View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0004695 | Spring engine | General | public | 2015-03-08 15:30 | 2017-09-24 19:14 |
| Reporter | msafwan | Assigned To | abma | ||
| Priority | normal | Severity | major | Reproducibility | always |
| Status | closed | Resolution | no change required | ||
| Product Version | 98.0 | ||||
| Summary | 0004695: Please replace spin-lock, its very bad | ||||
| Description | Okay I'm sorry about that last bug report. Sorry jK Okay, I read the link, it says spin-lock "burn system resource and make CPU unavailable for system task" unless.. you added "sleep", or replace it with "blocking". This clearly state spin-lock its bad because Spring did waste resources badly like I mentioned in previous bug report. Player also said "WorkerThreadSpinTime=0" fix stutter, this further says spin-lock is bad, it create stutter rather than make Spring faster. IMO spin-lock need to be fixed to be more efficient or replaced with something more efficient, like what the link says. | ||||
| Steps To Reproduce | Look at what people say, sometimes they mention about possible overheat, sometimes they said game became unplayable laggy after being in game too long. | ||||
| Additional Information | abma ask when it started: It started probably since Spring 96. abma asked what's my settings: Here is my settings recorded in this infolog: http://pastebin.com/0xjTwL9R abma's links on spin-lock: https://springrts.com/mantis/view.php?id=4693#c14115 | ||||
| Tags | No tags attached. | ||||
| Checked infolog.txt for Errors | |||||
|
|
GoogleFrog have changed default WorkerThreadSpinTime value a few times in ZKL, it is probably easier to convince him to set it to zero as default ZKL settings. |
|
|
I'll set WTST=0 as default in ZKL and inflict it on the masses with a widget somewhere during next week hopefully. However, even with WTST=0 i get weird behaviour (98.0.1-398 and later) where everything starts flickering as soon as i time out or when catching up. With nonzero WTST i get this even in realtime. |
|
|
@Anarchid: search for wrong opengl calls / DebugGL=1 / fix these errors |
|
|
also as note: WorkerThreadSpinTime = 1 is default where ZK seems to default to 5 which increases cpu load (but maybe performance, too) |
|
|
what does using Spin-lock achieve? I'm just saying it put lots of stress on my CPU. Fan always at max when Spring running. I don't really want to risk my PC, I don't have enough resource to replace it. Limit my CPU usage could be solution, but think about it... what about other people? |
|
|
the reason is written there: https://stackoverflow.com/questions/1107593/what-are-trade-offs-for-busy-wait-vs-sleep "lock" is used for syncing threads. |
|
|
Is people don't believe me about the heat issue? or just can't/won't fix at the moment? |
|
|
> Is people don't believe me about the heat issue? its not that devs are waiting for bug reports. they do what they think is most important and most fun. my personal todo list for spring is way to long, realistic seen i never can do/fix all whats on it. > or just can't/won't fix at the moment? its argueable if this is a bug of spring or not. its more a problem that your system can't run at 100% load for a long time, the cooling is to weak. imo basicly the system is broken. imo it only matters what is faster, busy waiting or sleep (and busy waiting seems to be the winner) |
|
|
okay, I Google-d (on spin-locks). It must be a bug... Dev-Spring had big CPU overhead and minimal speed gain. (the cost/gain ratio make less sense for my system) And, since spin-lock is best when task is closed together, could it be you guys made a mistake and spin-locked between empty game-frame? also why spin-lock when we set to 1 core? |
|
|
Quote from your infolog.txt: > VSync disabled enable vsync and if it still still stays at >= 100% cpu and below 60 fps your cpu is to slow. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2015-03-08 15:30 | msafwan | New Issue | |
| 2015-03-08 15:32 | msafwan | Tag Attached: free | |
| 2015-03-08 15:32 | msafwan | Tag Attached: hardware stressor | |
| 2015-03-08 15:33 | msafwan | Tag Detached: free | |
| 2015-03-09 09:58 | cleanrock | Note Added: 0014124 | |
| 2015-03-09 13:13 | Anarchid | Note Added: 0014126 | |
| 2015-03-09 23:58 | abma | Note Added: 0014135 | |
| 2015-03-09 23:59 | abma | Note Added: 0014136 | |
| 2015-03-10 00:00 | abma | Note Edited: 0014136 | |
| 2015-03-10 00:01 | abma | Note Edited: 0014136 | |
| 2015-03-10 01:21 | msafwan | Note Added: 0014137 | |
| 2015-03-10 01:28 | msafwan | Note Edited: 0014137 | |
| 2015-03-10 01:30 | msafwan | Note Edited: 0014137 | |
| 2015-03-10 01:30 | msafwan | Note Edited: 0014137 | |
| 2015-03-10 01:31 | abma | Note Added: 0014138 | |
| 2015-03-10 01:32 | abma | Note Edited: 0014138 | |
| 2015-03-10 01:47 | msafwan | Note Added: 0014139 | |
| 2015-03-10 01:53 | abma | Note Added: 0014140 | |
| 2015-03-10 04:36 | msafwan | Note Added: 0014141 | |
| 2015-03-10 04:38 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:39 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:42 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:43 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:44 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:44 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:45 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:46 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:46 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:47 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:48 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:51 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:53 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:55 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 04:56 | msafwan | Note Edited: 0014141 | |
| 2015-03-10 10:10 | abma | Note Added: 0014142 | |
| 2015-03-10 10:10 | abma | Status | new => closed |
| 2015-03-10 10:10 | abma | Assigned To | => abma |
| 2015-03-10 10:10 | abma | Resolution | open => no change required |
| 2015-03-10 10:12 | abma | Note Edited: 0014142 | |
| 2017-09-24 19:14 | Kloot | Tag Detached: hardware stressor |
