(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
[12:01 PM PT] Andrew Hellershanks: Hello, everyone.
[12:02 PM PT] Andrew Hellershanks: I have an errand to run this afternoon so I can't stay much past the hour.
[12:03 PM PT] Andrew Hellershanks: Hello, Kayaker.
[12:03 PM PT] Kayaker Magic: Hey Andrew!
[12:03 PM PT] Gavin.Hird @grid.xmir.org:8002: I view this as a 1680x1050 resolution but it is rendered at 3360 x 2088
[12:03 PM PT] Vincent.Sylvester @hg.zetaworlds.com: That explains the high usage then
[12:03 PM PT] Gavin.Hird @grid.xmir.org:8002: yes
[12:04 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Can that be turned off?
[12:04 PM PT] Gavin.Hird @grid.xmir.org:8002: it can but it will give a blurred world compared to the crisp Retian render, because the system compositor must upsample it
[12:05 PM PT] Ubit Umarov: on mac that 2x is os made
[12:05 PM PT] Gavin.Hird @grid.xmir.org:8002: All Macs with display have Retina displays
[12:05 PM PT] Ubit Umarov: on win the entire FS works on native disply res
[12:06 PM PT] Ubit Umarov: win has somethings to do rescale, but a total fail
[12:06 PM PT] Ubit Umarov: proper games ofc do it them selfs suing DLSS or amd equivalete
[12:06 PM PT] Gavin.Hird @grid.xmir.org:8002: another issue is that no Mac app can render directly to the disply
[12:07 PM PT] Gavin.Hird @grid.xmir.org:8002: it is always being composited to the display in a second pass by the system
[12:07 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Right that's what I remembered, there is something in between that can lead to higher usages
[12:07 PM PT] Ubit Umarov: in fact some games run a less than 1080 and display on 4K
[12:07 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Just never expected it to be over 30% more
[12:08 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Anyways I have some positive news to share
[12:08 PM PT] Ubit Umarov: you won lotery and going to share it all with us?
[12:08 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Spent last weekend doing upgrades to dotnet and rolled out a bunch of changes along
[12:09 PM PT] Ubit Umarov: pfff lotery would be more fun :p
[12:09 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Primarily something I been working on for a while after getting fed up with having to write stuff around to mitigate it
[12:09 PM PT] Vincent.Sylvester @hg.zetaworlds.com: https://i.imgur.com/a96j9BT.gif Have a look
[12:10 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Or in video form: https://www.youtube.com/watch?v=B48fj1KNmu0
[12:10 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Turns out with enough memory pointers and brute force the zoom level rendering for the map tiles doesn't have to be slow at all
[12:11 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Rendering 5000 images for 11 levels from 13000 individual tiles in 15 seconds
[12:12 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Kept the older style tile addition method around as well so doesn't have to do that each time a new region sends tiles, it's just for refreshing the whole set when you remove a bad tile
[12:13 PM PT] Vincent.Sylvester @hg.zetaworlds.com: While I was in there I also uncoupled the tile renderer from the main thread, so warp3d doesn't block the region console while it runs. Also means it doesn't block startup, just sends water and once the tile is done sends it
[12:14 PM PT] Vincent.Sylvester @hg.zetaworlds.com: This also meant I could set which renderer to use and operate on a copy of the scene data by reference which sped it up... not by much, but a little bit
[12:14 PM PT] Vincent.Sylvester @hg.zetaworlds.com: https://i.imgur.com/q9b8Gnh.gif
[12:15 PM PT] Kayaker Magic: I always disable map generation at startup, so startup is faster.
[12:15 PM PT] Kayaker Magic: Then send generate map commands in the middle of the night.
[12:15 PM PT] Andrew Hellershanks: Hello, Lyr
[12:15 PM PT] Lyr Lobo smiles and waves
[12:16 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Uncoupling it like this it doesn't block startup, but still renders on start. You can set the lesser MapImage at first and do Warp3D later
[12:16 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Only real downside is that being on own thread it cannot use more than 100% cpu
[12:17 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Though in case of warp3d that might be a good thing
[12:17 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Eventually it might be possible to reduce warp3d load with some more scary pointers
[12:18 PM PT] Vincent.Sylvester @hg.zetaworlds.com: One of the most complex regions I have took about 15 minutes to render a tile for a 2x2 var, so it definitely has room to improve, but given what it has to do for a render it may not be much
[12:19 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I plan to test this system in production for a while now to work out whatever glitches might still be in there and then Ubit gets the joy of dealing with another pull request :)
[12:19 PM PT] Ubit Umarov: see the lotery thing was a lot more fun
[12:19 PM PT] Ubit Umarov: :p
[12:20 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I only attract bugs, not lottery wins unfortunately ;)
[12:22 PM PT] Ubit Umarov: so usa is back in right time :p
[12:23 PM PT] Kayaker Magic hates daylight wastings time.
[12:24 PM PT] Andrew Hellershanks: I hate how the current time now means it gets dark by just after 5pm.
[12:26 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Hello darkness my old friend~
[12:26 PM PT] Andrew Hellershanks: Yup.
[12:26 PM PT] Andrew Hellershanks: :)
[12:27 PM PT] Andrew Hellershanks: Any other news related to OpenSimulator? Have you all registered for the upcoming OSCC?
[12:27 PM PT] orbert tatham: Ok, folks, sorry gotta go early today. See you next time maybe
[12:28 PM PT] Andrew Hellershanks: Orbert was in a hurry. I didn't get to say good bye before he went *poof*.
[12:29 PM PT] Lyr Lobo: Yes he had to dash
[12:29 PM PT] Lyr Lobo: Yes, registration is open and I'll share the schedule after we finish getting confirmations from the speakers. Our first training session is this Saturday
[12:30 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I did find another bug in the debug output for loghttp, but I have not quite found what exactly broke this time
[12:30 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Turning debug http in 6 and getting malformed requests
[12:30 PM PT] Vincent.Sylvester @hg.zetaworlds.com: log4net complaining as result
[12:30 PM PT] Gavin.Hird @grid.xmir.org:8002: not much to mention. working on viewer code pretty much every day now. lots of stupid updates because of compilers being upgraded
[12:31 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Godspeed to you :)
[12:31 PM PT] Gavin.Hird @grid.xmir.org:8002: not registered for OSCC. Weekends in December are hopeless for me
[12:31 PM PT] Ubit Umarov: yeah that is always fun gavin.Hird
[12:31 PM PT] Ubit Umarov: and worse undoc language changes..
[12:32 PM PT] Cuga.Rajal @rajal.org:9000: I'm registered.. I don;t have anything to present but I might be able to do a field trip
[12:32 PM PT] Cuga.Rajal @rajal.org:9000: what do you call them?
[12:32 PM PT] Lyr Lobo: Great!
[12:32 PM PT] Lyr Lobo: Events
[12:32 PM PT] Gavin.Hird @grid.xmir.org:8002: not so much that, but loads of ancient code that compilers barf over
[12:32 PM PT] Lyr Lobo: Events happen at a time convenient to you the week after the conference
[12:32 PM PT] Cuga.Rajal @rajal.org:9000: Lyr do you have a link handy to sign up for that?
[12:32 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Just pragma disable :)
[12:32 PM PT] Lyr Lobo: we also have zoom workshops for folks who need 60-90 minutes to demo or discuss design issues
[12:32 PM PT] Lyr Lobo: sure, one sec...
[12:33 PM PT] Lyr Lobo: https://conference.opensimulator.org/community-social-event-sign-up/
[12:33 PM PT] Andrew Hellershanks: I still need to register. I'll have to try and remember when to do that when I get back from my errand.
[12:33 PM PT] Lyr Lobo: Community Event signups
[12:33 PM PT] Cuga.Rajal @rajal.org:9000: great thx
[12:33 PM PT] Cuga.Rajal @rajal.org:9000: Vincent I'm wondering if that's related to the malformed request errors I was seeing a few weeeks ago
[12:34 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Far as I traced it's something in a try catch block that just writes logheader Exception
[12:34 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Think it's because the error contains square brackets that it tries to render a color maybe
[12:34 PM PT] Vincent.Sylvester @hg.zetaworlds.com: That's just guessing though
[12:35 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Normally that doesn't fail though
[12:36 PM PT] Cuga.Rajal @rajal.org:9000: I updated my operating system on the host computer for Opensim and I have not seen any of those malformed requests in about a month now
[12:36 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Well I caused them myself technically, but that should not break log out
[12:36 PM PT] Cuga.Rajal @rajal.org:9000: so that issue I was reporting might have been something weird on my system, or the offending client just stopped
[12:37 PM PT] Cuga.Rajal @rajal.org:9000: In my case the request had the effect of crashing the sunulator, requiring stopping and restarting
[12:37 PM PT] Cuga.Rajal @rajal.org:9000: bu tno issue since I updated
[12:38 PM PT] Cuga.Rajal @rajal.org:9000: so fingers crossed
[12:39 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I'll dig for the specific error, but it may just be so scuffed that it's not worth fixing
[12:40 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Ideally it doesn't get malformed nonsense over http ;)
[12:42 PM PT] Ubit Umarov: those things are on mantis?
[12:43 PM PT] Vincent.Sylvester @hg.zetaworlds.com: If I can find the broken part I'll mantis is. I don't like making tickets about fishing in the dark :)
[12:43 PM PT] Vincent.Sylvester @hg.zetaworlds.com: "something broke" isn't useful
[12:43 PM PT] Ubit Umarov: oh you didn't play icarus
[12:45 PM PT] Lyr Lobo smiles
[12:45 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I dismissed it originally cause I think I caused it myself. It technically hints at something being broken when it hits $[LOGHTTP]: Exception{exception} or something like that, which breaks log4net somehow
[12:46 PM PT] Vincent.Sylvester @hg.zetaworlds.com: That's as far as I looked into that cause it was caused by bad code elsewhere, technically broken, but only if you send it nonsense :)
[12:46 PM PT] Cuga.Rajal @rajal.org:9000: Vincent did the exception crash OS?
[12:46 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Thankfully not
[12:46 PM PT] Cuga.Rajal @rajal.org:9000: or just log something weird
[12:47 PM PT] Cuga.Rajal @rajal.org:9000: oh it crashed the logger but not the simulator?
[12:47 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I turned debug http in 6 to fetch why something was not giving me a response and it just barfed the moment it got the post request
[12:48 PM PT] Cuga.Rajal @rajal.org:9000: in dotnet6?
[12:48 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Said something about log4net renderer and $[LOGHTTP]: Exception{exception} if I recall correctly
[12:48 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Yes
[12:48 PM PT] Cuga.Rajal @rajal.org:9000: ahh.. so fixed in dotnet8?
[12:49 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Nah
[12:49 PM PT] Ubit Umarov: guess by dotnet8 you mean 0.9.3.0
[12:49 PM PT] Vincent.Sylvester @hg.zetaworlds.com: "debug http in 6" is a console command for outputting what it sees come in over http
[12:49 PM PT] Ubit Umarov: level 6
[12:50 PM PT] Cuga.Rajal @rajal.org:9000: ohhhhh
[12:50 PM PT] Ubit Umarov: a another name fo "useless spam"
[12:50 PM PT] Ubit Umarov: :p
[12:50 PM PT] Cuga.Rajal @rajal.org:9000: ok, I need to do that if my problem comes back. I'll ping you in IRC if that happens
[12:50 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Not entirely useless if used correctly
[12:50 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Can help debugging tools... for land... heh
[12:51 PM PT] Vincent.Sylvester @hg.zetaworlds.com: It's one of those things that probably comes up once in a blue moon when it gets totally bad data and you have http debug turned up like that, unlikely to happen in normal operation
[12:52 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Not sure why they are all string literals, but yeah probably not worth digging into
[12:52 PM PT] Vincent.Sylvester @hg.zetaworlds.com: It should have logged that error, but going through my log files I can't even find it, don't remember which simulator or robust it was even on heh
[12:53 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I only looked when it happened into code based on what it said it tried to output and found that loghttp exception exception thing, which could be the failure point or I am misled
[12:56 PM PT] Andrew Hellershanks: That's the issue sometimes. Is the exception handler catching a real issue or not.
[12:57 PM PT] Ubit Umarov: hard to say anything.. it can be jsut a simple typo or whatever
[12:57 PM PT] Vincent.Sylvester @hg.zetaworlds.com: It said something about log4net unable to render or something so my guess is just seeing something that messes with the way it does the colors since that was broken elsewhere before
[12:57 PM PT] Vincent.Sylvester @hg.zetaworlds.com: In any case not something I looked that far into cause like... pretty sure that it was just my bad code causing it
[1:00 PM PT] Andrew Hellershanks: Good luck tracking it down, Vincent.
[1:00 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I may just not bother if I don't think it'll come up more often
[1:01 PM PT] Vincent.Sylvester @hg.zetaworlds.com: It's in debug logging which normally is turned off, just use for debugging
[1:02 PM PT] Andrew Hellershanks: It is possible something changed and the debugging side hasn't caught up.
[1:05 PM PT] Lyr Lobo: Many thanks to you all for these great discussions! I must dash and will see you next week. *grins and waves*
[1:05 PM PT] Cuga.Rajal @rajal.org:9000: tc Lyr
[1:05 PM PT] Ubit Umarov: or as i said one of those very rare typos that did snick in...
[1:05 PM PT] Ubit Umarov: :p
[1:05 PM PT] Andrew Hellershanks: tc, Lyr.
[1:05 PM PT] Lyr Lobo waves and poofs
[1:05 PM PT] Andrew Hellershanks: I need to be going soon.
[1:05 PM PT] Ubit Umarov: cya Lyr :)
[1:07 PM PT] Ubit Umarov: wel any other opensim issue?
[1:07 PM PT] Andrew Hellershanks: That's usually my question. :)
[1:07 PM PT] Ubit Umarov: err so you are updating dayturn gavin?
[1:07 PM PT] Cuga.Rajal @rajal.org:9000: nothing new here
[1:08 PM PT] Gavin.Hird @grid.xmir.org:8002: as I said working on it pretty much every day now
[1:08 PM PT] Ubit Umarov: nice :)
[1:08 PM PT] Andrew Hellershanks: Sounds good.
[1:09 PM PT] Andrew Hellershanks: If there isn't anything else for today I'll call this meeting to a close.
[1:10 PM PT] Cuga.Rajal @rajal.org:9000: Take care everybody... see you soon
[1:10 PM PT] Andrew Hellershanks: Thank you all for coming. See you again next week.
