benchmark1 has outdated data, so it won't be fully accurate imo. We should move to a new one.
Account Details | |
---|---|
SteamID64 | 76561198046110893 |
SteamID3 | [U:1:85845165] |
SteamID32 | STEAM_0:1:42922582 |
Country | United States |
Signed Up | August 8, 2017 |
Last Posted | September 14, 2024 at 1:13 AM |
Posts | 1545 (0.6 per day) |
Game Settings | |
---|---|
In-game Sensitivity | |
Windows Sensitivity | |
Raw Input | 1 |
DPI |
|
Resolution |
|
Refresh Rate |
Hardware Peripherals | |
---|---|
Mouse | |
Keyboard | |
Mousepad | |
Headphones | |
Monitor |
So um, I had 1000 crate #85s that I bought almost two years ago for a total of $30, and um. Now I bought a Valve Index.
Moistbenchmark
Could you benchmark with mastercomfig low? Wondering how it'll work on these CPUs.
AriemastercomsWhat causes you to say that? Curious what you exactly mean by "state of the source server engine".
These are 4 non-default, very rarely changed variables on a 10 year old engine that no longer gets a lot of love from its creator.
If they work, only 1 out of 4 (clock correction) has any bearing on competitive play, the others are the concern of server operators, who should prefer sticking to defaults (used by nearly everyone) instead of rarely used (threaded) optimizations.
As a server operator I favor stability over changing variables that hardly anyone else runs. That way when one of my servers crashes, it's more likely to be something a lot of server operators have already seen, instead of some rarely unknown network threading bug in the orangebox server code.
League configs should concern themselves with settings that affect competitive play, not try to do server optimizations nor change server names and passwords.
Just because the engine isn't updated doesn't mean variables magically stop working.
sv_clockcorrection_msecs was changed to 30 in CS:GO, and is set to 0 in many CS:GO competitive leagues. It definitely works in TF2 too.
The other ones were optimizations that I agree should be server preference, and not part of a standard competitive config. However, optimizations can affect tickrate and thus networking reliability so the stability should be at least investigated to see if they can be used in normal competitive play.
ArieStuff that probably doesn't work considering the state of the source server engine:sv_clockcorrection_msecs "30" host_thread_mode "1" // turns on multithreading if the server supports it lzma_persistent_buffer "1" // keeps a persistent buffer for LZMA decompression to prevent memory from being allocated and unallocated repeatedly sv_parallel_sendsnapshot "1" // sends packets in parallel
What causes you to say that? Curious what you exactly mean by "state of the source server engine".
I've adjusted more networking settings, and would like to stress test with as many players as possible! Join if you can!
Paste in browser URL bar: steam://connect/tf.mastercomfig.com
Or, paste into console: connect tf.mastercomfig.com
You can also talk with me and others on Discord while we play! https://discord.gg/CuPb2zV
Amazing!
Testing some networking settings on a server:
steam://connect/tf.mastercomfig.com
connect tf.mastercomfig.com
Voice chat on Discord! https://discord.gg/CuPb2zV
v6.11.7 released with performance improvements, networking enhancements and quality of life improvements.
Can someone tell me the dealio with ESEA? Can I remove all references to it from mastercomfig documentation yet?
Tanniemine doesnt have it disabled, i copied it from credu stream and he told me there are ragdolls gibs and shadows but rn it seems to be fixed.
maybe ill have to wait for him to stream again and get his cfg
How did you copy it from credu's stream? Just download it from here: https://mastercomfig.com/download
sAvenSo does that mean you should gradually increase interp until the value is no longer yellow?
Usually there's too much variance to reliably get it from looking at the packet loss indicator. You should instead analyze the average server variance, and average sv tickrate, and determine the value from that, plus a fudge factor. I already did this with my interp value of 0.021, which should also not be very noticeable to humans. I only recommend this interp for projectiles and melee, and 0.0303 (interp ratio 2) for hitscan. This is because hitscan is fully lag compensated for interp, and there is not clientside offset like projectiles do with higher interp (because no travel time).
sAvenIs there any reason why r_worldlightmin is different for the very_low preset vs. the low preset even though r_worldlights is 0 for both?
I wasn't sure if r_worldlightmin was only for world lights.
zfnJust curious why the changelog says "Optimized particles", but the "ConVar.cl_particle_retire_cost" "0.0001" has disappeared from dxsupport_overide? Was it not working?
Just testing if this will affect CPU load. It might be more beneficial for dxsupport to decide based on your GPU. There are other particle optimizations, like disabling more particles using dxsupport. It might return in a future update.
whitepuzzleI've never had an issue with 15.2ms interp (no players stuttering or hitreg issues or anything) and the difference in projectile delay is noticeable.
Noticeable between 21 (or 17.9) vs 15.2ms? I kinda doubt that. Keep in mind that the delay isn't server side either, it's an artificial client delay that spawns your rocket further down the barrel to have better visual sync, since projectile spawns and destroys are not interpolated.
And just take a look at your net_graph 3. See whenever your lerp turns yellow. That means you are dropping packets.
v6.11.6 released with performance improvements and bug fixes.
sAvenSince TF2 just crashes when loading in-game matchmaking in DX8, would it be better to have "mat_dxlevel 90/95" in mm_overrides.cfg to save people a restart?
I don't want to set that, since it would also set for people already on DX9.
sAvenHow come my comfig.cfg doesn't have the comments and organization of the code changes for version 6.11.5?
https://imgur.com/a/oJfha0q
Comments are removed from the VPK version to save space. You can view the config files with comments on GitHub, and download them for offline access.
_KermitsAvenwhitepuzzleAlso, how to get 15.2ms interp on V6? I set snapshots to "lan" in modules.cfg, but I still have 21.0ms interp.
I think the answer was that 15.2 was too unreliable and resulted in packet loss.
I'm sure that on soldier/demo 15.2 is still worth it.
You can change it using snapshots_lan in user/game_overrides.cfg.
However, I seriously advise against using it: https://www.teamfortress.tv/51693/current-year-net-settings#4.
v6.11.5 released with performance improvements and bug fixes.
whitepuzzleI switched back from 7a to 6.11.4 (medium-low.vpk) and fences (like the one near snakewater spawn, or ammomod) look aids now even with MSAA enabled. How to fix it? On 7a, they only looked shit with MSAA disabled.
Use aa=medium in your modules.cfg.
Also, to respond to your edit, that wiki edit is straight from the mastercomfig wiki.
I have no reason to believe this launch option hasn't been in the game for a long while