Benchmarks and Performance comps

[ruhrbaron]

Dear PDX, sadly the beloved map games dont count as triple A titles yet so they typically dont make it in the journalistic hardware reviews. I think it'd be helpful if you stepped in to fill this gap. Doing so would address three problems in my mind:

a) the forums are full of questions along the lines of which hardware to get and a comprehensive database could help here greatly, also for us who think about upgrading but arent sure whether it is worth it
b) it would give the developers and us, the customers, a better idea of how the games run on our systems.
c) help us all to get an idea how the games run across different patches and hardware configurations over time and just might help you spot weirdness that you otherwise would only spot after a slow process of customer feedback. I think a lot of confusion in the performance debates could be addressed if said data was available.

i got this idea because Level1techs briefly mentioned in his AMD R7950X3D review that this CPU runs stellaris by far the best among all CPUs he has ever tested but frustratingly there's no data on that.

I would never expect you to test all the hardware but having low, mid and high end systems out of the various vendor combinations for each of the past 3-5 generations of hardware generations or so to compare across different resolutions, patches, cpus, gpus, ram, OSes and game combinations would be great.

I know you already are doing a bunch of internal testing for all kinds of purposes, so why not share the results with the community helping us make better hardware purchasing decisions

 

[KarroK]

Honestly, I'm perplexed that there isn't an built-in standard benchmark mode for Stellaris. This is why it's never benchmarked, because there is no uniform easy way to get proper metrics!

If CPU reviewers could just pop in a stellaris.exe --benchmark it would be right there next to factorio. It would allow players to benchmark their own systems, see what new hardware brings and bring a lot of "free" exposure of the game to the general public. Additionally development could use that same benchmark to check release candidate performance and look for any regressions.

so In short, a proper inbuilt benchmark would be:
- a marketing win
- a player win
- a stellaris-developer win!

Feels to me like a no-brainer that should've been present from the start, but as the saying goes, the best moment to plant a tree was 10 years ago, the second best moment to plant a tree is today.

 

[D313.m]

Just jumping in to say that, specifically for Stellaris, there's a way to "benchmark" the game using the "one_year" console command. I devised some standardised settings and created a standardised save file so that we can compare performance across different systems.

I originally publicised the "Paradox benchmarking Initiative" through the game subreddit, but this forum post is the top results for stellaris benchmark and I'd love for more people to participate in it.

The results and instructions are available in the Initiative's Discord server, but I understand some people may not want to join, so I created this Google Doc with the instructions. The doc contains a link to the results of the previous iteration of the benchmark (for v3.6.1), and will be updated as data is gathered for the v3.8.4 Gemini (current as of writing this) version.

Hopefully, one day a better first-party solution will be available, but for now this is what we have.

 

[elektrizikekswerk]

In Vic3 forum there is this thread:

Victoria 3 Performance Benchmark

There's been a lot of discussing Victoria 3's performance. Some people say it runs perfectly smooth the whole way through, others say it's unplayable after X number of years. Presumably the truth is somewhere in the middle. I decided to try and...

forum.paradoxplaza.com