i7 12700k to i9 14900k

[River]

Would this help a little or a lot in Cities Skylines 2 ?

I play at 3440x1440.

Video card is RTX 4090 if that matters sorry.

[taloN]

Ursprünglich geschrieben von gamer:

Would this help a little or a lot in Cities Skylines 2 ?

I play at 3440x1440.

Video card is RTX 4090 if that matters sorry.

If you don't know how to tweak the advanced graphics settings, it won't matter what kind of hardware you use.

[enderbeta]

Ursprünglich geschrieben von ¤ OpN ¤:

might as well 7950x3d..... When really a 13700k or 7800x3d are better bang for your buck the upgrade isnt really there but if you want the end of life chip before you have to upgrade the board again next gen then go for it.

Based on my testing with a i9-10900K, i9-11900K and a Ryzen 9 5950X it seems the more cores the better for the simulation. The rendering of the game in terms of FPS seems best on the i9-10900K. It’s a strange situation. I think overall the 5950X is only 4FPS slower on average but it runs cooler and has more headroom for the simulation.

I’m not sure I would bother with the 3D variant of the 7950X. In fact I would hang back and see if the 9950X3D has 3D cache on both chiplets.

I’m not sold on the 3D cache variant in the long term. Maybe if both chiplets had the 3D cache I would be more excited. For now I’m sticking to my 5950X. It’s holding out at 167,000 people in my city. Interestingly enough the i9-10900K is holding out just as well but it’s using everything it has to keep up. Considering it is a older design it’s impressive

[GnuHorn]

Given the processing profile for the game (it does spread out well across cores, and will have lots of repeat workload) the best bang for the buck would be a 5800X3d or 7800X3d when it comes to processor selection. I'm running it on a 5800X3d and my biggest annoyance is simulation bugs because the simulation itself with a 3080 Ti even at 100k may not have great performance, but not really experiencing much in the way of stuttering that others have experienced.

Part of it's my cooling solution is keeping me below the throttle point, another part of it is that this is exactly the kind of game that 3d v-cache would excel at. Much easier to handle the excessive API calls when most of the memory it's referencing most frequently will sit on the chip cache.

[enderbeta]

Ursprünglich geschrieben von GnuHorn:

Given the processing profile for the game (it does spread out well across cores, and will have lots of repeat workload) the best bang for the buck would be a 5800X3d or 7800X3d when it comes to processor selection. I'm running it on a 5800X3d and my biggest annoyance is simulation bugs because the simulation itself with a 3080 Ti even at 100k may not have great performance, but not really experiencing much in the way of stuttering that others have experienced.

Part of it's my cooling solution is keeping me below the throttle point, another part of it is that this is exactly the kind of game that 3d v-cache would excel at. Much easier to handle the excessive API calls when most of the memory it's referencing most frequently will sit on the chip cache.

Do you think the 3D cache of the 5800X3D would outweigh the extra 8c/16t of the 5950X?

[c0hKaine]

4090? 40 frames is what im offering take it or leave it

[GnuHorn]

Ursprünglich geschrieben von enderbeta:

Ursprünglich geschrieben von GnuHorn:

Given the processing profile for the game (it does spread out well across cores, and will have lots of repeat workload) the best bang for the buck would be a 5800X3d or 7800X3d when it comes to processor selection. I'm running it on a 5800X3d and my biggest annoyance is simulation bugs because the simulation itself with a 3080 Ti even at 100k may not have great performance, but not really experiencing much in the way of stuttering that others have experienced.

Part of it's my cooling solution is keeping me below the throttle point, another part of it is that this is exactly the kind of game that 3d v-cache would excel at. Much easier to handle the excessive API calls when most of the memory it's referencing most frequently will sit on the chip cache.

Do you think the 3D cache of the 5800X3D would outweigh the extra 8c/16t of the 5950X?

Hmm... That's a really interesting question.

I don't know for sure and wouldn't feel comfortable answering definitively without a benchmark. Typically the 3d v-cache is superior, even with more cores in a workload like this. Where you really see the advantage for more cores is productivity software.

Think about it like this:

Whenever the processor is running code, it identifies via algorithm which code is more likely called. That stays in the v-cache if it fits. So for a workload where you have lots of the same processes running over and over again - which happens frequently in games - the v-cache is superior because it doesn't incur a request to memory or storage to request that repeat data. So in an ideal situation, you're probably getting performance gains that would be superior to doubling the core count because of the how fast the cache return is.

Any workload where you have high multi-threading capability and where you have unique memory for every operation (editing/processing a video, running a database of large data sets, etc) is likely to benefit more from more cores.

So my gut is that the v-cache would be superior, but at the same time more cores will help you with some of the multi-threading workloads (meaning less thread contention). Having said that I'm not sure how well it scales beyond 8 cores, and the answer to that would likely give us some more insight as to the answer.

But each time the CPU can stay in the CPU package, the better the performance and the more bang you get for each core. So theoretically I would say the v-cache might be better for this workload.

[River]

Good to hear!

[Propeople]

I have a 12700 @ 5000 ghz all cores.....4090
150k people that get 100% at 3x speed..... 70% at 2x speed....

[Propeople]

Ah the enser is keep the 12700k until DDR6 Or pci 6.0.....

[Cosmic Sea]

Hardly any difference. Can maybe squeeze out an additional 25k pop before it craps out again.

[Propeople]

ah the test shows a little diference of 15% from 14700k to 12700k has videos on youtube