Chat log from the meeting on 2018-10-02

From OpenSimulator

(Difference between revisions)
Jump to: navigation, search
(Created page with " Category:Office Hour Logs")
 
 
Line 1: Line 1:
 +
[11:09] Ubit Umarov: so what bugs and simliar entities had you found on opensim during this last week ?
 +
<br>[11:11] Gavin.Hird @grid.xmir.org:8002: I like the ClientOnParcelGodMark commit, but I did not quite like how the land flags was set after the last commit, so I tightened a bit
 +
<br>[11:12] Ubit Umarov: yeah i just used the default..  fell free to sugest other set on that mantis
 +
<br>[11:13] Gavin.Hird @grid.xmir.org:8002: IMO when you set a parcel to a grid account, you don't really want it to be set for anyone to build on it
 +
<br>[11:13] Bill Blight: I  made a video to debunk some of the "slow mesh" rezzing bs, this building is all mesh, the furniture is all mesh and this video is in real time with an empty viewer cache ...
 +
<br>[11:13] Bill Blight: https://youtu.be/YFwy_XCyegw
 +
<br>[11:13] Ubit Umarov: yeah, on other hand a god should know better and fix it :)
 +
<br>[11:14] Ubit Umarov: well not god.. admin..
 +
<br>[11:14] Gavin.Hird @grid.xmir.org:8002: how about the admin opening it up rather :-)
 +
<br>[11:14] Ubit Umarov: bad choice of work, not hard to stop using
 +
<br>[11:14] Ubit Umarov: now..
 +
<br>[11:16] Andrew.Hellershanks @hg.osgrid.org: I started going through the ini files and looking for changes in the Allow_ statements for OSSL functions so I can add information to the OSSL wiki pages that will state when functions were added.
 +
<br>[11:17] Arielle.Popstar @hg.osgrid.org: nice  video Bill.  What was its point?
 +
<br>[11:17] Bill Blight: The point that people claim mesh loads slow in .9x
 +
<br>[11:17] Bill Blight: It doesn't
 +
<br>[11:17] Arielle.Popstar @hg.osgrid.org: they are
 +
<br>[11:18] Arielle.Popstar @hg.osgrid.org: go to event  plaza  when there are  30 avatars  there
 +
<br>[11:18] Arielle.Popstar @hg.osgrid.org: thats the point
 +
<br>[11:18] Bill Blight: I was there friday,  had no issues
 +
<br>[11:18] Arielle.Popstar @hg.osgrid.org: lucky you :)
 +
<br>[11:18] Ubit Umarov: Some ppl had issues...  viewers at 2fps
 +
<br>[11:19] Ubit Umarov: not much can be done with several avatars there with complexit > 300 000
 +
<br>[11:19] Bill Blight: I got a solid 25fps the whole time ...
 +
<br>[11:19] Arielle.Popstar @hg.osgrid.org: how long did it take  all mesh  avatars  to fully load for you Bill?
 +
<br>[11:19] Bill Blight: some were over 500k
 +
<br>[11:19] Bill Blight: about 2 minutes
 +
<br>[11:20] Ubit Umarov: yeap but many don't have a GPU like youes
 +
<br>[11:20] Leighton.Marjoram @grid.outandproud.life:8002: Bill I agree with your statement that people mis-attribute the problem to the server version 0.8 v 0.9
 +
<br>[11:20] Bill Blight: my GPU is considered low end for gamers these days ... LOL
 +
<br>[11:20] Ubit Umarov: it does not matter how fast regions can send info, if viewers can't handle it
 +
<br>[11:20] Arielle.Popstar @hg.osgrid.org: what gpu?
 +
<br>[11:20] Bill Blight: low end 1060
 +
<br>[11:20] Arielle.Popstar @hg.osgrid.org: same as  mine
 +
<br>[11:20] Ubit Umarov: in fact fast send may even make it worse, like happens with lludp
 +
<br>[11:21] Andrew.Hellershanks @hg.osgrid.org: That is a much higher level card than what I have.
 +
<br>[11:21] Gavin.Hird @grid.xmir.org:8002: I agree fast send can make it worse
 +
<br>[11:21] Ubit Umarov: when i allow 3000 on testing. i loose terrain patchs ocasionaly
 +
<br>[11:21] Bill Blight: yes packets tend to trip over themselves
 +
<br>[11:21] Gavin.Hird @grid.xmir.org:8002: there is only so much that can be done in a frame
 +
<br>[11:22] Ubit Umarov: viwers shoule send us http throttle levels also
 +
<br>[11:22] Dev Random: Is it true that all users are constrained by the slowest user? If I have a super-slow connection, will it affect scene updates for all, to keep everyone in sync?
 +
<br>[11:22] Ubit Umarov: so users can tune as do with lludp
 +
<br>[11:22] Ubit Umarov: singu does have a http control. but only internal
 +
<br>[11:23] Arielle.Popstar @hg.osgrid.org: i find  sing slower loading
 +
<br>[11:23] Bill Blight: well I find position updates are more affected by slow users than rendering
 +
<br>[11:23] Arielle.Popstar @hg.osgrid.org: what os you using  Bill?
 +
<br>[11:24] Andrew.Hellershanks @hg.osgrid.org: Ubit, which setting were you changing to 3,000?
 +
<br>[11:24] Ubit Umarov: during last SG conc at event, really many had the viewers totally overload
 +
<br>[11:24] Ubit Umarov: and of course blamed the region :)
 +
<br>[11:25] Bill Blight: Windows on my PC , linux on the servers
 +
<br>[11:25] Arielle.Popstar @hg.osgrid.org: well if  the  viewer  works  fine everywhere else...
 +
<br>[11:25] Ubit Umarov: lludp  bandwith andrew
 +
<br>[11:25] Arielle.Popstar @hg.osgrid.org: 7, 10?
 +
<br>[11:25] Bill Blight: 10
 +
<br>[11:26] Arielle.Popstar @hg.osgrid.org: why then 2 minutes for you and  12 for me?
 +
<br>[11:26] Ubit Umarov: Arielle plz dont' be silly..  "everywhere" does not have >50 avatars. many very high complexity
 +
<br>[11:26] Gavin.Hird @grid.xmir.org:8002: where in the code is lludp  bandwith set?
 +
<br>[11:26] Ubit Umarov: viewers set it gavin :p
 +
<br>[11:26] Bill Blight: there is a config option for the buffer I think
 +
<br>[11:26] Arielle.Popstar @hg.osgrid.org: majority of users  wouldnt get the differewnce  Ubit
 +
<br>[11:26] Ubit Umarov: well some of the llupd options are dead..
 +
<br>[11:26] Bill Blight: but it get's limited by the max the OS will allow
 +
<br>[11:26] Gavin.Hird @grid.xmir.org:8002: right, actually it has changed a bit in the LL code
 +
<br>[11:27] Gavin.Hird @grid.xmir.org:8002: it used to be 3000
 +
<br>[11:27] Gavin.Hird @grid.xmir.org:8002: now it is higher
 +
<br>[11:27] Ubit Umarov: in regions code is somewhere on lludpserver  or around :)
 +
<br>[11:27] Bill Blight: firewall, and operating systems have their own UDP throttles ..
 +
<br>[11:27] Arielle.Popstar @hg.osgrid.org: and  routers
 +
<br>[11:27] Ubit Umarov: well FS cries when you set more that 1500 ( warns )
 +
<br>[11:27] Ubit Umarov: and it does fail with 3000
 +
<br>[11:28] Gavin.Hird @grid.xmir.org:8002: my router is supposed to manage 930 Mbut/sec UDP across the NAT
 +
<br>[11:28] Gavin.Hird @grid.xmir.org:8002: Mbit/sec*
 +
<br>[11:29] Arielle.Popstar @hg.osgrid.org: your  router  is  a  higher end  one though
 +
<br>[11:29] Gavin.Hird @grid.xmir.org:8002: it cost $55
 +
<br>[11:29] Arielle.Popstar @hg.osgrid.org: i looked  at the  specs  and reviews
 +
<br>[11:29] Arielle.Popstar @hg.osgrid.org: i know, i saw
 +
<br>[11:29] Arielle.Popstar @hg.osgrid.org: too bad  it had no wifi
 +
<br>[11:30] Andrew.Hellershanks @hg.osgrid.org: I just noticed I had FS set to only 1000 bandwidth. I've upped that now to 1500.
 +
<br>[11:30] Gavin.Hird @grid.xmir.org:8002: they have an excellent range of wifi kit to help you with that
 +
<br>[11:30] Ubit Umarov: im using 1500 also ( the max they advice )
 +
<br>[11:31] Bill Blight: Even second life recommends you don't use wifi if you can avoid it though
 +
<br>[11:32] Andrew.Hellershanks @hg.osgrid.org: Hm... in Sing I have maximum bandwidth at 2000 and texture bandwidth at 3000 (the defaults for that viewer).
 +
<br>[11:32] Arielle.Popstar @hg.osgrid.org: i wonder  if that issue  wishbringer  is mentioning  might  be  related
 +
<br>[11:32] Gavin.Hird @grid.xmir.org:8002: SL viewer is now set to 2100 kbps as default and max is 10000
 +
<br>[11:32] Bill Blight: wifi on non MIMO routers is only a half duplex connection , it does not send and receive at the some t ime ..
 +
<br>[11:32] Bill Blight: same
 +
<br>[11:32] Ubit Umarov: and depends on the power yr microwave oven is :P
 +
<br>[11:33] Bill Blight: that too
 +
<br>[11:33] Bill Blight: and portable phones
 +
<br>[11:33] Bill Blight: LOL
 +
<br>[11:33] Ubit Umarov: and blutooth
 +
<br>[11:33] Arielle.Popstar @hg.osgrid.org: i see  .2% packet loss here
 +
<br>[11:33] Ubit Umarov: and how many ppl around yoy also have wifi
 +
<br>[11:34] Bill Blight: quick interlude, "Neighbor kept getting kicked off his wifi, found out everytime his phone ringed ... His phone base station was next to the wifi/router"
 +
<br>[11:34] Bill Blight: I have 0.0% packet loss
 +
<br>[11:34] Ubit Umarov: well they can coexit
 +
<br>[11:34] Bill Blight: yeah they can
 +
<br>[11:34] Bill Blight: 68ms ping
 +
<br>[11:34] Gavin.Hird @grid.xmir.org:8002: 0.0% packet loss and ping time around 160 ms
 +
<br>[11:34] Ubit Umarov: but not if the TX of one saturates the rx of the other :)
 +
<br>[11:35] Arielle.Popstar @hg.osgrid.org: 34 ms  for me
 +
<br>[11:35] Arielle.Popstar @hg.osgrid.org: 17
 +
<br>[11:35] Bill Blight: and this server is about 1500 miles from me
 +
<br>[11:35] Arielle.Popstar @hg.osgrid.org: ohv montreal?
 +
<br>[11:35] Bill Blight: OVH yes
 +
<br>[11:35] Arielle.Popstar @hg.osgrid.org: in montreal?
 +
<br>[11:35] Bill Blight: yes
 +
<br>[11:36] Arielle.Popstar @hg.osgrid.org: same  as  osgrid servers
 +
<br>[11:36] Arielle.Popstar @hg.osgrid.org: thats  like 20 ms for
 +
<br>[11:36] Bill Blight: 3 servers there 1 in France, and 1 in Germay ..
 +
<br>[11:37] Arielle Popstar: Ubit, did you  not do a commit  a  couple months ago to set  region z at -100?
 +
<br>[11:38] Arielle Popstar: i was  sure i saw a  commit note  to that effect
 +
<br>[11:38] Ubit Umarov: set sure i did not :)
 +
<br>[11:38] Gavin.Hird @grid.xmir.org:8002: he did
 +
<br>[11:38] Bill Blight: you did
 +
<br>[11:38] Bill Blight: LOL
 +
<br>[11:38] Bill Blight: http://opensimulator.org/viewgit/?a=commit&p=opensim&h=39c8db8eb1cbfdefc120b1bf5f5aef45f9635509
 +
<br>[11:38] Ubit Umarov: i did relaz ubode to allow it
 +
<br>[11:39] Ubit Umarov: relax
 +
<br>[11:39] Ubit Umarov: i did not set regions at z = -100 bill :p
 +
<br>[11:39] Bill Blight: this is true
 +
<br>[11:39] Bill Blight: I stand corrected
 +
<br>[11:40] Ubit Umarov: but viewers will do a mess
 +
<br>[11:40] Arielle Popstar: so that will allow  the titanic  to sink again?
 +
<br>[11:41] Ubit Umarov: titanic is at -4000m or more :p
 +
<br>[11:41] Gavin.Hird @grid.xmir.org:8002: most new viewer builds will support z=-200, older builds will not
 +
<br>[11:41] Arielle Popstar: heh  yes  but they will settle for it to  be  at -80 i think
 +
<br>[11:41] Ubit Umarov: last time i check light is all broken below 0
 +
<br>[11:41] Dev Random: heh.. I remember having that convo with Mr Blue a few years ago...
 +
<br>[11:41] Bill Blight: current FS still looses shaders if you go below 0
 +
<br>[11:42] Gavin.Hird @grid.xmir.org:8002: There are new code in the LL Cougar build that fix both lighting and shadow under water
 +
<br>[11:42] Gavin.Hird @grid.xmir.org:8002: it has been released in any viewer yet
 +
<br>[11:42] Ubit Umarov: well ll does not allow <0 terrain
 +
<br>[11:42] Bill Blight: but did FS force disable those for Opensim like other stuff they have done
 +
<br>[11:42] Gavin.Hird @grid.xmir.org:8002: it allows -200
 +
<br>[11:43] Ubit Umarov: not sure how
 +
<br>[11:43] Gavin.Hird @grid.xmir.org:8002: which is why newer viewer builds got updated
 +
<br>[11:43] Ubit Umarov: and max ?
 +
<br>[11:43] Gavin.Hird @grid.xmir.org:8002: max height?
 +
<br>[11:43] Ubit Umarov: yeap
 +
<br>[11:43] Gavin.Hird @grid.xmir.org:8002: I don't recall but maybe 6k meters
 +
<br>[11:44] Ubit Umarov: nahh
 +
<br>[11:44] Ubit Umarov: that could smoke the terrain encoder
 +
<br>[11:44] Gavin.Hird @grid.xmir.org:8002: I'll have to check the source
 +
<br>[11:45] Bill Blight: looks like I'm gonna have to build me a new viewer and see what has changed in the last month and a half
 +
<br>[11:45] Ubit Umarov: and raw import is 256m range only or 512m
 +
<br>[11:45] Gavin.Hird @grid.xmir.org:8002: const F32 REGION_HEIGHT_METERS = 4096.f;
 +
<br>[11:46] Dev Random: Ubit, how much would a new Terrain encoding break physics? e.g. If it were a "base height" and a collection of offsets instead of just a scaling?
 +
<br>[11:46] Ubit Umarov: internally terrain heighs are floats now
 +
<br>[11:47] Ubit Umarov: the eoncoding i was talking is sending to viewers
 +
<br>[11:47] Dev Random: so probably no effect, except for maybe not allowing big negatives....?
 +
<br>[11:48] Ubit Umarov: its limits to keep things coerent to viewers
 +
<br>[11:48] Ubit Umarov: not nice to let a physical thing get out of bounds
 +
<br>[11:48] Dev Random: ha.. yeah.. got that...
 +
<br>[11:49] Ubit Umarov: but there are resolution issues of course
 +
<br>[11:49] Ubit Umarov: floats have only +- 6 digits..
 +
<br>[11:49] Ubit Umarov: than is 1mm error on 1000m
 +
<br>[11:50] Ubit Umarov: if you decide to put things at 10 000m  thats 1cm error
 +
<br>[11:50] Ubit Umarov: ( also a isssue on region s size )
 +
<br>[11:50] Dev Random: yeah .. building on the moon might have some repercussions....
 +
<br>[11:51] Ubit Umarov: yeap :)
 +
<br>[11:51] Gavin.Hird @grid.xmir.org:8002: so you get a drafty skybox if you put at 10k meters
 +
<br>[11:52] Ubit Umarov: you may notice issue putting 2 linksets side by side
 +
<br>[11:52] Ubit Umarov: or editing a linkset
 +
<br>[11:52] Ubit Umarov: so yes there are limits..
 +
<br>[11:52] Dev Random: I'd wondered how much more flexibility might be wanted due to Vars... With more CPU power every year, Vars will get bigger, and people may want more resolution..
 +
<br>[11:52] Ubit Umarov: and makes no sense moving from floats to doubles
 +
<br>[11:52] Ubit Umarov: :)
 +
<br>[11:53] Ubit Umarov: this is a limit also on vars size
 +
<br>[11:53] Ubit Umarov: guess 1cm is the max error we can tolerate
 +
<br>[11:55] Arielle Popstar: whats the limit now?
 +
<br>[11:55] Ubit Umarov: ( i mean errors just from the numbers representation..  things get a bit worse if there is math on top of that )
 +
<br>[11:55] Ubit Umarov: 8km
 +
<br>[11:55] Ubit Umarov: but that is not that pratical
 +
<br>[11:55] Arielle Popstar: 30x30?
 +
<br>[11:55] Ubit Umarov: 2km seems workable
 +
<br>[11:56] Andrew Hellershanks: It is 5 minutes to the hour. Any other topics for today?
 +
<br>[11:56] Ubit Umarov: possible 1k a better choice :)
 +
<br>[11:56] Arielle Popstar: pfft
 +
<br>[11:56] Arielle Popstar: go back to megas  then
 +
<br>[11:56] Andrew Hellershanks: :)  You would have to bring back some code from the dead.
 +
<br>[11:57] Arielle Popstar: yeah  i dunno why they removed  it
 +
<br>[11:57] Arielle Popstar: megas  had their uses
 +
<br>[11:58] Andrew Hellershanks: IIRC, support for megas was a bit hackish then the vars kind of took over.
 +
<br>[11:59] Ubit Umarov: i was talking about overall regions a nd viewers performance on those limits ..
 +
<br>[11:59] Bill Blight: As far as I remember Megas used even more resources than vars ...
 +
<br>[11:59] Arielle Popstar: very little more
 +
<br>[11:59] Ubit Umarov: and on larger sizes better have the regions isolated
 +
<br>[11:59] Ubit Umarov: or viewers may smoke
 +
<br>[11:59] Gavin.Hird @grid.xmir.org:8002: 1km regions seems to be a reasonable limit for the current viewer code
 +
<br>[12:00] Bill Blight: I did a cluster of 9 16x16 regions , viewer was not happy but they worked
 +
<br>[12:00] Ubit Umarov: and you cna pack them side by side ( with the viewer restriction of just one per side )
 +
<br>[12:01] Ubit Umarov: well you cna limit the view range to 1st neib also etc
 +
<br>[12:01] Ubit Umarov: or should..
 +
<br>[12:02] Ubit Umarov: and of course.. all users should use minimal view range
 +
<br>[12:02] Ubit Umarov: that is always valid
 +
<br>[12:02] Gavin.Hird @grid.xmir.org:8002: 64 meters?
 +
<br>[12:02] Andrew Hellershanks: I know of one person who has a a 1x1 and a 2x2 (separated from each other) both up against var. I haven't heard of them experiencing any ill effects, althought it isn't recommended.
 +
<br>[12:02] Ubit Umarov: acording to the needs
 +
<br>[12:02] Arielle Popstar: 32
 +
<br>[12:03] Ubit Umarov: 1000m WILL NOT WORK in many cases
 +
<br>[12:03] Bill Blight: as long as there is only one per side that is fine Andrew
 +
<br>[12:03] Arielle Popstar: in fs
 +
<br>[12:03] Andrew Hellershanks: Bill, the 1x1 and 2x2 are on the same side of the var.
 +
<br>[12:03] Arielle Popstar: just need betterglasses
 +
<br>[12:03] Bill Blight waits for the crash
 +
<br>[12:03] Bill Blight: it is a viewer side issue
 +
<br>[12:03] Andrew Hellershanks: Bill, none so far.
 +
<br>[12:03] Ubit Umarov: view ranfe is a important download and viewer loac control
 +
<br>[12:03] Bill Blight: not an opensim one
 +
<br>[12:03] Andrew Hellershanks: ah, ok
 +
<br>[12:04] Ubit Umarov: of course, does not help on a cluster of avatars with high complexity
 +
<br>[12:04] Gavin.Hird @grid.xmir.org:8002: The Kokua code supports 8m draw distance
 +
<br>[12:04] Andrew Hellershanks: I keep view distance down around 64m to 96m
 +
<br>[12:04] Ubit Umarov: for that..  get a gtx 1080 or better :p
 +
<br>[12:04] Bill Blight: if they try crossing when both sims are in view on the same side the viewer get's confused what region they are crossing into
 +
<br>[12:04] Arielle Popstar: complexity higher then 3000?
 +
<br>[12:04] Ubit Umarov: missed a 0 :p
 +
<br>[12:04] Arielle Popstar: :)
 +
<br>[12:04] Ubit Umarov: err 2 zeros lol
 +
<br>[12:05] Andrew Hellershanks: Ubit, I wouldn't spend that kind of money on a video card.
 +
<br>[12:05] Ubit Umarov: 300000 or even 100 000 if ist a big party :)
 +
<br>[12:05] Gavin.Hird @grid.xmir.org:8002: you can use it for mining crypto
 +
<br>[12:06] Bill Blight: I had an avatar land at one of my parties with 999,999 complexity and everyone came to a grinding halt
 +
<br>[12:06] Arielle Popstar: what you could  do is get rid of the lower 2 LOD settings  on any mesh clothing  items  in the  db
 +
<br>[12:06] Arielle Popstar: or  auto set them to 0
 +
<br>[12:06] Ubit Umarov: btw there is a experimental  objects culling option on regions
 +
<br>[12:06] Ubit Umarov: it does not even send objects out of view to viewers
 +
<br>[12:06] Bill Blight: How do you tell a mesh clothing from a mesh building?  as you could wear a mesh building if you wanted
 +
<br>[12:06] Ubit Umarov: ( without it, all is sent eventually =
 +
<br>[12:07] Bill Blight: mesh is mesh to the DB
 +
<br>[12:07] Arielle Popstar: mesh clothing  has  rigging  info
 +
<br>[12:07] Gavin.Hird @grid.xmir.org:8002: the mesh building is probalby not rigged
 +
<br>[12:07] Arielle Popstar: a  building hopefully not
 +
<br>[12:07] Ubit Umarov: easy bill  a mesh building does not have 130 000 triangles :)
 +
<br>[12:07] Bill Blight: that is true
 +
<br>[12:07] Bill Blight: most sims don't have a complexity of over 999,9999
 +
<br>[12:07] Bill Blight: LOL
 +
<br>[12:07] Andrew Hellershanks: Ubit, that would depend on the building. ;)
 +
<br>[12:07] Gavin.Hird @grid.xmir.org:8002: actually I have seen mesh building exceeding 250k tris
 +
<br>[12:07] Bill Blight: yeah have
 +
<br>[12:08] Gavin.Hird @grid.xmir.org:8002: not in one piece, but the combined building
 +
<br>[12:08] Ubit Umarov: and buildings are larger... parts can't get out of view etc
 +
<br>[12:08] Ubit Umarov: a avatar.. bahh
 +
<br>[12:08] Gavin.Hird @grid.xmir.org:8002: plus add the furniture and plants
 +
<br>[12:08] Bill Blight: just making a point that most well built buildings have less complexity than some of these mesh avatars and clothing
 +
<br>[12:08] Gavin.Hird @grid.xmir.org:8002: they usually do
 +
<br>[12:08] Arielle Popstar: well  another  fix  is  to  put  a better  system avatar  up
 +
<br>[12:08] Gavin.Hird @grid.xmir.org:8002: but the interior can be excessive
 +
<br>[12:08] Arielle Popstar: then we  dont  need  all these  mesh bodies
 +
<br>[12:08] Ubit Umarov: that is what LL should had done
 +
<br>[12:08] Bill Blight: this is true
 +
<br>[12:09] Gavin.Hird @grid.xmir.org:8002: foliage
 +
<br>[12:09] Ubit Umarov: and NO downloads
 +
<br>[12:09] Ubit Umarov: like old ruth
 +
<br>[12:09] Ubit Umarov: all have them on disk
 +
<br>[12:09] Bill Blight: we had a plant that that one leaf was over 800k display weight and it had 24 leaves
 +
<br>[12:09] Ubit Umarov: but NO..  tghey made the crap mesh ones
 +
<br>[12:09] Arielle Popstar: so  why cant  that be  changed in opensim?
 +
<br>[12:09] Ubit Umarov: all diferent.. all needed to be downloaded one by one
 +
<br>[12:10] Ubit Umarov: wlel basicly bc we don't have a avatar :)
 +
<br>[12:10] Arielle Popstar: we  have  ruth
 +
<br>[12:10] Ubit Umarov: but thats is viewer side
 +
<br>[12:10] Arielle Popstar: or we  can buy one  from  s/l
 +
<br>[12:10] Bill Blight: the default avatar info comes with the viewer
 +
<br>[12:10] Ubit Umarov: from the complains i get ruth 2.0 seems 2 heavy
 +
<br>[12:10] Arielle Popstar: viewers are  opensource
 +
<br>[12:11] Arielle Popstar: no  she  isnt
 +
<br>[12:11] Bill Blight: ruth is not too heavy ... but a lot of the clothing being made for it is ..
 +
<br>[12:11] Arielle Popstar: 20k
 +
<br>[12:11] Ubit Umarov: but this model of each one having its own mesh avatar does not hold
 +
<br>[12:11] Arielle Popstar: less
 +
<br>[12:11] Ubit Umarov: unless you are alone on a region
 +
<br>[12:11] Ubit Umarov: i mean own unique mesh avatar
 +
<br>[12:11] Bill Blight: SL is a lag fest with all the mesh avatars
 +
<br>[12:11] Bill Blight: and people want to bring that to opensim?
 +
<br>[12:11] Ubit Umarov: that was a very bad move by LL imo
 +
<br>[12:12] Arielle Popstar: talking  a better  system  mesh Bill
 +
<br>[12:12] Bill Blight: LL has millions to throw at bandwidth and servers
 +
<br>[12:12] Bill Blight: and still can't make it NOT lag
 +
<br>[12:12] Gavin.Hird @grid.xmir.org:8002: according to HGB there are over 250k OpenSim regions, but only about 32k active users. So the chances you are alone in the region is pretty high
 +
<br>[12:12] Andrew Hellershanks: I think it was seen as a natural extension to having mesh objects.
 +
<br>[12:12] Ubit Umarov: as i said they should added a higher resolution ruth even making it optional on graphics settings
 +
<br>[12:12] Ubit Umarov: that should had been the proper way
 +
<br>[12:13] Ubit Umarov: but mesh avatars sell...  uploads cost... so.. well... we understand..
 +
<br>[12:14] Bill Blight: LL has a physics engine that costs millions  and millions of dollars, they have millions to throw at the rendering engine, as they are getting ready to change to DirectX and Metal, but we are supposed to just pull one out of thin air?
 +
<br>[12:14] Ubit Umarov: and of course not nothing can be done
 +
<br>[12:14] Ubit Umarov: Uniqueness is something all ladies want
 +
<br>[12:14] Ubit Umarov: ( and not only ladies :) )
 +
<br>[12:15] Gavin.Hird @grid.xmir.org:8002: actually LL only has about $150 to throw at bandwidth and server per region - in reality even less as there are a large number of non-paid regions
 +
<br>[12:15] Ubit Umarov: isn't metal a apple thing only ? mb Vulcan ?
 +
<br>[12:15] Bill Blight: they are going Metal for Mac and DirectX for windows
 +
<br>[12:16] Bill Blight: this is why they discontinued Linux support
 +
<br>[12:16] Gavin.Hird @grid.xmir.org:8002: metal is a Apple only tehc as directx is Microsoft only
 +
<br>[12:16] Ubit Umarov: no they did it bc vivox i think
 +
<br>[12:16] Ubit Umarov: vivox drop linux
 +
<br>[12:16] Andrew Hellershanks: Bill, I can't use the SL viewer any more.
 +
<br>[12:16] Gavin.Hird @grid.xmir.org:8002: Linux is 0.4% of their users
 +
<br>[12:16] Gavin.Hird @grid.xmir.org:8002: which is why they dropped it
 +
<br>[12:16] Bill Blight: That may have been a side reason but if you read their developer blog , that is not what Ebbe said
 +
<br>[12:17] Ubit Umarov: and more will drop linux
 +
<br>[12:17] Ubit Umarov: they made FS the linux suport team no ?
 +
<br>[12:18] Ubit Umarov: well i know we can't be following ll viewer all the time
 +
<br>[12:18] Bill Blight: well if there is going to be a linux one it is going to be third party
 +
<br>[12:18] Ubit Umarov: for us vulcan is a better option i guess
 +
<br>[12:19] Gavin.Hird @grid.xmir.org:8002: They are changing from Google breakpad to Bugsplat, and that is another paid for license
 +
<br>[12:19] Bill Blight: vulcan is the progression from OpenGL so yeah makes more sense
 +
<br>[12:19] Gavin.Hird @grid.xmir.org:8002: I wonder if FS will be able to absorb it
 +
<br>[12:19] Bill Blight: I'm thinking they are going to pull a lot of the opensource viewer code and make it in house, but that is just my feeling
 +
<br>[12:19] Ubit Umarov: it depends on how deep and specific the code changes are
 +
<br>[12:20] Ubit Umarov: if they optimize all for direct  well a pain to do on other
 +
<br>[12:20] Gavin.Hird @grid.xmir.org:8002: A Bugsplat license for 50,000 annual crashes is $139 / month
 +
<br>[12:21] Ubit Umarov: how is the kokuaOS project as a viewer for opensim gavin ?
 +
<br>[12:22] Bill Blight: No windows that's how it is ... :P
 +
<br>[12:22] Gavin.Hird @grid.xmir.org:8002: I am working on the macOS version almost daily, but I don't have anybody to do the Win and Linux builds, so they are frozen in time
 +
<br>[12:23] Arielle Popstar: werent the  IMS  people  looking into that?
 +
<br>[12:23] Gavin.Hird @grid.xmir.org:8002: really
 +
<br>[12:23] Ubit Umarov: better hide Arielle
 +
<br>[12:23] Ubit Umarov: lol
 +
<br>[12:23] Bill Blight: HAHAHAHA
 +
<br>[12:23] Arielle Popstar: sitting  with bar between us :)
 +
<br>[12:24] Gavin.Hird @grid.xmir.org:8002: fact is, keeping the viewer current is a lot of work
 +
<br>12:25] Ubit Umarov: actually it is been a while i ear things from IMA
 +
<br>[12:25] Ubit Umarov: how is tit going Selby ?
 +
<br>[12:26] Selby.Evans @grid.kitely.com:8002: Slow -- Working on an edu viewer now
 +
<br>[12:26] Ubit Umarov: i had a nice chat with lisa,  but months (?) ago now :)
 +
<br>[12:26] Selby.Evans @grid.kitely.com:8002: a bit like I described in strawman articles
 +
<br>[12:27] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/search?q=strawman
 +
<br>[12:27] Ubit Umarov: thanks selby.Evans
 +
<br>[12:28] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/01/edu-ima-ww-scenegate-updating-straw-man.html
 +
<br>[12:28] Selby.Evans @grid.kitely.com:8002: use this one
 +
<br>[12:28] Selby.Evans @grid.kitely.com:8002: A straw-man model for a virtual world viewer was presented with the objective of eliciting comments and suggestions for improvement.  Comments and suggestions are here used to produce a more refined model.  This article is explicitly licensed as public domain.
 +
<br>[12:29] Selby.Evans @grid.kitely.com:8002: First is minimanlist -- for new users
 +
<br>[12:30] Selby.Evans @grid.kitely.com:8002: Kitely is also developing grid- customized viewers for its VPG project
 +
<br>[12:30] Leighton.Marjoram @grid.outandproud.life:8002: What viewer code did these suggestions come from the article says "Straw-man strategy: You put it up and everyone tells you what is wrong with it."
 +
<br>[12:31] Leighton.Marjoram @grid.outandproud.life:8002: What was put up ??
 +
<br>[12:31] Bill Blight: Any actual code releases?
 +
<br>[12:31] Leighton.Marjoram @grid.outandproud.life:8002: Or is it a wish list
 +
<br>[12:31] Arielle Popstar: whats Kitely VPG?
 +
<br>[12:31] Bill Blight: Virtual Private Grid
 +
<br>[12:32] Arielle Popstar: kk
 +
<br>[12:32] Selby.Evans @grid.kitely.com:8002: No code releases yet -- Should have one soon -ww when released by cooperating company
 +
<br>[12:32] Selby.Evans @grid.kitely.com:8002: My articles are wish-lists
 +
<br>[12:34] Arielle Popstar: did  everyone see what  Dahlia was  doing  with web viewer?
 +
<br>[12:34] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/07/2018-edu-biz-kitely-will-offer-private.html
 +
<br>[12:34] Leighton.Marjoram @grid.outandproud.life:8002: Certainly did Arielle, looking forward to seeing where that goes.
 +
<br>[12:34] Andrew Hellershanks: I haven't. I know Cinder Roxley had been working on a modular viewer.
 +
<br>[12:35] Arielle Popstar: https://www.youtube.com/watch?v=1XNdOOVLbJs&feature=youtu.be
 +
<br>[12:35] Gavin.Hird @grid.xmir.org:8002: Yes I saw, but also noticed there were no dynamic updates which kindoff is the major feature of SL/Opensim compared to other virtual environments
 +
<br>[12:35] Arielle Popstar: you mean after  editing  something?
 +
<br>[12:36] Selby.Evans @grid.kitely.com:8002: I work with 2 web-worlds -- just a min
 +
<br>[12:36] Gavin.Hird @grid.xmir.org:8002: no, in the scene
 +
<br>[12:36] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/04/2018-ww-web-worlds-3d-virtual-worlds.html
 +
<br>[12:36] Selby.Evans @grid.kitely.com:8002: Cybaloounge is talking students now
 +
<br>[12:37] Selby.Evans @grid.kitely.com:8002: Two browser-based 3D virtual worlds are now in public beta.  Both are ready for initial use.  CybaLOUNGE  has an initial educational project in the build stage.  3DWebWorldz is presently seeking museum content as part of its preparation for educational services.
 +
<br>[12:37] Selby.Evans @grid.kitely.com:8002: web links in the article
 +
<br>[12:37] Leighton.Marjoram @grid.outandproud.life:8002: ty Selby
 +
<br>[12:37] Gavin.Hird @grid.xmir.org:8002: dynamic updates is usually what makes other engines tank in performance compared to the optimized content they are made for
 +
<br>[12:38] Gavin.Hird @grid.xmir.org:8002: static/optimized
 +
<br>[12:38] Arielle Popstar: ok
  
 
[[Category:Office Hour Logs]]
 
[[Category:Office Hour Logs]]

Latest revision as of 11:44, 2 October 2018

[11:09] Ubit Umarov: so what bugs and simliar entities had you found on opensim during this last week ?
[11:11] Gavin.Hird @grid.xmir.org:8002: I like the ClientOnParcelGodMark commit, but I did not quite like how the land flags was set after the last commit, so I tightened a bit
[11:12] Ubit Umarov: yeah i just used the default.. fell free to sugest other set on that mantis
[11:13] Gavin.Hird @grid.xmir.org:8002: IMO when you set a parcel to a grid account, you don't really want it to be set for anyone to build on it
[11:13] Bill Blight: I made a video to debunk some of the "slow mesh" rezzing bs, this building is all mesh, the furniture is all mesh and this video is in real time with an empty viewer cache ...
[11:13] Bill Blight: https://youtu.be/YFwy_XCyegw
[11:13] Ubit Umarov: yeah, on other hand a god should know better and fix it :)
[11:14] Ubit Umarov: well not god.. admin..
[11:14] Gavin.Hird @grid.xmir.org:8002: how about the admin opening it up rather :-)
[11:14] Ubit Umarov: bad choice of work, not hard to stop using
[11:14] Ubit Umarov: now..
[11:16] Andrew.Hellershanks @hg.osgrid.org: I started going through the ini files and looking for changes in the Allow_ statements for OSSL functions so I can add information to the OSSL wiki pages that will state when functions were added.
[11:17] Arielle.Popstar @hg.osgrid.org: nice video Bill. What was its point?
[11:17] Bill Blight: The point that people claim mesh loads slow in .9x
[11:17] Bill Blight: It doesn't
[11:17] Arielle.Popstar @hg.osgrid.org: they are
[11:18] Arielle.Popstar @hg.osgrid.org: go to event plaza when there are 30 avatars there
[11:18] Arielle.Popstar @hg.osgrid.org: thats the point
[11:18] Bill Blight: I was there friday, had no issues
[11:18] Arielle.Popstar @hg.osgrid.org: lucky you :)
[11:18] Ubit Umarov: Some ppl had issues... viewers at 2fps
[11:19] Ubit Umarov: not much can be done with several avatars there with complexit > 300 000
[11:19] Bill Blight: I got a solid 25fps the whole time ...
[11:19] Arielle.Popstar @hg.osgrid.org: how long did it take all mesh avatars to fully load for you Bill?
[11:19] Bill Blight: some were over 500k
[11:19] Bill Blight: about 2 minutes
[11:20] Ubit Umarov: yeap but many don't have a GPU like youes
[11:20] Leighton.Marjoram @grid.outandproud.life:8002: Bill I agree with your statement that people mis-attribute the problem to the server version 0.8 v 0.9
[11:20] Bill Blight: my GPU is considered low end for gamers these days ... LOL
[11:20] Ubit Umarov: it does not matter how fast regions can send info, if viewers can't handle it
[11:20] Arielle.Popstar @hg.osgrid.org: what gpu?
[11:20] Bill Blight: low end 1060
[11:20] Arielle.Popstar @hg.osgrid.org: same as mine
[11:20] Ubit Umarov: in fact fast send may even make it worse, like happens with lludp
[11:21] Andrew.Hellershanks @hg.osgrid.org: That is a much higher level card than what I have.
[11:21] Gavin.Hird @grid.xmir.org:8002: I agree fast send can make it worse
[11:21] Ubit Umarov: when i allow 3000 on testing. i loose terrain patchs ocasionaly
[11:21] Bill Blight: yes packets tend to trip over themselves
[11:21] Gavin.Hird @grid.xmir.org:8002: there is only so much that can be done in a frame
[11:22] Ubit Umarov: viwers shoule send us http throttle levels also
[11:22] Dev Random: Is it true that all users are constrained by the slowest user? If I have a super-slow connection, will it affect scene updates for all, to keep everyone in sync?
[11:22] Ubit Umarov: so users can tune as do with lludp
[11:22] Ubit Umarov: singu does have a http control. but only internal
[11:23] Arielle.Popstar @hg.osgrid.org: i find sing slower loading
[11:23] Bill Blight: well I find position updates are more affected by slow users than rendering
[11:23] Arielle.Popstar @hg.osgrid.org: what os you using Bill?
[11:24] Andrew.Hellershanks @hg.osgrid.org: Ubit, which setting were you changing to 3,000?
[11:24] Ubit Umarov: during last SG conc at event, really many had the viewers totally overload
[11:24] Ubit Umarov: and of course blamed the region :)
[11:25] Bill Blight: Windows on my PC , linux on the servers
[11:25] Arielle.Popstar @hg.osgrid.org: well if the viewer works fine everywhere else...
[11:25] Ubit Umarov: lludp bandwith andrew
[11:25] Arielle.Popstar @hg.osgrid.org: 7, 10?
[11:25] Bill Blight: 10
[11:26] Arielle.Popstar @hg.osgrid.org: why then 2 minutes for you and 12 for me?
[11:26] Ubit Umarov: Arielle plz dont' be silly.. "everywhere" does not have >50 avatars. many very high complexity
[11:26] Gavin.Hird @grid.xmir.org:8002: where in the code is lludp bandwith set?
[11:26] Ubit Umarov: viewers set it gavin :p
[11:26] Bill Blight: there is a config option for the buffer I think
[11:26] Arielle.Popstar @hg.osgrid.org: majority of users wouldnt get the differewnce Ubit
[11:26] Ubit Umarov: well some of the llupd options are dead..
[11:26] Bill Blight: but it get's limited by the max the OS will allow
[11:26] Gavin.Hird @grid.xmir.org:8002: right, actually it has changed a bit in the LL code
[11:27] Gavin.Hird @grid.xmir.org:8002: it used to be 3000
[11:27] Gavin.Hird @grid.xmir.org:8002: now it is higher
[11:27] Ubit Umarov: in regions code is somewhere on lludpserver or around :)
[11:27] Bill Blight: firewall, and operating systems have their own UDP throttles ..
[11:27] Arielle.Popstar @hg.osgrid.org: and routers
[11:27] Ubit Umarov: well FS cries when you set more that 1500 ( warns )
[11:27] Ubit Umarov: and it does fail with 3000
[11:28] Gavin.Hird @grid.xmir.org:8002: my router is supposed to manage 930 Mbut/sec UDP across the NAT
[11:28] Gavin.Hird @grid.xmir.org:8002: Mbit/sec*
[11:29] Arielle.Popstar @hg.osgrid.org: your router is a higher end one though
[11:29] Gavin.Hird @grid.xmir.org:8002: it cost $55
[11:29] Arielle.Popstar @hg.osgrid.org: i looked at the specs and reviews
[11:29] Arielle.Popstar @hg.osgrid.org: i know, i saw
[11:29] Arielle.Popstar @hg.osgrid.org: too bad it had no wifi
[11:30] Andrew.Hellershanks @hg.osgrid.org: I just noticed I had FS set to only 1000 bandwidth. I've upped that now to 1500.
[11:30] Gavin.Hird @grid.xmir.org:8002: they have an excellent range of wifi kit to help you with that
[11:30] Ubit Umarov: im using 1500 also ( the max they advice )
[11:31] Bill Blight: Even second life recommends you don't use wifi if you can avoid it though
[11:32] Andrew.Hellershanks @hg.osgrid.org: Hm... in Sing I have maximum bandwidth at 2000 and texture bandwidth at 3000 (the defaults for that viewer).
[11:32] Arielle.Popstar @hg.osgrid.org: i wonder if that issue wishbringer is mentioning might be related
[11:32] Gavin.Hird @grid.xmir.org:8002: SL viewer is now set to 2100 kbps as default and max is 10000
[11:32] Bill Blight: wifi on non MIMO routers is only a half duplex connection , it does not send and receive at the some t ime ..
[11:32] Bill Blight: same
[11:32] Ubit Umarov: and depends on the power yr microwave oven is :P
[11:33] Bill Blight: that too
[11:33] Bill Blight: and portable phones
[11:33] Bill Blight: LOL
[11:33] Ubit Umarov: and blutooth
[11:33] Arielle.Popstar @hg.osgrid.org: i see .2% packet loss here
[11:33] Ubit Umarov: and how many ppl around yoy also have wifi
[11:34] Bill Blight: quick interlude, "Neighbor kept getting kicked off his wifi, found out everytime his phone ringed ... His phone base station was next to the wifi/router"
[11:34] Bill Blight: I have 0.0% packet loss
[11:34] Ubit Umarov: well they can coexit
[11:34] Bill Blight: yeah they can
[11:34] Bill Blight: 68ms ping
[11:34] Gavin.Hird @grid.xmir.org:8002: 0.0% packet loss and ping time around 160 ms
[11:34] Ubit Umarov: but not if the TX of one saturates the rx of the other :)
[11:35] Arielle.Popstar @hg.osgrid.org: 34 ms for me
[11:35] Arielle.Popstar @hg.osgrid.org: 17
[11:35] Bill Blight: and this server is about 1500 miles from me
[11:35] Arielle.Popstar @hg.osgrid.org: ohv montreal?
[11:35] Bill Blight: OVH yes
[11:35] Arielle.Popstar @hg.osgrid.org: in montreal?
[11:35] Bill Blight: yes
[11:36] Arielle.Popstar @hg.osgrid.org: same as osgrid servers
[11:36] Arielle.Popstar @hg.osgrid.org: thats like 20 ms for
[11:36] Bill Blight: 3 servers there 1 in France, and 1 in Germay ..
[11:37] Arielle Popstar: Ubit, did you not do a commit a couple months ago to set region z at -100?
[11:38] Arielle Popstar: i was sure i saw a commit note to that effect
[11:38] Ubit Umarov: set sure i did not :)
[11:38] Gavin.Hird @grid.xmir.org:8002: he did
[11:38] Bill Blight: you did
[11:38] Bill Blight: LOL
[11:38] Bill Blight: http://opensimulator.org/viewgit/?a=commit&p=opensim&h=39c8db8eb1cbfdefc120b1bf5f5aef45f9635509
[11:38] Ubit Umarov: i did relaz ubode to allow it
[11:39] Ubit Umarov: relax
[11:39] Ubit Umarov: i did not set regions at z = -100 bill :p
[11:39] Bill Blight: this is true
[11:39] Bill Blight: I stand corrected
[11:40] Ubit Umarov: but viewers will do a mess
[11:40] Arielle Popstar: so that will allow the titanic to sink again?
[11:41] Ubit Umarov: titanic is at -4000m or more :p
[11:41] Gavin.Hird @grid.xmir.org:8002: most new viewer builds will support z=-200, older builds will not
[11:41] Arielle Popstar: heh yes but they will settle for it to be at -80 i think
[11:41] Ubit Umarov: last time i check light is all broken below 0
[11:41] Dev Random: heh.. I remember having that convo with Mr Blue a few years ago...
[11:41] Bill Blight: current FS still looses shaders if you go below 0
[11:42] Gavin.Hird @grid.xmir.org:8002: There are new code in the LL Cougar build that fix both lighting and shadow under water
[11:42] Gavin.Hird @grid.xmir.org:8002: it has been released in any viewer yet
[11:42] Ubit Umarov: well ll does not allow <0 terrain
[11:42] Bill Blight: but did FS force disable those for Opensim like other stuff they have done
[11:42] Gavin.Hird @grid.xmir.org:8002: it allows -200
[11:43] Ubit Umarov: not sure how
[11:43] Gavin.Hird @grid.xmir.org:8002: which is why newer viewer builds got updated
[11:43] Ubit Umarov: and max ?
[11:43] Gavin.Hird @grid.xmir.org:8002: max height?
[11:43] Ubit Umarov: yeap
[11:43] Gavin.Hird @grid.xmir.org:8002: I don't recall but maybe 6k meters
[11:44] Ubit Umarov: nahh
[11:44] Ubit Umarov: that could smoke the terrain encoder
[11:44] Gavin.Hird @grid.xmir.org:8002: I'll have to check the source
[11:45] Bill Blight: looks like I'm gonna have to build me a new viewer and see what has changed in the last month and a half
[11:45] Ubit Umarov: and raw import is 256m range only or 512m
[11:45] Gavin.Hird @grid.xmir.org:8002: const F32 REGION_HEIGHT_METERS = 4096.f;
[11:46] Dev Random: Ubit, how much would a new Terrain encoding break physics? e.g. If it were a "base height" and a collection of offsets instead of just a scaling?
[11:46] Ubit Umarov: internally terrain heighs are floats now
[11:47] Ubit Umarov: the eoncoding i was talking is sending to viewers
[11:47] Dev Random: so probably no effect, except for maybe not allowing big negatives....?
[11:48] Ubit Umarov: its limits to keep things coerent to viewers
[11:48] Ubit Umarov: not nice to let a physical thing get out of bounds
[11:48] Dev Random: ha.. yeah.. got that...
[11:49] Ubit Umarov: but there are resolution issues of course
[11:49] Ubit Umarov: floats have only +- 6 digits..
[11:49] Ubit Umarov: than is 1mm error on 1000m
[11:50] Ubit Umarov: if you decide to put things at 10 000m thats 1cm error
[11:50] Ubit Umarov: ( also a isssue on region s size )
[11:50] Dev Random: yeah .. building on the moon might have some repercussions....
[11:51] Ubit Umarov: yeap :)
[11:51] Gavin.Hird @grid.xmir.org:8002: so you get a drafty skybox if you put at 10k meters
[11:52] Ubit Umarov: you may notice issue putting 2 linksets side by side
[11:52] Ubit Umarov: or editing a linkset
[11:52] Ubit Umarov: so yes there are limits..
[11:52] Dev Random: I'd wondered how much more flexibility might be wanted due to Vars... With more CPU power every year, Vars will get bigger, and people may want more resolution..
[11:52] Ubit Umarov: and makes no sense moving from floats to doubles
[11:52] Ubit Umarov: :)
[11:53] Ubit Umarov: this is a limit also on vars size
[11:53] Ubit Umarov: guess 1cm is the max error we can tolerate
[11:55] Arielle Popstar: whats the limit now?
[11:55] Ubit Umarov: ( i mean errors just from the numbers representation.. things get a bit worse if there is math on top of that )
[11:55] Ubit Umarov: 8km
[11:55] Ubit Umarov: but that is not that pratical
[11:55] Arielle Popstar: 30x30?
[11:55] Ubit Umarov: 2km seems workable
[11:56] Andrew Hellershanks: It is 5 minutes to the hour. Any other topics for today?
[11:56] Ubit Umarov: possible 1k a better choice :)
[11:56] Arielle Popstar: pfft
[11:56] Arielle Popstar: go back to megas then
[11:56] Andrew Hellershanks: :) You would have to bring back some code from the dead.
[11:57] Arielle Popstar: yeah i dunno why they removed it
[11:57] Arielle Popstar: megas had their uses
[11:58] Andrew Hellershanks: IIRC, support for megas was a bit hackish then the vars kind of took over.
[11:59] Ubit Umarov: i was talking about overall regions a nd viewers performance on those limits ..
[11:59] Bill Blight: As far as I remember Megas used even more resources than vars ...
[11:59] Arielle Popstar: very little more
[11:59] Ubit Umarov: and on larger sizes better have the regions isolated
[11:59] Ubit Umarov: or viewers may smoke
[11:59] Gavin.Hird @grid.xmir.org:8002: 1km regions seems to be a reasonable limit for the current viewer code
[12:00] Bill Blight: I did a cluster of 9 16x16 regions , viewer was not happy but they worked
[12:00] Ubit Umarov: and you cna pack them side by side ( with the viewer restriction of just one per side )
[12:01] Ubit Umarov: well you cna limit the view range to 1st neib also etc
[12:01] Ubit Umarov: or should..
[12:02] Ubit Umarov: and of course.. all users should use minimal view range
[12:02] Ubit Umarov: that is always valid
[12:02] Gavin.Hird @grid.xmir.org:8002: 64 meters?
[12:02] Andrew Hellershanks: I know of one person who has a a 1x1 and a 2x2 (separated from each other) both up against var. I haven't heard of them experiencing any ill effects, althought it isn't recommended.
[12:02] Ubit Umarov: acording to the needs
[12:02] Arielle Popstar: 32
[12:03] Ubit Umarov: 1000m WILL NOT WORK in many cases
[12:03] Bill Blight: as long as there is only one per side that is fine Andrew
[12:03] Arielle Popstar: in fs
[12:03] Andrew Hellershanks: Bill, the 1x1 and 2x2 are on the same side of the var.
[12:03] Arielle Popstar: just need betterglasses
[12:03] Bill Blight waits for the crash
[12:03] Bill Blight: it is a viewer side issue
[12:03] Andrew Hellershanks: Bill, none so far.
[12:03] Ubit Umarov: view ranfe is a important download and viewer loac control
[12:03] Bill Blight: not an opensim one
[12:03] Andrew Hellershanks: ah, ok
[12:04] Ubit Umarov: of course, does not help on a cluster of avatars with high complexity
[12:04] Gavin.Hird @grid.xmir.org:8002: The Kokua code supports 8m draw distance
[12:04] Andrew Hellershanks: I keep view distance down around 64m to 96m
[12:04] Ubit Umarov: for that.. get a gtx 1080 or better :p
[12:04] Bill Blight: if they try crossing when both sims are in view on the same side the viewer get's confused what region they are crossing into
[12:04] Arielle Popstar: complexity higher then 3000?
[12:04] Ubit Umarov: missed a 0 :p
[12:04] Arielle Popstar: :)
[12:04] Ubit Umarov: err 2 zeros lol
[12:05] Andrew Hellershanks: Ubit, I wouldn't spend that kind of money on a video card.
[12:05] Ubit Umarov: 300000 or even 100 000 if ist a big party :)
[12:05] Gavin.Hird @grid.xmir.org:8002: you can use it for mining crypto
[12:06] Bill Blight: I had an avatar land at one of my parties with 999,999 complexity and everyone came to a grinding halt
[12:06] Arielle Popstar: what you could do is get rid of the lower 2 LOD settings on any mesh clothing items in the db
[12:06] Arielle Popstar: or auto set them to 0
[12:06] Ubit Umarov: btw there is a experimental objects culling option on regions
[12:06] Ubit Umarov: it does not even send objects out of view to viewers
[12:06] Bill Blight: How do you tell a mesh clothing from a mesh building? as you could wear a mesh building if you wanted
[12:06] Ubit Umarov: ( without it, all is sent eventually =
[12:07] Bill Blight: mesh is mesh to the DB
[12:07] Arielle Popstar: mesh clothing has rigging info
[12:07] Gavin.Hird @grid.xmir.org:8002: the mesh building is probalby not rigged
[12:07] Arielle Popstar: a building hopefully not
[12:07] Ubit Umarov: easy bill a mesh building does not have 130 000 triangles :)
[12:07] Bill Blight: that is true
[12:07] Bill Blight: most sims don't have a complexity of over 999,9999
[12:07] Bill Blight: LOL
[12:07] Andrew Hellershanks: Ubit, that would depend on the building. ;)
[12:07] Gavin.Hird @grid.xmir.org:8002: actually I have seen mesh building exceeding 250k tris
[12:07] Bill Blight: yeah have
[12:08] Gavin.Hird @grid.xmir.org:8002: not in one piece, but the combined building
[12:08] Ubit Umarov: and buildings are larger... parts can't get out of view etc
[12:08] Ubit Umarov: a avatar.. bahh
[12:08] Gavin.Hird @grid.xmir.org:8002: plus add the furniture and plants
[12:08] Bill Blight: just making a point that most well built buildings have less complexity than some of these mesh avatars and clothing
[12:08] Gavin.Hird @grid.xmir.org:8002: they usually do
[12:08] Arielle Popstar: well another fix is to put a better system avatar up
[12:08] Gavin.Hird @grid.xmir.org:8002: but the interior can be excessive
[12:08] Arielle Popstar: then we dont need all these mesh bodies
[12:08] Ubit Umarov: that is what LL should had done
[12:08] Bill Blight: this is true
[12:09] Gavin.Hird @grid.xmir.org:8002: foliage
[12:09] Ubit Umarov: and NO downloads
[12:09] Ubit Umarov: like old ruth
[12:09] Ubit Umarov: all have them on disk
[12:09] Bill Blight: we had a plant that that one leaf was over 800k display weight and it had 24 leaves
[12:09] Ubit Umarov: but NO.. tghey made the crap mesh ones
[12:09] Arielle Popstar: so why cant that be changed in opensim?
[12:09] Ubit Umarov: all diferent.. all needed to be downloaded one by one
[12:10] Ubit Umarov: wlel basicly bc we don't have a avatar :)
[12:10] Arielle Popstar: we have ruth
[12:10] Ubit Umarov: but thats is viewer side
[12:10] Arielle Popstar: or we can buy one from s/l
[12:10] Bill Blight: the default avatar info comes with the viewer
[12:10] Ubit Umarov: from the complains i get ruth 2.0 seems 2 heavy
[12:10] Arielle Popstar: viewers are opensource
[12:11] Arielle Popstar: no she isnt
[12:11] Bill Blight: ruth is not too heavy ... but a lot of the clothing being made for it is ..
[12:11] Arielle Popstar: 20k
[12:11] Ubit Umarov: but this model of each one having its own mesh avatar does not hold
[12:11] Arielle Popstar: less
[12:11] Ubit Umarov: unless you are alone on a region
[12:11] Ubit Umarov: i mean own unique mesh avatar
[12:11] Bill Blight: SL is a lag fest with all the mesh avatars
[12:11] Bill Blight: and people want to bring that to opensim?
[12:11] Ubit Umarov: that was a very bad move by LL imo
[12:12] Arielle Popstar: talking a better system mesh Bill
[12:12] Bill Blight: LL has millions to throw at bandwidth and servers
[12:12] Bill Blight: and still can't make it NOT lag
[12:12] Gavin.Hird @grid.xmir.org:8002: according to HGB there are over 250k OpenSim regions, but only about 32k active users. So the chances you are alone in the region is pretty high
[12:12] Andrew Hellershanks: I think it was seen as a natural extension to having mesh objects.
[12:12] Ubit Umarov: as i said they should added a higher resolution ruth even making it optional on graphics settings
[12:12] Ubit Umarov: that should had been the proper way
[12:13] Ubit Umarov: but mesh avatars sell... uploads cost... so.. well... we understand..
[12:14] Bill Blight: LL has a physics engine that costs millions and millions of dollars, they have millions to throw at the rendering engine, as they are getting ready to change to DirectX and Metal, but we are supposed to just pull one out of thin air?
[12:14] Ubit Umarov: and of course not nothing can be done
[12:14] Ubit Umarov: Uniqueness is something all ladies want
[12:14] Ubit Umarov: ( and not only ladies :) )
[12:15] Gavin.Hird @grid.xmir.org:8002: actually LL only has about $150 to throw at bandwidth and server per region - in reality even less as there are a large number of non-paid regions
[12:15] Ubit Umarov: isn't metal a apple thing only ? mb Vulcan ?
[12:15] Bill Blight: they are going Metal for Mac and DirectX for windows
[12:16] Bill Blight: this is why they discontinued Linux support
[12:16] Gavin.Hird @grid.xmir.org:8002: metal is a Apple only tehc as directx is Microsoft only
[12:16] Ubit Umarov: no they did it bc vivox i think
[12:16] Ubit Umarov: vivox drop linux
[12:16] Andrew Hellershanks: Bill, I can't use the SL viewer any more.
[12:16] Gavin.Hird @grid.xmir.org:8002: Linux is 0.4% of their users
[12:16] Gavin.Hird @grid.xmir.org:8002: which is why they dropped it
[12:16] Bill Blight: That may have been a side reason but if you read their developer blog , that is not what Ebbe said
[12:17] Ubit Umarov: and more will drop linux
[12:17] Ubit Umarov: they made FS the linux suport team no ?
[12:18] Ubit Umarov: well i know we can't be following ll viewer all the time
[12:18] Bill Blight: well if there is going to be a linux one it is going to be third party
[12:18] Ubit Umarov: for us vulcan is a better option i guess
[12:19] Gavin.Hird @grid.xmir.org:8002: They are changing from Google breakpad to Bugsplat, and that is another paid for license
[12:19] Bill Blight: vulcan is the progression from OpenGL so yeah makes more sense
[12:19] Gavin.Hird @grid.xmir.org:8002: I wonder if FS will be able to absorb it
[12:19] Bill Blight: I'm thinking they are going to pull a lot of the opensource viewer code and make it in house, but that is just my feeling
[12:19] Ubit Umarov: it depends on how deep and specific the code changes are
[12:20] Ubit Umarov: if they optimize all for direct well a pain to do on other
[12:20] Gavin.Hird @grid.xmir.org:8002: A Bugsplat license for 50,000 annual crashes is $139 / month
[12:21] Ubit Umarov: how is the kokuaOS project as a viewer for opensim gavin ?
[12:22] Bill Blight: No windows that's how it is ... :P
[12:22] Gavin.Hird @grid.xmir.org:8002: I am working on the macOS version almost daily, but I don't have anybody to do the Win and Linux builds, so they are frozen in time
[12:23] Arielle Popstar: werent the IMS people looking into that?
[12:23] Gavin.Hird @grid.xmir.org:8002: really
[12:23] Ubit Umarov: better hide Arielle
[12:23] Ubit Umarov: lol
[12:23] Bill Blight: HAHAHAHA
[12:23] Arielle Popstar: sitting with bar between us :)
[12:24] Gavin.Hird @grid.xmir.org:8002: fact is, keeping the viewer current is a lot of work
12:25] Ubit Umarov: actually it is been a while i ear things from IMA
[12:25] Ubit Umarov: how is tit going Selby ?
[12:26] Selby.Evans @grid.kitely.com:8002: Slow -- Working on an edu viewer now
[12:26] Ubit Umarov: i had a nice chat with lisa, but months (?) ago now :)
[12:26] Selby.Evans @grid.kitely.com:8002: a bit like I described in strawman articles
[12:27] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/search?q=strawman
[12:27] Ubit Umarov: thanks selby.Evans
[12:28] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/01/edu-ima-ww-scenegate-updating-straw-man.html
[12:28] Selby.Evans @grid.kitely.com:8002: use this one
[12:28] Selby.Evans @grid.kitely.com:8002: A straw-man model for a virtual world viewer was presented with the objective of eliciting comments and suggestions for improvement. Comments and suggestions are here used to produce a more refined model. This article is explicitly licensed as public domain.
[12:29] Selby.Evans @grid.kitely.com:8002: First is minimanlist -- for new users
[12:30] Selby.Evans @grid.kitely.com:8002: Kitely is also developing grid- customized viewers for its VPG project
[12:30] Leighton.Marjoram @grid.outandproud.life:8002: What viewer code did these suggestions come from the article says "Straw-man strategy: You put it up and everyone tells you what is wrong with it."
[12:31] Leighton.Marjoram @grid.outandproud.life:8002: What was put up ??
[12:31] Bill Blight: Any actual code releases?
[12:31] Leighton.Marjoram @grid.outandproud.life:8002: Or is it a wish list
[12:31] Arielle Popstar: whats Kitely VPG?
[12:31] Bill Blight: Virtual Private Grid
[12:32] Arielle Popstar: kk
[12:32] Selby.Evans @grid.kitely.com:8002: No code releases yet -- Should have one soon -ww when released by cooperating company
[12:32] Selby.Evans @grid.kitely.com:8002: My articles are wish-lists
[12:34] Arielle Popstar: did everyone see what Dahlia was doing with web viewer?
[12:34] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/07/2018-edu-biz-kitely-will-offer-private.html
[12:34] Leighton.Marjoram @grid.outandproud.life:8002: Certainly did Arielle, looking forward to seeing where that goes.
[12:34] Andrew Hellershanks: I haven't. I know Cinder Roxley had been working on a modular viewer.
[12:35] Arielle Popstar: https://www.youtube.com/watch?v=1XNdOOVLbJs&feature=youtu.be
[12:35] Gavin.Hird @grid.xmir.org:8002: Yes I saw, but also noticed there were no dynamic updates which kindoff is the major feature of SL/Opensim compared to other virtual environments
[12:35] Arielle Popstar: you mean after editing something?
[12:36] Selby.Evans @grid.kitely.com:8002: I work with 2 web-worlds -- just a min
[12:36] Gavin.Hird @grid.xmir.org:8002: no, in the scene
[12:36] Selby.Evans @grid.kitely.com:8002: https://virtualoutworlding.blogspot.com/2018/04/2018-ww-web-worlds-3d-virtual-worlds.html
[12:36] Selby.Evans @grid.kitely.com:8002: Cybaloounge is talking students now
[12:37] Selby.Evans @grid.kitely.com:8002: Two browser-based 3D virtual worlds are now in public beta. Both are ready for initial use. CybaLOUNGE has an initial educational project in the build stage. 3DWebWorldz is presently seeking museum content as part of its preparation for educational services.
[12:37] Selby.Evans @grid.kitely.com:8002: web links in the article
[12:37] Leighton.Marjoram @grid.outandproud.life:8002: ty Selby
[12:37] Gavin.Hird @grid.xmir.org:8002: dynamic updates is usually what makes other engines tank in performance compared to the optimized content they are made for
[12:38] Gavin.Hird @grid.xmir.org:8002: static/optimized
[12:38] Arielle Popstar: ok

Personal tools
General
About This Wiki