So this is still a thing, also after the new update 2.20?

any reason this is a distinct build and not standard ? any downsides ? stability issues ?

Wew lad, you convinced me to finally try a beta.

Hello, im Dora the Explorer and I seem to have misplaced that "MASSIVE 20-70 % performance increase"
Can You help me find it ?
https://www.youtube.com/watch?v=z2vJndnMZ-w

Originally posted by 76561198146637018:

Hello, im Dora the Explorer and I seem to have misplaced that "MASSIVE 20-70 % performance increase"
Can You help me find it ?
https://www.youtube.com/watch?v=z2vJndnMZ-w

So, troll or whatever?

Dunno why ya bother "testing" this on extreme AI / bad mods / bad scripting, you will never get good performance. The results from January til now for people enabling profiling and playing GOOD servers is 100% of all people that enabled it, came back to our place, and played, and commented got VASTLY higher FPS/smoothness....

There are places on some maps that will never be good... and when ya play ♥♥♥♥ like exile with 2 dozens mods -- you can live with 4/4/6 on view distance for ground/vic/air....

thinking you will max ♥♥♥♥ out is absurd in those benchmarks but if you don't EXTREME it all, performance profiling has helped several dozen people play with better results on our exile and antistasi servers....

But keep on ♥♥♥♥♥♥♥ about this old game -- ignore put absurd benchmark tests out there when theres like a handful of good servers left in the world.... run ♥♥♥♥ on extreme with insane view distances thinking you will get the FPS ya want -- it won't happen...

... but claiming performance/profiling didn't make a difference from January til now is stupid.

PS that xeon e5 blows -- and we run our servers off a 4.9Ghz Ryzen 9 with 12-16 gigs dedicated. My rig is an older 3900x with a 3080 and the smoothness and fps gets/got a massive bump for using profiling since January. No idea what the last update did, but I didn't disable the profiling...

Originally posted by MetalMilitia:

Originally posted by 76561198146637018:

Hello, im Dora the Explorer and I seem to have misplaced that "MASSIVE 20-70 % performance increase"
Can You help me find it ?
https://www.youtube.com/watch?v=z2vJndnMZ-w

So, troll or whatever?

Dunno what yer "testing" this on, if its extreme AI / bad mods / bad scripting, you will never get good performance. The results from January til now for people enabling profiling and playing GOOD servers is 100% of all people that enabled it, came back to our place, and played, and commented got VASTLY higher FPS/smoothness....

There are places on some maps that will never be good... and when ya play ♥♥♥♥ like exile with 2 dozens mods -- you can live with 4/4/6 on view distance for ground/vic/air....

thinking you will max ♥♥♥♥ out is absurd, but if you don't EXTREME it all, performance profiling has helped several dozen people play with better results on our exile and antistasi servers....

i specifically made 3 different tests in case yaab wasn't enough for u even though yaab is considered golden standard even by bohemian quacks themselves, then i made it all in a single video showing that there wasn't any mods / settings or other tricks in it, and yet u are not satisfied, could u kindly make ur own perf benchmark with ur own rig runing yab then ? in both standard and profiling mod to show me and all of the rest of us that it actually works at the very least on ur rig ?

Originally posted by 76561198146637018:

Originally posted by MetalMilitia:

So, troll or whatever?

Dunno what yer "testing" this on, if its extreme AI / bad mods / bad scripting, you will never get good performance. The results from January til now for people enabling profiling and playing GOOD servers is 100% of all people that enabled it, came back to our place, and played, and commented got VASTLY higher FPS/smoothness....

There are places on some maps that will never be good... and when ya play ♥♥♥♥ like exile with 2 dozens mods -- you can live with 4/4/6 on view distance for ground/vic/air....

thinking you will max ♥♥♥♥ out is absurd, but if you don't EXTREME it all, performance profiling has helped several dozen people play with better results on our exile and antistasi servers....

i specifically made 3 different tests in case yaab wasn't enough for u even though yaab is considered golden standard even by bohemian quacks themselves, then i made it all in a single video showing that there wasn't any mods / settings or other tricks in it, and yet u are not satisfied, could u kindly make ur own perf benchmark with ur own rig runing yab then ? in both standard and profiling mod to show me and all of the rest of us that it actually works at the very least on ur rig ?

I mean if you don't want to take someone at their word with 26,000+ hours playing and dev'ing -- then I guess I'm just full of ♥♥♥♥. When I posted in January (I've been enabling profiling on and off since it STARTED) -- I had numerous other people come try it on several places we run/ran. Every single one of them got a performance bump, and FPS increase -- even people on old QUAD cores with 8gigs of RAM.

That benchmark is also last updated 2021....

It's entirely possible this last ARMA update pushed what profiling WAS doing from January til now, but what settings ya want specifically -- view/object/ultra?

Originally posted by MetalMilitia:

Originally posted by 76561198146637018:

i specifically made 3 different tests in case yaab wasn't enough for u even though yaab is considered golden standard even by bohemian quacks themselves, then i made it all in a single video showing that there wasn't any mods / settings or other tricks in it, and yet u are not satisfied, could u kindly make ur own perf benchmark with ur own rig runing yab then ? in both standard and profiling mod to show me and all of the rest of us that it actually works at the very least on ur rig ?

I mean if you don't want to take someone at their word with 26,000+ hours playing and dev'ing -- then I guess I'm just full of ♥♥♥♥. When I posted in January (I've been enabling profiling on and off since it STARTED) -- I had numerous other people come try it on several places we run/ran. Every single one of them got a performance bump, and FPS increase -- even people on old QUAD cores with 8gigs of RAM.

That benchmark is also last updated 2021....

yes ur words mean nothing to me when they have no basis behind it, if performance build was so great and helped so much people then surely it would be spread trough out community and everyone would know about it, especially after 6 months since jan, did it happen ? nope
xeon is quite bad cpu indeed for a game that relies on a single core performance and yet ur thread claimed to give atleast 20% not 0 but 20, that means that even on ♥♥♥♥♥♥ xeons performance there would atleast 20% increase according to ur words, hell even if it was 10 it would be fine but it gave 0, negative in 1 of yaab runs in fact
running tests on ultra ? i specifically tested it on high preset and durring whole bench gpu util didn't go higher than 70%
benchmark updated in 2021 what does it freaking have to do with anything ? it is still the MOST popular way to quickly and consistently test ur performance in arma 3, some dev u are the more u talk the more of a quack u seem

#edit, for some reason i doubled quoted, fixed it

Originally posted by MetalMilitia:

It's entirely possible this last ARMA update pushed what profiling WAS doing from January til now, but what settings ya want specifically -- view/object/ultra?

perhaps, but don't u think that u should at the very least test ur claims urself before sharing them ? especially as a "DEV"

Originally posted by 76561198146637018:

Originally posted by MetalMilitia:

I mean if you don't want to take someone at their word with 26,000+ hours playing and dev'ing -- then I guess I'm just full of ♥♥♥♥. When I posted in January (I've been enabling profiling on and off since it STARTED) -- I had numerous other people come try it on several places we run/ran. Every single one of them got a performance bump, and FPS increase -- even people on old QUAD cores with 8gigs of RAM.

That benchmark is also last updated 2021....

yes ur words mean nothing to me when they have no basis behind it, if performance build was so great and helped so much people then surely it would be spread trough out community and everyone would know about it, especially after 6 months since jan, did it happen ? nope
xeon is quite bad cpu indeed for a game that relies on a single core performance and yet ur thread claimed to give atleast 20% not 0 but 20, that means that even on ♥♥♥♥♥♥ xeons performance there would atleast 20% increase according to ur words, hell even if it was 10 it would be fine but it gave 0, negative in 1 of yaab runs in fact
running tests on ultra ? i specifically tested it on high preset and durring whole bench gpu util didn't go higher than 70%
benchmark updated in 2021 what does it freaking have to do with anything ? it is still the MOST popular way to quickly and consistently test ur performance in arma 3, some dev u are the more u talk the more of a quack u seem

#edit, for some reason i doubled quoted, fixed it

Ok well, I guess everyone I've told to enable it is fulla ♥♥♥♥.

30 fps on the demo is pretty crap so testing it on that processor is also terrible results.

I ran it r9/3900x/3080 50.1 with ultra (This rig is going on 5 years old I think?)
Ran it performance profiling, 52.7 in ultra

So 5.2% change same settings? No tweaks no mods.

My point is that is a ♥♥♥♥ "benchmark", that is creating a completely unrealistic scenario for most live servers -- the key being LIVE servers. Hosting rigs matter, mods matter, AI counts matter, scripting matters.... and the CPU speed yer runnin these tests on ain't worth a mention to benchmarks of any kind. A lot of ♥♥♥♥ that helps is mods like CHVD -- which is built into most modded servers so an individual can change view/object/ distances based on ground, vehicle, or air -- and scale them to perform better.

Every player I had try it told me noticeable difference in FPS and smoothness...

Again, I have no idea if they just added this to default ARMA for that big patch, as I've been working mainly on reforger assets, but I don't think like 75 people we had try it in those months were bullshitting about results.

Originally posted by MetalMilitia:

Originally posted by 76561198146637018:

yes ur words mean nothing to me when they have no basis behind it, if performance build was so great and helped so much people then surely it would be spread trough out community and everyone would know about it, especially after 6 months since jan, did it happen ? nope
xeon is quite bad cpu indeed for a game that relies on a single core performance and yet ur thread claimed to give atleast 20% not 0 but 20, that means that even on ♥♥♥♥♥♥ xeons performance there would atleast 20% increase according to ur words, hell even if it was 10 it would be fine but it gave 0, negative in 1 of yaab runs in fact
running tests on ultra ? i specifically tested it on high preset and durring whole bench gpu util didn't go higher than 70%
benchmark updated in 2021 what does it freaking have to do with anything ? it is still the MOST popular way to quickly and consistently test ur performance in arma 3, some dev u are the more u talk the more of a quack u seem

#edit, for some reason i doubled quoted, fixed it

Ok well, I guess everyone I've told to enable it is fulla ♥♥♥♥.

30 fps on the demo is pretty crap so testing it on that processor is also terrible results.

I ran it r9/3900x/3080 50.1 with ultra (This rig is going on 5 years old I think?)
Ran it performance profiling, 52.7 in ultra

So 5.2% change same settings? No tweaks no mods.

My point is that is a ♥♥♥♥ "benchmark", that is creating a completely unrealistic scenario for most live servers -- the key being LIVE servers. Hosting rigs matter, mods matter, AI counts matter, scripting matters.... and the CPU speed yer runnin these tests on ain't worth a mention to benchmarks of any kind. A lot of ♥♥♥♥ that helps is mods like CHVD -- which is built into most modded servers so an individual can change view/object/ distances based on ground, vehicle, or air -- and scale them to perform better.

Every player I had try it told me noticeable difference in FPS and smoothness...

Again, I have no idea if they just added this to default ARMA for that big patch, as I've been working mainly on reforger assets, but I don't think like 75 people we had try it in those months were bullshitting about results.

how come an offline benchmark that plays out same iteration of events from different perspectives is LESS REALISTIC than a ♥♥♥♥♥♥♥ live server where actual ppl play with mods and have random ♥♥♥♥ happening due to not being constant, ur diggin ur own grave mate the more u talk the deeper it gets, want to prove me wrong i give u last chance, make ur own video footage similar to mine in yaab, ♥♥♥♥♥♥ or not its the best CONSISTENT benchmark we have the most popular one too, then u will have something but empty words to reinforce ur claims

I also make these posts because it's sad to watch ♥♥♥♥ die out, especially when they have made progress with multicore usage and performance issues -- at least as CRAPPILY as they can after all these f'ing years.

Most people use built on mods, set their ♥♥♥♥ to 4/4/6k for ground/vehicle/air and go from there -- also with terrain built in options, and sync to 80% view options.... makes a massive difference for live servers.

The other thing about them updating this, it wrecks current servers -- often to the point where the people who run or have run them for years say "♥♥♥♥ it" I'm done and then hundreds disappear every patch, update, sitrep -- whatever. Fixing the ♥♥♥♥ gets old, but its love of the game, nostalgia, and some of the coolest people I met since a2dayz to epoch, to expoch, to exile, to antistasi.

Reforger is just getting there for those kinda mods, cause listening to ps5 kids and doing cap points is dumb.

Arma itself is dumb lol.. My playtime with no mods would be about 3 minutes since dayz beta.

Originally posted by 76561198146637018:

Originally posted by MetalMilitia:

Ok well, I guess everyone I've told to enable it is fulla ♥♥♥♥.

30 fps on the demo is pretty crap so testing it on that processor is also terrible results.

I ran it r9/3900x/3080 50.1 with ultra (This rig is going on 5 years old I think?)
Ran it performance profiling, 52.7 in ultra

So 5.2% change same settings? No tweaks no mods.

My point is that is a ♥♥♥♥ "benchmark", that is creating a completely unrealistic scenario for most live servers -- the key being LIVE servers. Hosting rigs matter, mods matter, AI counts matter, scripting matters.... and the CPU speed yer runnin these tests on ain't worth a mention to benchmarks of any kind. A lot of ♥♥♥♥ that helps is mods like CHVD -- which is built into most modded servers so an individual can change view/object/ distances based on ground, vehicle, or air -- and scale them to perform better.

Every player I had try it told me noticeable difference in FPS and smoothness...

Again, I have no idea if they just added this to default ARMA for that big patch, as I've been working mainly on reforger assets, but I don't think like 75 people we had try it in those months were bullshitting about results.

how come an offline benchmark that plays out same iteration of events from different perspectives is LESS REALISTIC than a ♥♥♥♥♥♥♥ live server where actual ppl play with mods and have random ♥♥♥♥ happening due to not being constant, ur diggin ur own grave mate the more u talk the deeper it gets, want to prove me wrong i give u last chance, make ur own video footage similar to mine in yaab, ♥♥♥♥♥♥ or not its the best CONSISTENT benchmark we have the most popular one too, then u will have something but empty words to reinforce ur claims

switching from ULTRA to STANDARD YAAB resulted in 53.2 and 55.1 fps respectively

... and yes, thank you for helping me dig the grave -- I got owned badly....

...when yer benchmarking on a processor is LITERALLY from 2012... and to quote a search on it heh...

"The Xeon E5-2690, while once a powerful processor, is generally considered terrible for modern gaming or everyday tasks due to its age and lower performance compared to newer CPUs."[/b]

So i'd assume they (bohemia) added what (some of) performance / profiling was doing, but also -- you should set your parameters in the launcher and test things as well. Maybe with the right tweaks you can breach that 30FPS some day.

Way to "own" me on all this, ya know -- informing people for months to enable it with positive results for pretty much EVERYONE who tried it. Helping fellow players, who personally thanked me for informing them, getting people involved each summer sale every year, or any time ARMA goes on sale, helping new folks....

Get a processor from this decade and maybe the results would suit you a bit better. Hell man, my 3900X is a dinosaur at this point...

----- here is what chat GPT had to say BTW regarding profiling ----

Performance Gains & Community Feedback

Reddit reports from late 2024 to early 2025 show significant improvements:

Many users report frame rate increases of ~20 fps in large-scale battles (~60–120 fps in scenarios that previously ran at 30–60 fps)
reddit.com
.

In mission benchmarks like YAAB, typical improvements of 8–20 % are observed
reddit.com
.

Multi-threading enhancements (asynchronous AI, lights, JIP processing, etc.) are central to these FPS gains

-------------

So ♥♥♥♥ like a live server, with controlled ai sizes and missions and running off something like we have, a 4.9ghz R9 w/ 16 gigs dedicated JUST to each server...

...those increases when you can control yer view distance, and optimize further settings can easily reach 20-70%.

----------------

But I am completely owned by someone benchmarking on a 2012 processor setup.

Full breakdown if you want to read it

Early 2024 (Profiling activated)

March 2024: First major profiling branch update (v08), optimizations in smoke rendering and AI multithreading.

Users saw 8–20% FPS boost in missions like YAAB
feedback.bistudio.com+15reddit.com+15steamcommunity.com+15
.

Some reported smoother large battles—even though frame ceilings remained
steamhost.cn/steamcommunity_com
.

Late 2024

November 2024: A high-vote Reddit post called the profiling branch "an absolute game changer," reporting ~20% frame rate increases and smoother missions, albeit with occasional instability .

Also in November: Players noticed that profiling didn’t raise peak FPS but improved the minimum FPS and stability—reducing stutters in heavy AI/multiplayer scenarios
dev.arma3.com+15reddit.com+15reddit.com+15
.

Crash reports: Some profiling builds (e.g., v6–v7) caused visual glitches or crashes, prompting frequent rollbacks
steamhost.cn/steamcommunity_com+6reddit.com+6olympus-entertainment.com+6
.

🟢 Early 2025 (Multithreading enhancements)

Jan 2025:

One user reported about 11% average FPS gain in YAAB benchmark
steamhost.cn/steamcommunity_com+4reddit.com+4reddit.com+4
.

Others saw "double-ish FPS" in specific scenarios, especially flying or large-scale environments
reddit.com+3reddit.com+3reddit.com+3
.

However, some users with heavy AI setups experienced no gain—likely hardware-limited or scenario-dependent
reddit.com
.

Profiling now optimized memory allocation, thread core affinity, and fixed certain Linux AI bugs (e.g., animation hang) .

June 2025 — Stable 2.20 Integration

June 17, 2025: Arma 3 update 2.20 (build 18897844) released, officially integrating profiling branch multithreading and performance improvements
reddit.com+15dev.arma3.com+15steamdb.info+15
.

Engine overhaul: moved to Enfusion job system (graph-based threading), enabling more efficient use of CPU cores
dev.arma3.com
.

Deprecation of 32-bit and support for newer compilers (C++23), further boosting performance