Benchmark test scores

I was just wondering what benchmark test scores others got and what they are running.

1172.4

Rtx3070
I7 5820k
16gb ram
Ssd hard drive

There was an old topic with lots of benchmarks scores, but it think its really hidden now, if you want higher scores you need a stronger cpu to pair with that gpu.

Current patch version my score is 1227 with ryzen 7 and rx 5600xt it was higher few patches ago, but performance has decayed with latest versions it seems.

1 Like

I don’t remember exactly but it was around ~1300.

Ryzen 5900X
32gb 3600Mhz CL16 RAM
(synced to infinity fabric)

up to date bios and hardware drivers
clean install of Windows
no background tasks open

RTX 3070
SSD
(but mostly the CPU matters…)

1 Like

You have to remember that the engine can only use 1 core of the PC. There are a lot of good modern CPUs that have relatively weak individual cores.

That is not true as far as I know. It’s not the Age of Kings engine. You don’t even know what technologies are used exactly, if you don’t start to decompile the app. A msvc app takes advantage of a multicore cpu by default nowdays.

Take into consideration, that your FPS are locked to your monitor in many cases.
I have a ryzen 3700 and rx470 4gb and 13xx with the current patch.
If I use the default AMD driver settings and a 60hz screen I have a score of 11xx.

In general the benchmark test is poorly designed, because

  • people can pass the test, but run out of video or main memory in a 4v4 and provoce the yellow clock
  • people can runs apps in parallel, most commenly voip, browser and streaming software, and can even provoce the yellow clock in dark age, despite passing the benchmark test
  • people can circumvent the benchmark test by using hd mods, while setting everything to mininum

The only way to make a statement wether a user can run the game or not, is to log the computation time for every tick while playing ranked and decide on this, if the player is legit or not.

It is actually still the genie engine from the original game. This came up a lot in the beta.

I mean, it’s obviously a re-engineering of the aok genie engine, but how do you know that it’s still compiled to run on single core, if it’s an msvc app?
Also there are new features, like the “catch back to simulation” if you drop internet for a few secs, or the change of package routing and tickrate… so why are you sure that there are no concurrencies running multicore now?

You are just guessing.

I restricted the game to run on 4, 2, or 1 core by os and got a massive performance degraduation. Like going from 250fps with 16 cores to 50fps with 1 core with a post imp save game.

I’m not guessing… this was confirmed by the dev team during the beta as well when several people tested how the game was running on each of their components and noticed that it was only pulling from 1 core. It’s not an updated engine, it’s literally the exact same one.

Even if some code, you refer to be the engine, may be the same one, the app is not the same and certainly benefits from multicore cpus.
Or do you think Ensemble Studios time traveled into the future and back, just so that the Definitive Edition 20 years later uses MS WPF to mess up the GUI?

Is this the beta?

You really think they use the binary code of AoC 1.0c for Definitive Edition 11?
As soon as they click on the build button in Visual Studio magic things happen and the app runs faster on a multicore CPU, especially if Winforms/WPF/UWP is involved for the GUI. Even if the same code is used for the game simulation.

The game is not optimized to perfection to make use of high end hardware, that is for sure.
But it’s not the same like AoK/AoC.

Well the discussion about de only 1 core wasn’t only for the beta, it went on afterwards since that wasn’t changed. If that was changed in a patch I’m pretty sure someone would have brought that up. The only thing I can remember from patchnotes is when the improved vram allocation.

Not just that. The game is super poorly optimized.

Typically single core performances increases each generation by quite a high difference. In fact this is the main reason why people buy newer generation’s cpus. So not sure what cpus you are talking about.

I upgraded mine for this reason. Modern is likely misleading here. I mean modern as in within the past five years. I had an older i7 for example that could run most games fine but did have performance issues on DE

You can look it up… you’re being hyper dismissive of something that is very well documented.

The genie engine is the same source code for aoe2 DE, hd and vanilla, you can find the genie editor in the tools section in the game files.

DE is not a newer game, it is just massive game mod, the code has been changed but those changes are still minimal, just few tweaks here and there, lets say that making the game relay server based instead of peer to peer has been the biggest change to the code, cause the quality life improvements have been successfully ported to the old game since those changes don’t require changes to the source code.

The game uses two threads, cause sound, device inputs, network all of those are being done on another thread or core, but the game simulation only runs in one single core/thread, if you lock the game to run in one of them only, then the performance will be hurt.

Aoe2 crashes are really often on any game version, the most common crash is when one core reaches 100% usage, the simulation always uses up to 90% of the load, if any other thing is running and using that very same core then it might hit the 100% and insta crash or in the best cases just massive lag.

So far the only game that got a different treatment was aoe 1DE, during development they made it multicore, that game doesn’t even stress the cpu, despite using the same genie engine, but there were some issues and lack of cash so the original team was disbanded after that remake and they left that game without further support or improvements, despite being superior than the aoe2 DE version, i am quite sure they are using a different version of the genie engine for aoe2 DE cause water really moves in aoe 1 DE but in aoe2 DE is just a tricky bubble effect that works as a layer to the texture(the same that the hd version has), but in my test for aoe1 DE i never experienced the same lag behavior in long games compared to aoe2.

Then send me the docs, please.
Because I can’t find them.

Again it doesn’t mean that they did not build the app with a recent compiler and it also does not mean that did not changed things, even when using vast parts of the original source code.
Quake III is considered to have a different engine than Quake II, despite it uses much of it’s code.
At this point it’s more about language.

Well you are actually saying that some engine related things are spread between threads, which was not the case in 1.0c.
Also the rendering and GUI was remade also with new technologies which most likely take advantage of many CPU cores.

Stop with it, the game uses one core/thread for the simulation(playing), that is what matters, it is not multicore, if you get less performance is due sound and input devices in the previous version the cpu had less load cause of light files, now they weight 10-20 times more, it has been discussed during the beta and first months, you are a bit too late.

The game has the same performance hit than all the previous versions, it can’t handle 4x4 with trade.

Sorry, I really don’t see that.
I have a ryzen 3700 and an rx470 and i always have 100+ fps, even in a 4v4 with 1600 pop.
But it’s super hard to find a ranked game, where other people don’t provoce the yellow clock by that stage of the game.

I start to truggle on 8*225 = 1800 pop black forest with choke points and hand cannoneers.

I am not denying that, nor that this is a severe bottle neck for many people.

Still the things which are worked around the ‘simulation’ (network, rendering, gui, sound) may most likely profit from a multicore cpu.