04-17-2021, 02:02 AM
As many of you are well aware, roughly two months ago the performance of the game took a drastic drop despite having been stable in the past even with activity levels much higher than our current amount. This has meant a lot of stressful debugging on our end with the eventual conclusion that we lack the ability to see what's causing it with the usual CPU profile tool.
When we did update the engine recently, it showed us that an internal engine proc called sendmaps() is the cause. So we're doing something recent to trigger that which we can't diagnose right now, and we're waiting on an update to give us the tools to figure out what's happening.
![[Image: rbAUbsy.png]](https://i.imgur.com/rbAUbsy.png)
To put it in perspective, here you see that the total CPU is 97%. High! But 95% of that comes from the internal sendmaps() function, with our own code only being 2% or so.
On Monday we should be getting improved profiling that'll hopefully let us pinpoint where the problem is.
It's temporary, but for now we'll have to endure it...
When we did update the engine recently, it showed us that an internal engine proc called sendmaps() is the cause. So we're doing something recent to trigger that which we can't diagnose right now, and we're waiting on an update to give us the tools to figure out what's happening.
![[Image: rbAUbsy.png]](https://i.imgur.com/rbAUbsy.png)
To put it in perspective, here you see that the total CPU is 97%. High! But 95% of that comes from the internal sendmaps() function, with our own code only being 2% or so.
On Monday we should be getting improved profiling that'll hopefully let us pinpoint where the problem is.
It's temporary, but for now we'll have to endure it...