which CPU for rimworld？

[Shooting star]

5800X3D? 13600KF?or wait for 7800x3d？
I'm considering upgrading from a 3900x to a 5800x3d or 13600KF D4，but I can find no user reported information or benchmarks that highlight the performance increase seen or expected.

[Steelfleece]

You'll be looking at how well each individual core performs, so I'd wager the Intel over the AMD in that comparison.

[Wasted]

AMD Ryzen 7 5800X3D probably makes the most sense for OP as its about $320 and will work with existing AM4 socket/build.

for a new value build i suggest:

AMD 7600X / 7700x
or
Intel 13600K(f) / 13700k

probably best value options for rimworld , it really depends on how much the cost is for the total build.

Ryzen 7 7800X3D will arrive on April 6. if you can buy it immediately when it comes out its probably the option i would take if i was building a new rig. For OP and anyone with existing AM4 socket it just depends if you want a new pc or not. for just rimworld 5800x3d cpu/cpu cooler swap will probably give you what you need.

[corisai]

Intel for gaming of course. Single core performance >>> many cores.

[brian_va]

Recently got an 11600kf, works well. Id imagine the 13600 would be better still. It'll be interesting to see how amds x3d chips work out and if Intel responds.

[Zane87]

Originally posted by Shooting star:

5800X3D? 13600KF?or wait for 7800x3d？
I'm considering upgrading from a 3900x to a 5800x3d or 13600KF D4，but I can find no user reported information or benchmarks that highlight the performance increase seen or expected.

The 5800x3D looks very promising.
Personally got a water-cooled 13700k last month and my TPS have more than doubled compared to the previous state of a 10750H.

Very pleasantly surprised, but I don't doubt the AMD one would rock as well.

[Mr.Elendig]

if you want to switch platform, am5 kinda makes more sense than 13th gen intel.

[VoiD]

recently got an i9 13900k, it made a world of difference in poorly optimized games (paradox, rimworld with very big late game colonies, etc...)

[Zane87]

Originally posted by VoiD:

recently got an i9 13900k, it made a world of difference in poorly optimized games (paradox, rimworld with very big late game colonies, etc...)

Oh so damn true. Stellaris too is on a whole different level now and so much more enjoyable to actually continue playing.

[XIII]

I dont think you need to buy a CPU for Rimworld, assuming yours isnt utterly ancient. This game has pretty low requirements.

Like people recommending 5800 3Ds and 13900Ks gotta be on drugs. What are you doing, 1000 colonist playthroughs?

Originally posted by Steelfleece:

You'll be looking at how well each individual core performs, so I'd wager the Intel over the AMD in that comparison.

Originally posted by corisai:

Intel for gaming of course. Single core performance >>> many cores.

Also, the idea that Intel CPUs always have more per core performance than AMD is just gibberish. Both of them have mostly 6-8 core CPUs, so obviously they are neck to neck with per core performance (and no E-cores dont matter for games).

Its like your thinking of Bulldozer CPUs from 10 years ago. With Ryzen its not really a thing.

[Zane87]

Originally posted by XIII:

I dont think you need to buy a CPU for Rimworld, assuming yours isnt utterly ancient. This game has pretty low requirements.

Like people recommending 5800 3Ds and 13900Ks gotta be on drugs. What are you doing, 1000 colonist playthroughs?

Originally posted by Steelfleece:

You'll be looking at how well each individual core performs, so I'd wager the Intel over the AMD in that comparison.

Originally posted by corisai:

Intel for gaming of course. Single core performance >>> many cores.

Also, the idea that Intel CPUs always have more per core performance than AMD is just gibberish. Both of them have mostly 6-8 core CPUs, so obviously they are neck to neck with per core performance (and no E-cores dont matter for games).

Its like your thinking of Bulldozer CPUs from 10 years ago. With Ryzen its not really a thing.

Lategame performance on large colonies, especially modded ones, can be pretty bad even with just 30-50 pawns, depending on what else you use.
Each mod generates overhead as well, drastically reducing headroom for stable 60 TPS, and for most of us that's not enough either.


The last part about the CPU core count as an argument for architectural differences that result in better or worse per core performance is.... Utterly meaningless. That's not an argument for anything, at all.

Use benchmarks like cinebench or Passmark to get an idea (and please don't start "all ...mark benches are favoring Intel yadda yadda", use multiple benchmarks to compare). Right now, the rocket lake chips and 7950x are extremely good in single core performance.

What you can't do is compare just the core count or clock speed. It's not working on the same chip manufacturer between generations and it's even less of a measure for comparing completely different architectures. It's honestly just not a measure for anything at all.

[XIII]

Originally posted by Zane87:

Lategame performance on large colonies, especially modded ones, can be pretty bad even with just 30-50 pawns, depending on what else you use.
Each mod generates overhead as well, drastically reducing headroom for stable 60 TPS, and for most of us that's not enough either.

I guess if you massively overload your game with mods, you might get issues no matter what you do. Even then high end CPUs will be limited in effect despite being 3 times as expensive and much harder to cool.

But frankly, I think at >20 pawns Rimworld already crumbles. At that point the personality of the pawns goes missing in the crowd and you might as well play Dwarf Fortress, thats way better with large colony management.

Rimworld shines when the charachters of the pawns stand out and its actually important to keep certain pawns stable, if they arent just replaceable. I think thats also the way the game is intended to work.

The last part about the CPU core count as an argument for architectural differences that result in better or worse per core performance is.... Utterly meaningless. That's not an argument for anything, at all.

Use benchmarks like cinebench or Passmark to get an idea (and please don't start "all ...mark benches are favoring Intel yadda yadda", use multiple benchmarks to compare). Right now, the rocket lake chips and 7950x are extremely good in single core performance.

What you can't do is compare just the core count or clock speed. It's not working on the same chip manufacturer between generations and it's even less of a measure for comparing completely different architectures. It's honestly just not a measure for anything at all.

Yeah, that makes no sense because you completely missed the point of what I said.

AMD and Intel CPUs are competitive in performance, despite having the same core counts. That implies that the cores are effectively of similar speed in games.
And thats just how it works out in games, even those with lack of multi-threading.

If you know the history, the idea of Intels faster cores mostly comes from AMDs early attempts at 6-8 weaker cores, when Intels mainstream stuff was still at 4-6 cores.

If we wanna be really specific, then cinebenchmark isnt a great evaluator for games performance either.

[corisai]

Originally posted by XIII:

Rimworld shines when the charachters of the pawns stand out and its actually important to keep certain pawns stable, if they arent just replaceable. I think thats also the way the game is intended to work.

Except Biotech DLC is all about being replaceble - children & mechanoids :)

Originally posted by XIII:

AMD and Intel CPUs are competitive in performance, despite having the same core counts. That implies that the cores are effectively of similar speed in games.

Wrong - you need to run a dedicated single core tests to say it for sure.

[XIII]

Originally posted by corisai:

Except Biotech DLC is all about being replaceble - children & mechanoids :)

That ment "replaceable" in the sense of "I got so many pawns it doesnt matter if one or two die". Kills the drama of the game.

Originally posted by XIII:

AMD and Intel CPUs are competitive in performance, despite having the same core counts. That implies that the cores are effectively of similar speed in games.

Wrong - you need to run a dedicated single core tests to say it for sure.

What are you even talking about? You can just look at games benchmarks, and that shows why your original claim of Intel being better for games is clearly wrong. Thats a starting point.

[Zane87]

Originally posted by XIII:

I guess if you massively overload your game with mods, you might get issues no matter what you do. Even then high end CPUs will be limited in effect despite being 3 times as expensive and much harder to cool.

Sure, but that doesn't mean that a better CPU significantly shifts that threshold and thus allows for a better experience regardless.

Originally posted by XIII:

But frankly, I think at >20 pawns Rimworld already crumbles. At that point the personality of the pawns goes missing in the crowd and you might as well play Dwarf Fortress, thats way better with large colony management.

Rimworld shines when the charachters of the pawns stand out and its actually important to keep certain pawns stable, if they arent just replaceable. I think thats also the way the game is intended to work.

I do agree that this is the intention of the game and Tynan's vision. And if that's how you play, all power to you really.

(Un)fortunately it also is a pretty great colony sim and that's how personally I enjoy it much more. So for me, bigger colonies are a lot more fun and that's where the lag begins. Better CPU helped me really a lot here.

Originally posted by XIII:

Yeah, that makes no sense because you completely missed the point of what I said.

AMD and Intel CPUs are competitive in performance, despite having the same core counts. That implies that the cores are effectively of similar speed in games.
And thats just how it works out in games, even those with lack of multi-threading.

That's kinda the status quo with current gen CPU, in the past it was very different, with oftentimes Intel winning the Single core game and AMD utterly rocking the multi core performance benchmarks.

And not every user will or can afford current gen CPU. Hence my remark to not just go by feeling but by hard measures/benchmarks.

Originally posted by XIII:

If you know the history, the idea of Intels faster cores mostly comes from AMDs early attempts at 6-8 weaker cores, when Intels mainstream stuff was still at 4-6 cores.

Absolutely true, I totally agree here

Originally posted by XIII:

If we wanna be really specific, then cinebenchmark isnt a great evaluator for games performance either.

And here as well. Still because of the whole "mark bench Intel bias" thing I recommend to include it for less alleged bias to AMD. But that's why I meant use multiple benchmarks.

Even then I wish more YouTubers and tech channels would use Stellaris, RimWorld and such with fixed save games and multiple runs (to account for random events in those games) for their CPU tests and not just your usual action game that nearly never is CPU bottlenecked.

Until this eventually (probably never) happens we kinda are stuck with abstract benchmarks that aren't very realistic. But it's better than nothing to get an idea of comparative advantages to other CPU and that's actually pretty neat

[XIII]

Originally posted by Zane87:

That's kinda the status quo with current gen CPU, in the past it was very different, with oftentimes Intel winning the Single core game and AMD utterly rocking the multi core performance benchmarks.

And not every user will or can afford current gen CPU. Hence my remark to not just go by feeling but by hard measures/benchmarks.

When someones says "just go with Intel" for older CPUs, then thats the same as going by feeling. Its not an informed way to make a buying decision.

And here as well. Still because of the whole "mark bench Intel bias" thing I recommend to include it for less alleged bias to AMD. But that's why I meant use multiple benchmarks.

What do you mean, User-Benchmark? Thats site is actually moronic, anyone using it pretty much disqualifies itself. Even redditors have learned that lesson at this point.

But no benchmark programs is overly useful for game performance.

Even then I wish more YouTubers and tech channels would use Stellaris, RimWorld and such with fixed save games and multiple runs (to account for random events in those games) for their CPU tests and not just your usual action game that nearly never is CPU bottlenecked.

Until this eventually (probably never) happens we kinda are stuck with abstract benchmarks that aren't very realistic. But it's better than nothing to get an idea of comparative advantages to other CPU and that's actually pretty neat

Pretty much any competent CPU benchmark makes sure the game is CPU-bottlenecked for the tests, thats why they turn down resolution and graphics.