I have switched back and forth several times for testing and there is a significant drop in performance switching from Forge 1.12.2 Build 2674 to Build 2678 on a multiplayer server. However, I'm not seeing any error messages on the console or experiencing any crashes. It's just a performance drop. 

Edited May 4, 20187 yr by Fandomaniac
clarify title description

Why is this an issue if you know one version works better than the other just stick to that version.

9 hours ago, MDW01 said:

Why is this an issue if you know one version works better than the other just stick to that version.

That's not helpful. Problems should be fixed, not just avoided.

 

With that said, there really isn't enough information to do anything here.

If you are experiencing a clear difference, can you post some server profiling results?

For both versions, run /debug start, then, after a minute or two, run /debug stop.

 

This will create a file in the debug subdirectory - upload them to pastebin/whatever and link them here.

 

2678 is a development build performance drops are to be expected. If you have actual information or the issue continues in future builds create an issue on the GitHub.

"Performance drops" is not a valid report. We need some form of actual numbers, videos showing issues, SOMETHING that can actually be tracked or reproduced.

Also, luckally you're not like some people and pushing out a 500 build change and complaining, you only have 4 builds. And looking at the history I have a idea which one could be effecting it.

So try 2675 vs 2674. If you can reproduce, provide logs, The sampler mod may help: https://minecraft.curseforge.com/projects/sampler

 

There were 3 mod updates these last few days. After the updates, the performance is now back where it was with 2674. Next time I encounter an issue like this, I will wait for any new mod updates and re-test before submitting a report. I'll also try to compile as much data as I can via /debug, etc.

Edited May 7, 20187 yr by Fandomaniac
Addendum