| henriok | Greetings! | 10:30 |
|---|---|---|
| klys | hello | 10:31 |
| henriok | I’ve been following Libre-Soc for a time now, and Luke seems eager to make a decent Wikipedia article. I agree. That’s probably due now. | 10:39 |
| henriok | I’m here to help, if you need it. I have 16 years of editing/authoring Wikipedia articles in this aprticular area. | 10:39 |
| klys | lkcl, ping. | 10:50 |
| klys | i'm in between conversations atm. | 10:50 |
| lkcl | klys, ack | 11:13 |
| lkcl | henriok, appreciated | 11:14 |
| klys | :) | 11:14 |
| klys | what will the title be, Libre-SOC ? | 11:14 |
| lkcl | yes. or perhaps OpenPOWER_Libre-SOC if that's the convention | 11:15 |
| lkcl | there's a stack of articles here for references https://libre-soc.org/3d_gpu/articles/ | 11:15 |
| lkcl | i cut/paste started from here https://en.wikipedia.org/wiki/User:Lkcl/Libre-SOC | 11:15 |
| lkcl | henriok, klys ^ | 11:16 |
| lkcl | it is possible for other people to edit User:talk pages, right? | 11:16 |
| lkcl | you just have to have a login, is that correct? | 11:17 |
| henriok | I had a discussion with myself how to name the Microwatt article.. just ”Microwatt” (no) or ”IBM Microwatt” (no, too narrow and not in spirit) so I settled with ”Openpower Microwatt” recetly changed (which is OK) to ”OpenPOWER Microwatt”. | 11:19 |
| henriok | I think ”Libre-SOC” is good. | 11:19 |
| klys | I'm not logged in and it has [edit] stanzas, yes | 11:19 |
| henriok | just Libre-SOC that is. Including OpenPOWER there is not what I see Libre-SOC to be, unless you shift stewardship of the project to the OpenPOWER Foundation. | 11:20 |
| henriok | Other peopla can use Talk pages yes. | 11:20 |
| lkcl | henriok, Libre-SOC is the official name of the project. | 11:21 |
| lkcl | ahh got it about OPF | 11:21 |
| henriok | lkcl: Your style of editing is to change often in small incremets, no? | 11:21 |
| lkcl | henriok, yes | 11:23 |
| lkcl | means that anyone else editing the page doesn't end up with conflicts | 11:23 |
| * lkcl just doing a bit about "history" | 11:25 | |
| henriok | yes, Wikipedia isnt too good at merging concurrent edits. at all. | 11:40 |
| lkcl | yehyeh, so i try to keep it very short (and local) | 11:41 |
| lkcl | hmm this is challenging not to do stuff that i've done rather than what's been published by other people | 11:41 |
| lkcl | like, we can't refer to the video demonstrating running the Litex BIOS on an FPGA, because i was the one that did it | 11:41 |
| lkcl | and it's not been picked up by news / press / OPF / Phoronix / etc | 11:42 |
| * lkcl taking a break, out for a walk, leave it with you | 11:44 | |
| henriok | Yes :) I understand that. You are amazingly ”by the book”. I love it. Most first party people wouldn’t even understand this issue with Wikipedia. | 12:24 |
| lkcl | henriok: been there... :) | 12:25 |
| lkcl | henriok: as an aside, i worked out the other day how many variants are added by SVP64 to each instruction | 13:21 |
| lkcl | it's measured in the HUNDREDs. per instruction! | 13:21 |
| lkcl | if you count the vector length, VL=2..64, it's literally in the thousands. | 13:22 |
| lkcl | ew=8/16/32/default, on both src operands and dest operands, that's x4 x4 = x16 | 13:22 |
| lkcl | scalar/vector per operand (src/dst) it varies between 4x and 16x (2-operand to 4-operand) | 13:23 |
| lkcl | dang that's 256 already | 13:23 |
| lkcl | then there's "modes" - that's about another... x10 (saturation signed/unsigned) no wait more than that because you get 8 BO modes for CR pred-result | 13:24 |
| lkcl | probably around x30 | 13:25 |
| lkcl | then there's predication as well (2 different types), and twin predication | 13:25 |
| lkcl | holy cow | 13:25 |
| lkcl | between 1,000 and a whopping 32,000 variants. | 13:26 |
| lkcl | *of a single instruction* | 13:26 |
| henriok | lkcl: Made some significant changes/Additions to the article. You might want to reload. | 13:26 |
| lkcl | henriok, keeping an eye out https://en.wikipedia.org/w/index.php?title=User%3ALkcl%2FLibre-SOC&type=revision&diff=1021198661&oldid=1021183840 | 14:14 |
| lkcl | found a link that can be considered "notable" https://openpowerfoundation.org/cores-and-more/ | 14:23 |
| lkcl | henriok: we can't run linux yet (no MMU, working on that), and i haven't been able to test Zephyr yet | 14:37 |
| henriok | lkcl: That’s a vestige from the Microwatt article. I didn’t know the current state of Libre-SOC in that regard. | 15:19 |
| henriok | Also.. fabbing.. what’s the current status on that? | 15:19 |
| henriok | lkcl: The article is good to publish in my oppinion. | 15:27 |
| lkcl | henriok: cool. | 15:38 |
| lkcl | ahh 1 sec let me find something about the new Jun 9 tapeout | 15:38 |
| lkcl | nuts, it's not gone out as a public discussion. | 15:39 |
| lkcl | henriok: it's just a "move", right? change from "User" to "(Article)" along with the page-name? | 15:45 |
| lkcl | looks like it https://en.wikipedia.org/wiki/Help:How_to_move_a_page | 15:46 |
| henriok | No, easiest is to just create the new page in the open name space,and paste in the text. | 15:46 |
| henriok | I have bad experiences with moving any pages. | 15:46 |
| henriok | It’s mostly done if one have a need to keep a discussion and edit history intact. I dont think we need that now. | 15:47 |
| henriok | You cankeep your copy in your User space. It wont list in seraches. | 15:48 |
| lkcl | i did the development of the LMDB page, i was quite pissed when someone took a copy, and did not preserve the history of 3 *weeks* of over 100 edits that i'd made | 15:54 |
| henriok | If you want to preserve the history, then do a move. | 15:58 |
| lkcl | done | 15:59 |
| lkcl | thank you for your help! | 15:59 |
| henriok | Excellent! | 16:00 |
| lkcl | i noticed google searches for Microwatt put the wikipedia link pretty much right at the top | 16:01 |
| henriok | This meta page is usually a nice page to keep an eye on https://en.wikipedia.org/wiki/Special:WhatLinksHere/Libre-SOC | 16:08 |
| lkcl | oo nice, will put that on my user page | 16:10 |
| *** henriok_ is now known as henriok | 16:19 | |
| lkcl | ha! the MMU <-> dcache interaction works, no data corruption | 16:32 |
| lkcl | disabled the actual cache, physical memory only, but it works | 16:32 |
| lkcl | it's a huge first step | 16:32 |
| jn__ | nice | 16:55 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!