1.17.1 versus 1.16.4 Performance

[babycat]

TLDR:​

Gameplay might see slightly lower FPS while standing still, but when moving around the performance will be smoother due to improved chunk loading algorithms in 1.17.1 resulting in an overall better experience for the average player.

Hardware Configuration & Settings​

CPU: i9-10900k @ 4.95GHz
Allocated Ram: 12GB
GPU: RTX 3080 10GB
Render Resolution: 3440 x 1440
Render Distance: 8 Chunks
Graphics: Fancy

Minecraft Mods:
+ Fabric
+ Optifine

Shader Pack:
Seus's Path-Traced Global Illumination (PTGI) Half-Render-Resolution (HRR) Version 2.1

Testing Methodology​

Without PTGI​

1. Warp/Teleport to Location (Commands used are in table)
2. Wait for chunks to load and for the FPS to stabilize
3. Record Data

With PTGI​

1. Teleport to Location
2. Switch to Spectator to Force Load lots of Chunks
3. Teleport back to reference point
4. Wait for FPS to stabilize

Performance Data​

Location	1.16.4 - No PTGI	1.16.4 + PTGI	1.17.1 - No PTGI	1.17.1 + PTGI
Ascalon Docks /Warp West	700 FPS		500 FPS
Ascalon RI /minecraft:tp babycat 6848 87 4563 -110 11	670 FPS		490 FPS
Valentia /minecraft:tp babycat 2135 123 3876 -45 40	550 FPS	47 FPS	300 FPS	47 FPS
Kalros Docks /Warp North	730 FPS		340 FPS
Kalros RI /minecraft:tp babycat 3110 158 4978 22.5 30	575 FPS		400 FPS
Silverhand /minecraft:tp babycat 6754 132 4407 -115 0	340 FPS	57 FPS	330 FPS	38 FPS
Garama Docks /Warp South	840 FPS		430 FPS
Garama RI /minecraft:tp babycat 3406 85 1269 133 -4	920 FPS	77 FPS	418 FPS	44 FPS
Tartooine /minecraft:tp babycat 2961 161 2850 25 22.5	1200 FPS		900 FPS
	670 FPS Average	60 FPS Average	456 FPS Average (-30%)	43 FPS Average

Conclusions & Discussion:​

The 1.17.1 update sees approximately 30% lower FPS when standing still, but the game's improved handling of chunk loading is likely to be at fault for this. In 1.16.4 Minecraft struggles to keep the world loaded resulting in numerous chunk errors which result in higher performance from reduced draw calls.

Generally, having spent a great deal of time working on 1.17.1 on the LokaMC build server I have seen significantly better performance on it than I have on the 1.16.4 server, so I think it is safe to say that most players will have a more playable experience on 1.17.1 than they currently do in 1.16.4.

Additionally, due to the improved chunk handling, I think it is likely that player's framerates will be more smooth due to spending less time loading/unloading chunks/subchunks.

 

[Z3K33]

Me with a i53400 And intergrated graphics :sob:

 

[babycat]

Me with a i53400 And intergrated graphics :sob:

It is possible (and I think likely) that with 1.17.1 and Optifine/Sodium performance will be more consistent as well due to improved chunk handling.

I think it doesn't make much sense to worry.

 

[FoxyBearGames]

Good to finally see Mojang taking steps to finally improve Minecraft's performance. 1.17 is definitely a good thing overall I'd say, and as another plus the game looks a lot sharper as well.

 

[adawee]

this doesn't mean that fps is going to drop right?

 

[sborm]

this doesn't mean that fps is going to drop right?

FPS will drop by roughly 30%, but to compensate there'll be much smoother chunk loading, the chunks loading like this will cause a more smooth experience overall with possibly less FPS drops due to chunks loading.

 

[TheFreshLemon]

@Gokayy I think we might actually negative fps

 

[babycat]

this doesn't mean that fps is going to drop right?

FPS will drop by roughly 30%, but to compensate there'll be much smoother chunk loading, the chunks loading like this will cause a more smooth experience overall with possibly less FPS drops due to chunks loading.

One thing that also needs to be noted about my testing is that it was done while
1. Standing Still, and
2. Without any other players around

With these factors in play, the performance drop will likely be less significant due to the increased demand and Minecraft's already good handling (relatively speaking) of players and other entities.

 

[TheFreshLemon]

One thing that also needs to be noted about my testing is that it was done while
1. Standing Still, and
2. Without any other players around

With these factors in play, the performance drop will likely be less significant due to the increased demand and Minecraft's already good handling (relatively speaking) of players and other entities.

Loka Conquest = lots of players so gg

 

[babycat]

Loka Conquest = lots of players so gg

It will likely be more stable though due to the improved chunk loading, so there will be less FPS fluctuation as you move about the battlefield.

 

[Lurnn]

It will likely be more stable though due to the improved chunk loading, so there will be less FPS fluctuation as you move about the battlefield.

so far in my experience it's just been worse

 

[TheFreshLemon]

I lost 200 fps now (crying) and my ping = around 100 now so 1.17 = regedits

 

[PartayArc]

This chart wouldve been more accruate if it was tested on a larger scale or used the same player count that appears during the peak hours

 

[Z3K33]

I cant go without a second Of me just having a lagspike When i just upgraded my internet So this updating experience is not fun

 

[casperistrash]

you get under 1000 fps at garama ri with those specs? oof

 

[SilentStormSix]

my ping randomly spikes to 200 and goes back down, also weird issue with chests forcing me out of them (same with literally any container block)

 

[Lurnn]

my ping randomly spikes to 200 and goes back down, also weird issue with chests forcing me out of them (same with literally any container block)

same on the second part

 

[TheFreshLemon]

Yall wanna join my new ffa clan

 

[koi0001]

TLDR:​

Gameplay might see slightly lower FPS while standing still, but when moving around the performance will be smoother due to improved chunk loading algorithms in 1.17.1 resulting in an overall better experience for the average player.

Hardware Configuration & Settings​

CPU: i9-10900k @ 4.95GHz
Allocated Ram: 12GB
GPU: RTX 3080 10GB
Render Resolution: 3440 x 1440
Render Distance: 8 Chunks
Graphics: Fancy

Minecraft Mods:
+ Fabric
+ Optifine

Shader Pack:
Seus's Path-Traced Global Illumination (PTGI) Half-Render-Resolution (HRR) Version 2.1

Testing Methodology​

Without PTGI​

1. Warp/Teleport to Location (Commands used are in table)
2. Wait for chunks to load and for the FPS to stabilize
3. Record Data

With PTGI​

1. Teleport to Location
2. Switch to Spectator to Force Load lots of Chunks
3. Teleport back to reference point
4. Wait for FPS to stabilize

Performance Data​

Location	1.16.4 - No PTGI	1.16.4 + PTGI	1.17.1 - No PTGI	1.17.1 + PTGI
Ascalon Docks /Warp West	700 FPS		500 FPS
Ascalon RI /minecraft:tp babycat 6848 87 4563 -110 11	670 FPS		490 FPS
Valentia /minecraft:tp babycat 2135 123 3876 -45 40	550 FPS	47 FPS	300 FPS	47 FPS
Kalros Docks /Warp North	730 FPS		340 FPS
Kalros RI /minecraft:tp babycat 3110 158 4978 22.5 30	575 FPS		400 FPS
Silverhand /minecraft:tp babycat 6754 132 4407 -115 0	340 FPS	57 FPS	330 FPS	38 FPS
Garama Docks /Warp South	840 FPS		430 FPS
Garama RI /minecraft:tp babycat 3406 85 1269 133 -4	920 FPS	77 FPS	418 FPS	44 FPS
Tartooine /minecraft:tp babycat 2961 161 2850 25 22.5	1200 FPS		900 FPS
	670 FPS Average	60 FPS Average	456 FPS Average (-30%)	43 FPS Average

Conclusions & Discussion:​

The 1.17.1 update sees approximately 30% lower FPS when standing still, but the game's improved handling of chunk loading is likely to be at fault for this. In 1.16.4 Minecraft struggles to keep the world loaded resulting in numerous chunk errors which result in higher performance from reduced draw calls.

Generally, having spent a great deal of time working on 1.17.1 on the LokaMC build server I have seen significantly better performance on it than I have on the 1.16.4 server, so I think it is safe to say that most players will have a more playable experience on 1.17.1 than they currently do in 1.16.4.

Additionally, due to the improved chunk handling, I think it is likely that player's framerates will be more smooth due to spending less time loading/unloading chunks/subchunks.

Would you consider adding new spawn to list of comparisons?

 

[babycat]

Would you consider adding new spawn to list of comparisons?

This would be theoretically possible, but hard as I don't have a 1.16 version of the new spawn. I can ask Mag though if I can get a download of that world since I am a build team member though.