| 00:14.00 | Noshei | http://en.wikipedia.org/wiki/Kingdoms_of_Amalur:_Reckoning |
| 01:14.44 | *** join/#riftuidev jmerlin (~jmerlin@205.185.121.225) |
| 01:21.55 | seebs | Huh, looks sorta neat. |
| 01:22.13 | seebs | I quite liked Rise of Nations, which was an early Big Huge Games project. Although I did use them as an example of exactly how NOT to use XML. |
| 01:22.20 | *** join/#riftuidev Imhothar_ (~miranda@ppp-93-104-169-54.dynamic.mnet-online.de) |
| 02:10.11 | *** join/#riftuidev Imhothar (~miranda@ppp-188-174-3-114.dynamic.mnet-online.de) |
| 02:16.12 | *** join/#riftuidev Imhothar_ (~miranda@ppp-188-174-168-223.dynamic.mnet-online.de) |
| 02:19.21 | *** join/#riftuidev Imhothar_ (~miranda@ppp-188-174-30-60.dynamic.mnet-online.de) |
| 02:45.12 | *** join/#riftuidev Imhothar_ (~miranda@ppp-188-174-142-190.dynamic.mnet-online.de) |
| 08:22.53 | *** join/#riftuidev Wildtide (~quassel@host86-174-70-33.range86-174.btcentralplus.com) |
| 08:23.49 | Wildtide | mornin' |
| 08:31.45 | Thorarin | Morning |
| 14:11.07 | *** join/#riftuidev Imhothar (~miranda@ppp-188-174-20-164.dynamic.mnet-online.de) |
| 14:27.11 | ZorbaTHut | wildtide, sounds about right. I played around with more complicated transactional systems for a bit but decided that I didn't really want to go to the trouble of ACID compliance for a video game api :V There's probably clever things you can do with lock entries to make it safer, note. |
| 14:37.42 | *** join/#riftuidev Imhothar_ (~miranda@188.174.20.164) |
| 15:18.16 | Wildtide | hi ZorbaTHut. I came to the same conclusion - the chances that two people would manage to update an entry in the time it takes to download/change/upload is so remote, it's not worth worrying about |
| 15:19.03 | ZorbaTHut | Yep. If you're feeling super paranoid, have the client check again in a minute and see if the data is still there. If so, it's probably safe :V |
| 15:19.28 | ZorbaTHut | If it becomes a real problem, it probably wouldn't be *too* tough to set up an extremely basic atomic transaction system |
| 15:19.56 | Wildtide | i'm pretty sure it's not worth it |
| 15:19.57 | ZorbaTHut | (of the form "take these keys which should contain these values, replace them with these new values" - if the original values don't match, you get an error and nothing happens) |
| 15:20.09 | ZorbaTHut | yeah, agreed. But if it turns out to be a major issue then let me know :) |
| 15:20.14 | Wildtide | will do :) |
| 15:22.02 | Wildtide | i did think some sort of sequential version number could work. you download the data and it's stamped as v123. when you write back, you provide 123 to the API, and if the data isn't version 123 when it goes to update, it fails |
| 15:22.24 | Wildtide | and on a successful update, you increment it |
| 15:22.32 | ZorbaTHut | that's pretty much what the atomic thing would do - you'd just embed the version number in the data, and it would make sure you had the right data |
| 15:22.39 | Wildtide | yep |
| 15:23.10 | Wildtide | it's just not something we can do outside of the API. but, as we said, not worth worrying about :) |
| 15:23.42 | Wildtide | had a go at Puzzle Pirates btw. good fun :) |
| 15:27.19 | ZorbaTHut | yeah, it's a well-crafted game :) |
| 15:27.27 | ZorbaTHut | haven't played it in forever, myself |
| 15:30.52 | Wildtide | i only got chance this morning, so have only played a few games. was waiting for the washing machine repair man... |
| 15:31.11 | Wildtide | who turned up, said he needed to order a part and left, and i spent the rest of the day in a laundrette :( |
| 15:31.19 | ZorbaTHut | welp. |
| 15:31.40 | ZorbaTHut | if you make enough to get into monster hunting, that's kinda the "raid" of the game. you'll need a badge (I forget which one) for it, though |
| 15:35.03 | Wildtide | not sure i'll be getting that hardcore with it. now, if my wife sees it, could be a different story. she's already hooked on online puzzle sites |
| 15:37.04 | ZorbaTHut | yeah, it's only a few hours to get that far, but I'll admit that I wouldn't have the time for it myself, so :V |
| 15:47.32 | Imhothar | yay the boss is back? |
| 15:48.26 | *** join/#riftuidev g0urra (~g0urra@unaffiliated/gourra) |
| 16:50.42 | ZorbaTHut | I never really left, but I'm also still not back on addons :( Someday. |
| 16:51.36 | Imhothar | argh |
| 16:52.58 | Imhothar | are you on SL full-time or got time for addons here and then? |
| 16:54.12 | ZorbaTHut | Full-time rendering work - I've got a queue of work, and right now my goal is to get as far through it as I can before feature freeze. If I have time for addons, it'd be after the queue is done. (realistically: this probably isn't going to happen until after 2.0) |
| 16:54.29 | ZorbaTHut | there's one or two small things I want to get in before 2.0 that I might prioritize over rendering, but they really will be small. |
| 16:55.11 | ZorbaTHut | Mostly stuff like "the game behavior has changed slightly and the API needs slight adjustments to not shed features". The .absorb member on Inspect.Unit.Detail() is a good example of that. |
| 16:55.55 | Imhothar | ah well |
| 16:56.12 | Imhothar | one would think there's enough rendering geeks in SF |
| 16:56.15 | ZorbaTHut | yeah, I agree completely |
| 16:56.26 | ZorbaTHut | apparently none who can get through our hiring gauntlet ;) |
| 16:57.20 | ZorbaTHut | in all fairness, AAA rendering is a *horribly* nasty subject. It would not surprise me if you could fit all the really good rendering people in the world inside a single floor of a moderately large building. |
| 16:57.32 | ZorbaTHut | and I'm not convinced I would be on that floor yet, but hopefully someday |
| 16:58.44 | Wildtide | if that's your aim, couldn't you persuade "da management" to hire an addon guy instead? |
| 16:59.11 | Imhothar | well on the one hand you got the people who know all the math and invent things like trapezoidal shadow maps, and on the other hand you got the people who know how to implement the the dozens of algorithms without blowing up the graphics card |
| 16:59.19 | ZorbaTHut | I strongly suspect you could fit all the really good addon API developers inside a single rather small room in that building. |
| 16:59.28 | Wildtide | lol, good point |
| 16:59.38 | ZorbaTHut | And it wouldn't surprise me at all if I were the world authority in MMO addon development :P |
| 17:00.31 | Wildtide | nope, you don't get yer crown until i can hide unit frames ;p |
| 17:00.58 | ZorbaTHut | for what it's worth, it's on my todo list, it's just blocking on a few other things first :) |
| 17:01.14 | Wildtide | i'm only kidding :) |
| 17:01.23 | Imhothar | you wish |
| 17:01.46 | Wildtide | i really would prefer you to concentrate on making SL a success, 'cos if it's not, our addons will become somewhat pointless :) |
| 17:01.59 | ZorbaTHut | it's kinda sad, I have all these things listed that wouldn't be too much work and that I'd really like to add . . . and I've got so many of them that I could probably spend two entire patches just providing small fixes |
| 17:02.07 | ZorbaTHut | so, hey, someday |
| 17:03.06 | Imhothar | all you need is some freelance nerds from over the pond to help out over VPN ;) |
| 17:04.48 | *** join/#riftuidev jmerlin (~jmerlin@205.185.121.225) |
| 17:05.22 | Wildtide | yeah, we've said it before. open up your repo to us, it couldn't possibly be anything other than a complete success! |
| 17:05.47 | Imhothar | totally |
| 17:07.31 | Imhothar | just imagine the luxury: instead of complaining about bugs we simply fix them ourselves. wouldn't that be great? |
| 17:08.52 | ZorbaTHut | well I certainly can't think of anything wrong with this plan |
| 17:09.07 | Thorarin | lol |
| 17:09.23 | Imhothar | it's win-win |
| 17:13.06 | Thorarin | I was considering donating some code to Trion, but not for addons :P |
| 17:14.58 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:14.58 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:16.09 | seebs | FWIW, I did confirm (using a new machine) that Rift has a very noticable and disproportionate performance penalty for AMD hardware, that is pretty far out of line with what I see in other games. |
| 17:17.12 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:17.12 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:18.23 | Thorarin | Hmm, it's performing quite similar on my hardware and my girlfriend's, same-gen nVidia and AMD cards |
| 17:19.01 | seebs | Well, my data points are (1) laptop with geforce 560M, (2) desktop with Radeon HD 6950, (3) dekstop with GeForce 670. |
| 17:19.36 | seebs | The laptop and the AMD desktop get very similar performance in Rift; in everything but Rift, the laptop is maybe 1/3-1/2 the speed of the desktop. |
| 17:20.03 | seebs | And the 670 is getting >2x the performance in Rift of the 6950, while for other stuff it appears to be maybe 30% or so faster. |
| 17:20.41 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:20.41 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:23.08 | Thorarin | Hmm, 6950 outperforms my GTX 570 at least |
| 17:24.24 | Thorarin | Both are somewhat overclocked models |
| 17:25.23 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:25.23 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:26.20 | seebs | Interesting. |
| 17:26.44 | seebs | What threw me was that since I mostly play Rift, I got used to using the same settings on them. Then I installed TSW, and the desktop was fine and the laptop got maybe 10fps. WHOOPS. |
| 17:27.00 | Thorarin | World event end boss last night, around 100 people there probably |
| 17:27.09 | Thorarin | We both got like 7 fps :S |
| 17:27.56 | *** join/#riftuidev Torhal_ (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:27.57 | *** mode/#riftuidev [+v Torhal_] by ChanServ |
| 17:31.57 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:31.57 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:34.18 | *** join/#riftuidev Torhal|work (~torhal_wo@173-14-207-161-huntsville.hfc.comcastbusiness.net) |
| 17:34.18 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 17:46.11 | *** join/#riftuidev Noshei (~quassel@c-67-161-146-244.hsd1.co.comcast.net) |
| 17:58.17 | *** join/#riftuidev Torhal|work (~torhal_wo@76.164.170.2) |
| 17:58.17 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 20:50.58 | jmerlin | Rift is severely held back by its engine as far as playability in environments like that are concerned. |
| 21:15.33 | seebs | Once you're under heavy load, it's very much CPU bound on basically any hardware. |
| 21:16.00 | seebs | When I had Rift drop from 30fps to 3fps, I watched a resource meter, and basically the GPU went from nearly-100% utilization to about 10%. So, yeah. |
| 21:16.53 | jmerlin | Sure, but even an i7 running at 3.5GHz is the bottleneck for ~30 people |
| 21:17.17 | jmerlin | On minimum graphical settings, which would imply there's like an O(n^5) algorithm somewhere chewing up CPU unnecessarily |
| 21:17.44 | jmerlin | The vertcounts in Rift are low, the settings turned off means normal calculations and other such CPU-intensive post-processing isn't happening, so something is fundamentally broken. |
| 21:19.00 | jmerlin | If I had the client source I'd gladly do some profiling and figure out what's eating all the CPU. |
| 21:30.39 | seebs | I am not sure whether it's the rendering or the game logic. |
| 21:30.44 | seebs | Yet another reason I really wish we had demorecord. |
| 21:40.56 | jmerlin | But do they care? :o |
| 21:50.33 | seebs | I suspect they do, but it |
| 21:50.44 | seebs | it's not always easy to get the time it takes to track something down properly. |
| 21:51.02 | seebs | It can be really hard to get the right tools to enable the results you want. |
| 21:51.23 | jmerlin | PIN + a few minutes running profiles |
| 21:53.07 | jmerlin | Just need debugging files to get any useful data from it, but those include source. They're not going to hire someone with the skillset necessary without some ridiculous compensation, though. They're too busy hacking linux and BSD kernels or something. |
| 21:53.35 | jmerlin | And they certainly won't let a non-Trion employee do it, heaven forbid they actually let the community make the game better. |
| 21:56.38 | Noshei | I can easily say it's not a CPU issue in populated areas |
| 21:57.11 | Noshei | I can sit here and watch my CPU usage not change at all when entering populated areas, while my GPU usage goes down |
| 21:57.20 | Noshei | and Rift barely touches my CPU |
| 21:58.14 | Noshei | oh and new version of EasyMail is now out, allows you to send the same message to multiple people |
| 21:58.45 | jmerlin | Noshei: what GPU? |
| 21:59.16 | Noshei | 2x480 |
| 22:00.32 | seebs | For me, the game tends to keep one CPU core pegged pretty much no matter what. |
| 22:00.43 | seebs | Which means that it's not going to change under load, but GPU usage will drop. |
| 22:01.01 | seebs | The only time you'd expect to see CPU load change for that would be if you weren't already maxing out that thread. |
| 22:02.22 | seebs | I should figure out what to use to see CPU/GPU usage. I had an AMD thing for the AMD/ATI machine, I don't know what does the same stuff on intel/nVidia. |
| 22:03.28 | jmerlin | I have 2x580s and my GPUs are never under any load in Rift, (compared with any modern, decent engine), in fact I typically have a higher GPU load playing bejeweled and watching Netflix at the same time. |
| 22:04.20 | jmerlin | The CPU is pretty much maxing out any cores its threads are running on, and if you drop to minimum graphics, you still have terrible FPS in crowded areas. Run out into an empty space in front of Meridian, for example, and your FPS shoots up to 150+. |
| 22:04.35 | Noshei | seebs: make sure you have MainThreadCPU = 0 is the rift.cfg file |
| 22:04.53 | Noshei | otherwise it will only utilize a single thread |
| 22:05.21 | seebs | The last time I saw a purple name comment on that, they said this was not so. |
| 22:05.38 | jmerlin | That config should do what it says: change the core the game uses on its primary thread. |
| 22:05.41 | seebs | The game always runs multiple threads, but one of the threads is The Main Thread. |
| 22:05.48 | seebs | And there is only one Main Thread. |
| 22:06.07 | seebs | That line changes whether it tries to stick on a particular core, or whether it's allowed to get swapped around. |
| 22:06.27 | seebs | But you still can't actually get it running on more than one core, so all you're adding is more expensive context switches. |
| 22:07.05 | seebs | You're still getting the same aggregate CPU usage overall, with that thread running basically 100% of the time, and other threads running much of the time. |
| 22:07.16 | seebs | I suppose I can go double-check, but this was the case in the past for me. |
| 22:10.38 | Noshei | so just doing a quick check, my CPU usage actually goes down in meridian, compared to non populated zones |
| 22:12.16 | jmerlin | I might run Rift in PIN tonight and get the hot spots, then send those to someone (since I can't post on the forum). I find it doubtful that there's anyone at Trion who can take an address from the build and get the function in source, so they'll probably just ignore it anyway. |
| 22:14.24 | seebs | My vague theory is that the really heavy load cases in fights where CPU usage is pegged and GPU usage drops have to do with processing the client is doing -- say, stuff like managing which animations and effects to display. A bunch of people just standing around might be a higher GPU load with less CPU effort. |
| 22:15.58 | Noshei | yeah I gotta try and get MSI afterburner working on Windows 8 to see if there is any changes on GPU usage |
| 22:16.21 | jmerlin | That stuff should be of zero consequence to a modern CPU |
| 22:23.29 | seebs | I haven't bought anything from MSI in a long time. Long time ago, they stiffed me on a rebate. So I took 'em to small claims court. So... They took all the info from the rebate submission (which they had after all) and used it to fill out fake credit apps and sign me up for a bazillion mailing lists. |
| 22:49.39 | *** join/#riftuidev Noshei_ (~quassel@c-67-161-146-244.hsd1.co.comcast.net) |