Network and performance audit

[Madis0]

I do want to be Fabulously Optimized to be as private and as performant as it can. Therefore it would be wise to audit its network requests with something like Fiddler and performance with something like spark.

I may take this task on myself at some point but if there are more knowledgeable people here who know how to properly test these things out and interpret the results, I would appreciate it.

[isXander]

I would love to see an in-depth benchmark of Vanilla - FO compared to just Vanilla - Sodium

[Madis0]

Benchmarking is not exactly the purpose of this (the purpose is fixing issues, not showcasing the benefits), but sure, volunteers can do that as well.

[voidpointer0x00]

I do want to be Fabulously Optimized to be as private and as performant as it can. Therefore it would be wise to audit its network requests with something like Fiddler and performance with something like spark.

I may take this task on myself at some point but if there are more knowledgeable people here who know how to properly test these things out and interpret the results, I would appreciate it.

Sounds good, but it's a bit unclear to me: what exactly do you want to do? Do you want to configure the FO using Spark&Fiddler to maximise the performance or do you want to just compare Vanilla vs FO performance?

In the first case, I bet multiple people can test various configuration options on their hardware, so that we get somewhat objective data and configure the modpack accordingly. But I reckon we'll need some defined tests with the same environment.

[Madis0]

what exactly do you want to do? Do you want to configure the FO using Spark&Fiddler to maximise the performance or do you want to just compare Vanilla vs FO performance?

The former. As you said, comparisons vary by hardware, but fixing issues should positively affect everyone.

In the first case, I bet multiple people can test various configuration options on their hardware, so that we get somewhat objective data and configure the modpack accordingly.

Ehh, I don't think option changes are the real differentiator at this point. Sure, it may show a difference between enabling and disabling a feature, but overall having mods that unintentionally harm performance is probably a bigger issue.

[voidpointer0x00]

Ehh, I don't think option changes are the real differentiator at this point. Sure, it may show a difference between enabling and disabling a feature, but overall having mods that unintentionally harm performance is probably a bigger issue.

Now I see what you mean :)

[Madis0]

Attempted a network audit with Fiddler Classic and mitmproxy - neither showed the requests coming from the client itself. Also made sure that HTTPS interceptor is set up, tried JVM flags and Prism Launcher proxy.

Wireshark with Winshark was suggested - I tried Wireshark only and it lacked proper filters, so I didn't get anything good.

Got the best results with opensnitch. As there is a cape mod, these requests were expected.

References: https://discord.com/channels/859124104644788234/859124104644788237/1175483638503194624 (FO) https://discord.com/channels/507304429255393322/789950149395742740/1175485471485341756 (Fabricord)