: regenerate | 05:30 |
| stekern | JohnSelf: how should we know when you haven't asked the question? | 05:31 |
| JohnSelf | haha | 05:32 |
| JohnSelf | sorry | 05:32 |
| JohnSelf | I was wondering how far along you have got with the design of the kncminer units | 05:34 |
| JohnSelf | i.e. what stage you are at now, and what target you have for production | 05:36 |
| JohnSelf | in terms of the target date | 05:36 |
| stekern | hmm, what does kncminer have to do with openrisc? | 05:41 |
| JohnSelf | my mistake | 05:45 |
| JohnSelf | but perhaps you could tell if the openrisc designs are advanced enough to be mass produced? | 05:47 |
| stekern | or1200 is stable enough for that, yes | 05:56 |
| JohnSelf | i thought so, great | 05:57 |
| JohnSelf | K | 05:57 |
| JohnSelf | I was hoping someone from orsoc would be here. | 05:57 |
| stekern | olofk and LoneTech works at orsoc | 05:58 |
| JohnSelf | Well if either of you come online, I'd love a rough estimate for kncminer units 1.0 | 06:00 |
| JohnSelf | Thanks for your help stekern | 06:01 |
| stekern | what is kncminer? | 06:09 |
| JohnSelf | some company that is claiming it will ship 500 orsoc-designed asic units within months | 06:21 |
| JohnSelf | I don't have much technical education but that seems unrealistic to me given the problems other asic developers have had. | 06:22 |
| JohnSelf | But I'm interested to hear from someone who is working on it. | 06:22 |
| stekern | juliusb: I have prepared this patch for (pronto)espresso: http://pastie.org/7672453 | 11:45 |
| stekern | it's a follow-up on this: https://github.com/openrisc/mor1kx/commit/98d426d13c216ca512051f8dec4feff7e2a8e755 | 11:46 |
| stekern | which move the lsu op code decoding into the decode module | 11:47 |
| stekern | for espresso, there is no practical change, but I prepared the patch to keep the lsu code more concistent between the pipelines | 11:48 |
| stekern | in hindsight, I probably should have broke that cappuccino/lsu commit in two, one that adds the decoding to mor1kx_decode and one that does the changes to lsu (like the patch I've now prepared for espresso) | 11:49 |
| stekern | anyways, what's your opinion on the espresso patch? | 11:50 |
| @juliusb | stekern: i like the patch | 13:35 |
| @juliusb | I assume it works against the tests? | 13:35 |
| @juliusb | btw hello all from the London chip hack | 14:24 |
| stekern | juliusb: yup, all tests passes | 14:29 |
| stekern | I'm playing a bit with the adder again, I noticed that the adder in or1200 gets inferred as a lpm_add_sub when compiled with the altera tools, but I can't for my life get the one in mor1kx to do that | 14:32 |
| stekern | not sure if it makes a difference, but worth a try | 14:32 |
| @juliusb | hmm, using macros is nice | 14:34 |
| @juliusb | btw please commit that patch if you like | 14:34 |
| @juliusb | looks good to me | 14:36 |
| stekern | I will, on the run right now | 14:43 |
| mor1kx | [mor1kx] skristiansson pushed 1 new commit to master: https://github.com/openrisc/mor1kx/commit/8bba14a4082d3fbdea86483d5c21b2933358cd81 | 15:49 |
| mor1kx | mor1kx/master 8bba14a Stefan Kristiansson: (pronto)espresso/lsu: move length end extension decoding to decode... | 15:49 |
| stekern | looks like you can do git pushes from this phone at least ;) | 15:50 |
| stekern | hmm, the code examples from altera has the result registered, maybe that's required for lpm_add_sub | 16:30 |
| stekern | well, I think I leave that alone for now, maybe we should just put the adder as a seperate module, that way users can more easily put their own (optimised) adder in there | 17:07 |
| --- Log closed Sun Apr 21 00:00:24 2013 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!