|
|
|
[
Permlink
| « Hide
]
Funk Schnook added a comment - 15/Sep/07 06:27 PM
OK I just left the viewer running for an hour sitting in my skybox not even moving. Mem usage is pretty high for doing nothing right? (as seen in sl-mem-usage2-rc1.18.3.4.jpg)
I just attached another screenshot of my task manager running 1.18.0.6 for an hour. The mem usage is only about 300MB compared to 1.18.3s 1.2GB. There is a huge difference here.
Both clients are using exactly the same settings. I ran both clients for an hour simply sitting my av in my skybox. Funk, could you try this with one of my viewers based on the RC (18xm or later)? I just tested Nicholaz's viewer for an hour (BleedingEdge-n based on 1.18.3(4)) and dont see this massive memory leak.
His viewer's memory usage is much closer to 1.18.0.6 (check my screenshot) Our QA was unable to repro.
I have a bit more info. I had all the clients set to a 512 draw distance (I do this regularly), so being 500M up didnt really make a difference. I was still loading every avatar texture of people tping into the sim. The sim also gets quite busy. If you want to repro this, I suggest doing it on my sim "FNKY Cake" between 11am and 4pm SLT when its busiest.
I can repro this every time, and it doesn't even take an hour to get to the state mentioned above. It happens within the first 15 minutes or so Well, given the fact that it doesn't seem to occur in my viewer and what Funk says there with the busy sim, I may add that there is a memory leak (memory actually lost) and a memory bubble (memory added to a list infinitely, not forgotten but also not flushed) in the viewer. The first is there since Voice, the 2nd is there for a long time ... there was a patch by me for the 2nd which collided with changes that happened in Voice (actually with exactly the change which also caused an additional leak). Repro: Just go to any busy sim with lots of people and animations (Bad Girls for example), ideally many people passing through (new ones coming and going). Or of course the one Funk suggested. Soft Linden has more details on the leaking etc., otherwise feel free to ask. I've seen this on every viewer after 1.18.0.6 on x86_64. It seems to be leaking VBOs. Try turning off VBO and see if that helps?
Seg, yeah someone mentioned that and I tried it. It didnt help.
Nicholaz has fixed this in his viewer though. Is there any reason why his patch cant be used? Im getting this issue in windlight too. I use a draw distance of 512 and within the hour, my task manager reports over 1GB of mem usage. Once again, Im on a semi popular sim where many avs are coming in and out (I think this has something to do with it)
Removed "WindLight": A "First Look" version should ONLY be selected if the issue only affects that version.
Hi!
I have bought a dual core CPU (AMD X2 6000+, AsRock AliveXFire-eSATA2 Motherboard, 2 GB RAM) yesterday and have the same problem since i run SL on my dual core PC. (No problems on my old AMD X64 3000+ on a Asus board) The increasing of memory usage is a bit slower when i limit the client to one CPU. I hope this comment helps Mojo Torley, this effects windlight too.
Hardware Overview: Machine Name: iMac GeForce FX 5200: Chipset Model: GeForce FX 5200 I am also having the same symptoms as described above. At my home location, seems to run smooth as silk. As soon as I teleport to a busier place, a steady decrease in performance. Within 20, not uncommon for a gig of ram to become utilized. This happens with the most recent candidate release as well as the prior. Both Windlight and regular viewers.
HELP! Model Name: iMac Model Identifier: iMac6,1 Chipset Model: NVIDIA GeForce 7600 GT I have this same issue. RAM usage steadily increases from around 250MB upto 1.2GB over a 90 minute time period. I then quit and return to regain the excessive memory.
At the moment RAM usage after 20 minutes log on time is at 723MB. Second Life 1.19.0 (5) Feb 28 2008 17:18:12 (Second Life Release) You are at 189567.8, 259871.4, 35.3 in Hathian located at sim5625.agni.lindenlab.com (8.2.34.181:12035) CPU: AMD (Unknown model) (2010 MHz) Moving my comments to this issue, as the more appropriate one. Formerly re-opened under
Description of issues remains the same for now. Seeming memory management anomalies and crashiness persist in 1.19.1 RC1 as of 20 Mar 08. Am wondering if memory leakage has worsened under 1.19.1 RC0 – Page file is now nearing 2 GB on my (older) system. It's almost always been large, but leaks were addressed with some success when this issue originally surfaced back in Spring/Summer 2007. It seems like probable new leaks are emerging as the present 1.19.1 (0) Release Candidate seems to become increasingly unstable and slow with time online. Then again, it could be my boat anchor of a machine, running Win XP SP2 with a mere 1GB of RAM. Still, the new RC seems pretty fast and stable when first loaded, and that page file does seem to keep creeping up in size over time. I've been mainly messing with the new media calls via LSL, so it's also possible (I'd guess) that the root of the problem lies in this first hack at expanding the uses of Parcel Media streams to handle Webpage display on a prim. Perhaps the leak is in how I am calling pages, loading them, and refreshing the same page repeatedly by way of calls to the parcel media functions in LSL? Also, the crash logger has gotten a habit lately of getting into an infinite loop when it says that it's transmitting logs, and eventually has to be terminated manually in most instances. Crashes, though, are not a universal experience with this RC. At least not for my installation. I am crashing to blue screen after 45 minutes to 1 hour of use. More often if I am building and using textures heavily.
I would have added the crash log, but evidently one was not created by the SL client as it hit blue screen. When I logged back in I was not given the option of sending a crash report. Apparrently the log is just created if the client crashes without going to blue screen. Second Life 1.20.15 (92456) Jul 18 2008 10:58:42 (Second Life Release) CPU: Intel Core 2 Series Processor (1828 MHz) I have the same issue as posted above.
Spoke with my PC manufacturer, and bought 2 GB more RAM just to try and eliminate the out of memory error. CPU: Intel Pentium III/Pentium III Xeon (0.25 micron process) with external L2 cache (2666 MHz) Support ticket from resident Loki Ball unable to access Jira system:
"I continue to get memory spikes that remain at 100% on my cpu. It doesn't matter if I set the affinity or not. It will spike at 100% and just hold that. Doesn't matter if I relog, doesn't matter if I clear cache. None of that works. Reinstalling doesn't help at all either as I'm sure your fully aware of already. I'm actually looking for a working solution to this so I don't end up frying another system using SL. Here are some of the specs on my system.. CPU: AMD (Unknown model) (2611 MHz) |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Reopened
Critical
High memory usage after short periods of use
