Turning Off CPU Cores Actually Help in Cities: Skylines 2

[yangzilata]

• In Short:

Can we actually boost Cities: Skylines 2’s performance by tweaking CPU cores, or is the game’s built-in core management already doing the best job?

---

Since Cities: Skylines 2 is super CPU-heavy, would manually adjusting CPU cores actually help? A lot of modern systems let us turn specific cores on or off, and some games actually run better when you tweak them. But is that the case here, or is it better to just let the game handle it?

- Pros & cons of manually adjusting CPU cores
- How much of a difference it actually makes
- Whether C:S2 benefits from parallel processing tweaks

References
​

I came across an interesting article titled "ADJUST YOUR LAPTOP'S P-CORES AND E-CORES FOR BETTER PERFORMANCE AND BATTERY LIFE" According to this article, disabling some CPU cores can improve performance in certain scenarios.

For example, they found that turning off some cores could lead to higher clock speeds, making apps load faster and improving performance in games like Final Fantasy XV and Shadow of the Tomb Raider. But the results depended on the task—some stuff like video editing still needed all cores running.

What Are Your Thoughts?​

• This got me wondering—does Cities: Skylines 2 actually benefit from this kind of manual CPU tweaking? Could disabling some cores improve performance, or does the game already manage this well enough on its own?
• Has anyone here tried adjusting CPU cores for C:S2? If so, did it make a noticeable difference?
• What setup worked best for you?

Since C:S2 is a pretty complex game, maybe disabling too many cores isn’t the best move. But at the same time, some games do benefit from this, so I’m curious.

Would love to hear your experiences!

A Part of Article​

• Disabling cores obviously comes at the cost of overall CPU performance, and it’s rare that you’ll want to go all the way down to only two Performance-cores. But disabling some cores can also improve performance in certain scenarios.
• We tested a number of different configurations and found that with two Performance Cores and all Efficient-cores disabled, the remaining four Performance-cores were able to boost to higher clock speeds, leading to higher performance in certain situations. Again, if we look at PCMark 10’s Modern Office test, we found four Performance cores were faster in the App Start-Up, Web Browsing, Spreadsheet, and Writing tests. That said, other tests — like Video Conferencing, Photo Editing, Rendering and Visualization, and Video Editing — performed better with the full might of the CPU, so the configuration you choose may depend on the task at hand.
• This setting can also improve performance in some games. In our testing we found the Final Fantasy XV benchmark performed about 3% better with only four Performance-cores enabled, while Shadow of the Tomb Raider achieved 3.5% higher framerates with this configuration in CPU-bound scenarios. Counter-Strike: Global Offensive, on the other hand, saw minimal performance changes between our different configurations, so your mileage may vary from game to game — though remember, these configurations also improved efficiency, so even if the performance jump is small, you may be able to game longer on battery too, which is a win-win.

Related Thread: If you're into this kind of stuff, I asked a similar question for Crusader Kings III too.

 

[MTO35]

• In Short:

Can we actually boost Cities: Skylines 2’s performance by tweaking CPU cores, or is the game’s built-in core management already doing the best job?

---

Since Cities: Skylines 2 is super CPU-heavy, would manually adjusting CPU cores actually help? A lot of modern systems let us turn specific cores on or off, and some games actually run better when you tweak them. But is that the case here, or is it better to just let the game handle it?

- Pros & cons of manually adjusting CPU cores
- How much of a difference it actually makes
- Whether C:S2 benefits from parallel processing tweaks

References
​

I came across an interesting article titled "ADJUST YOUR LAPTOP'S P-CORES AND E-CORES FOR BETTER PERFORMANCE AND BATTERY LIFE" According to this article, disabling some CPU cores can improve performance in certain scenarios.

For example, they found that turning off some cores could lead to higher clock speeds, making apps load faster and improving performance in games like Final Fantasy XV and Shadow of the Tomb Raider. But the results depended on the task—some stuff like video editing still needed all cores running.

What Are Your Thoughts?​

• This got me wondering—does Cities: Skylines 2 actually benefit from this kind of manual CPU tweaking? Could disabling some cores improve performance, or does the game already manage this well enough on its own?
• Has anyone here tried adjusting CPU cores for C:S2? If so, did it make a noticeable difference?
• What setup worked best for you?

Since C:S2 is a pretty complex game, maybe disabling too many cores isn’t the best move. But at the same time, some games do benefit from this, so I’m curious.

Would love to hear your experiences!

A Part of Article​

• Disabling cores obviously comes at the cost of overall CPU performance, and it’s rare that you’ll want to go all the way down to only two Performance-cores. But disabling some cores can also improve performance in certain scenarios.
• We tested a number of different configurations and found that with two Performance Cores and all Efficient-cores disabled, the remaining four Performance-cores were able to boost to higher clock speeds, leading to higher performance in certain situations. Again, if we look at PCMark 10’s Modern Office test, we found four Performance cores were faster in the App Start-Up, Web Browsing, Spreadsheet, and Writing tests. That said, other tests — like Video Conferencing, Photo Editing, Rendering and Visualization, and Video Editing — performed better with the full might of the CPU, so the configuration you choose may depend on the task at hand.
• This setting can also improve performance in some games. In our testing we found the Final Fantasy XV benchmark performed about 3% better with only four Performance-cores enabled, while Shadow of the Tomb Raider achieved 3.5% higher framerates with this configuration in CPU-bound scenarios. Counter-Strike: Global Offensive, on the other hand, saw minimal performance changes between our different configurations, so your mileage may vary from game to game — though remember, these configurations also improved efficiency, so even if the performance jump is small, you may be able to game longer on battery too, which is a win-win.

Related Thread: If you're into this kind of stuff, I asked a similar question for Crusader Kings III too.

I think this could definitely lead to better results in some applications. However, it's probably not advisable to try this with Cities Skylines 2, as beyond a certain population and complexity, the game turns into a kind of Cinebench R23 continuous stress test.
The only thing that helped me (Intel i9 13900K) achieve better and more consistent performance was adjusting the clock speeds of the P and E cores (I lowered the clock speeds for the P cores in the BIOS/UEFI and increased them slightly for the E cores), so that the system became more stable during gaming. With extreme CPU settings, I did achieve clock speeds of almost 6 GHz, but in my case this often led to the CPU heating up far too quickly and then clocking back down again, resulting in minor stuttering. Now I just play with a maximum clock speed of 5.4 GHz (which is usually not reached) on the P cores and around 4.6 GHz on the E cores with Intel specifications for PL1 and PL2, and the game runs much smoother than before.

 

[nilesism]

• In Short:

Can we actually boost Cities: Skylines 2’s performance by tweaking CPU cores, or is the game’s built-in core management already doing the best job?

---

Since Cities: Skylines 2 is super CPU-heavy, would manually adjusting CPU cores actually help? A lot of modern systems let us turn specific cores on or off, and some games actually run better when you tweak them. But is that the case here, or is it better to just let the game handle it?

- Pros & cons of manually adjusting CPU cores
- How much of a difference it actually makes
- Whether C:S2 benefits from parallel processing tweaks

References
​

I came across an interesting article titled "ADJUST YOUR LAPTOP'S P-CORES AND E-CORES FOR BETTER PERFORMANCE AND BATTERY LIFE" According to this article, disabling some CPU cores can improve performance in certain scenarios.

For example, they found that turning off some cores could lead to higher clock speeds, making apps load faster and improving performance in games like Final Fantasy XV and Shadow of the Tomb Raider. But the results depended on the task—some stuff like video editing still needed all cores running.

What Are Your Thoughts?​

• This got me wondering—does Cities: Skylines 2 actually benefit from this kind of manual CPU tweaking? Could disabling some cores improve performance, or does the game already manage this well enough on its own?
• Has anyone here tried adjusting CPU cores for C:S2? If so, did it make a noticeable difference?
• What setup worked best for you?

Since C:S2 is a pretty complex game, maybe disabling too many cores isn’t the best move. But at the same time, some games do benefit from this, so I’m curious.

Would love to hear your experiences!

A Part of Article​

• Disabling cores obviously comes at the cost of overall CPU performance, and it’s rare that you’ll want to go all the way down to only two Performance-cores. But disabling some cores can also improve performance in certain scenarios.
• We tested a number of different configurations and found that with two Performance Cores and all Efficient-cores disabled, the remaining four Performance-cores were able to boost to higher clock speeds, leading to higher performance in certain situations. Again, if we look at PCMark 10’s Modern Office test, we found four Performance cores were faster in the App Start-Up, Web Browsing, Spreadsheet, and Writing tests. That said, other tests — like Video Conferencing, Photo Editing, Rendering and Visualization, and Video Editing — performed better with the full might of the CPU, so the configuration you choose may depend on the task at hand.
• This setting can also improve performance in some games. In our testing we found the Final Fantasy XV benchmark performed about 3% better with only four Performance-cores enabled, while Shadow of the Tomb Raider achieved 3.5% higher framerates with this configuration in CPU-bound scenarios. Counter-Strike: Global Offensive, on the other hand, saw minimal performance changes between our different configurations, so your mileage may vary from game to game — though remember, these configurations also improved efficiency, so even if the performance jump is small, you may be able to game longer on battery too, which is a win-win.

Related Thread: If you're into this kind of stuff, I asked a similar question for Crusader Kings III too.

I noticed better framerates and simulation speed in CS1 when I disabled hyperthreading in the bios. I tested further by running in a VM (via unraid) and gave the instance only real cores and it would consistently be better outcomes (even with the hypervisor overhead).

I noticed on the release version of CS2 it had similar results, using only real cores and using PBO2 to slightly overclock those cores, resulted in better scores. However hyperthread had less measured impact.

It's been 1.5 years since then and I have not reproduced the test with the anniversary update.

In the other thread we have been able to identify a shortcoming in 1x speed in that it does not scale to use all cores (only 4x and 8x do).

So it makes sense. If you play 1x, you may be better off giving the game ~10 of your fastest cores and hiding the others.

With AMD you can "draw" a curve that will apply a scaling adjustment in addition to limiting the maximum wattage. I found adding small voltage more effective for multithread scores than limiting total wattage.

 

[yangzilata]

I lowered the clock speeds

value?

 

[yangzilata]

adding small voltage more effective

what temp does it reach?

 

[MTO35]

value?

No, I meant the clock speeds of the individual cores.
So, I carefully tried adjusting the processor affinity for Cities Skylines 2 via the task manager and later with a tool (Core Director) so that with my processor, only 4 P-cores and 8 E-cores were available for the game. The game started and I got into the main menu, but when loading an existing save game or a new one, the game froze immediately after loading with a UnityEngineError message in both attempts.
I actually wanted to test what would happen if I only assigned P cores and left out E cores. But then I didn't test it further.
I even tried making only E-cores available to the application out of curiosity, and the game simply wouldn't start at all.
Maybe halving the cores was already too much, or I could have configured it in the UEFI. But I can't say whether my system would have still run stably.

 

[nilesism]

what temp does it reach?

I dont have the board set up any more. It was a 5950x with a ASUS x470f then later a MSI B550. The MSI had two power ports at the top of the board and larger VRM heatsinks so I assume the board itself draws more power.

My understanding of PBO2 is it significantly reduces temperatures (power draw). I used a Noctua DH-15 140mm dual tower dual fan air cooler. I think I went pretty low with the power limits. In short, you pick the temp your cooler or A/C can tolerate and the chip will scale.

My experience with a 5800x3D is that chip was already tuned from the factory and had very little headroom for overclock or PBO2. The 5950x however started at Cinebench r23 scores in the 21k range and reached 33k with the voltage boost and curve.

This translated directly to more population in CS2 before things dropped into the .2 smooth speed range, from memory firmly 30% but up to 60% more population depending on your tolerance to jitter / fps.

Here is a guide pretty similar to my setup with a lot of the background.

Overclocking AMD Zen3 with PBO2 Done Right: A full guide

Overclocking Ryzen 5600X/5800X/5900X/5950X using PBO2This guide aims to overclock Zen3 CPUs with Precision Boost Overdrive 2 (PBO2), gaining extra performance in the single-core score(a

cubicy.icu

The only difference is with CS2 I ended up having to add voltage by a very small amount (I want to say to the chipset) for stability rather than reduce. The chip would pass Corecycler, Cinebench and prime95 but then crash in game without a little more. I cant remember which setting or how much, and it was different between the two boards. I followed the suggestions in the video referenced here, but the full original video from the AMD technical marketing manager is worth a watch.

I would wager that by doing a per-core adjustment I was basically doing what was suggested in OP (making some cores work harder, making the game favor the cores that score the best). That said, I never tried entirely disabling some cores and making the best performing work even harder.

 

[yangzilata]

No, I meant the clock speeds of the individual cores.
So, I carefully tried adjusting the processor affinity for Cities Skylines 2 via the task manager and later with a tool (Core Director) so that with my processor, only 4 P-cores and 8 E-cores were available for the game. The game started and I got into the main menu, but when loading an existing save game or a new one, the game froze immediately after loading with a UnityEngineError message in both attempts.
I actually wanted to test what would happen if I only assigned P cores and left out E cores. But then I didn't test it further.
I even tried making only E-cores available to the application out of curiosity, and the game simply wouldn't start at all.
Maybe halving the cores was already too much, or I could have configured it in the UEFI. But I can't say whether my system would have still run stably.

It’s cool to learn how different setups can impact CS:2

 

[yangzilata]

I dont have the board set up any more. It was a 5950x with a ASUS x470f then later a MSI B550. The MSI had two power ports at the top of the board and larger VRM heatsinks so I assume the board itself draws more power.

My understanding of PBO2 is it significantly reduces temperatures (power draw). I used a Noctua DH-15 140mm dual tower dual fan air cooler. I think I went pretty low with the power limits. In short, you pick the temp your cooler or A/C can tolerate and the chip will scale.

My experience with a 5800x3D is that chip was already tuned from the factory and had very little headroom for overclock or PBO2. The 5950x however started at Cinebench r23 scores in the 21k range and reached 33k with the voltage boost and curve.

This translated directly to more population in CS2 before things dropped into the .2 smooth speed range, from memory firmly 30% but up to 60% more population depending on your tolerance to jitter / fps.

Here is a guide pretty similar to my setup with a lot of the background.

Overclocking AMD Zen3 with PBO2 Done Right: A full guide

Overclocking Ryzen 5600X/5800X/5900X/5950X using PBO2This guide aims to overclock Zen3 CPUs with Precision Boost Overdrive 2 (PBO2), gaining extra performance in the single-core score(a

cubicy.icu

The only difference is with CS2 I ended up having to add voltage by a very small amount (I want to say to the chipset) for stability rather than reduce. The chip would pass Corecycler, Cinebench and prime95 but then crash in game without a little more. I cant remember which setting or how much, and it was different between the two boards. I followed the suggestions in the video referenced here, but the full original video from the AMD technical marketing manager is worth a watch.

I would wager that by doing a per-core adjustment I was basically doing what was suggested in OP (making some cores work harder, making the game favor the cores that score the best). That said, I never tried entirely disabling some cores and making the best performing work even harder.

Even small changes can have such a big impact. Appreciate your detailed approach!