FM24 Performance Benchmarking Thread

[Bainsy]

4 hours ago, yohanes1985 said:

Benchmarks 1 and 3 are Single Threaded tests, the results are not far from the M4 Pro, this is why the M4 chip is special in single thread scores.
I ran Benchmark 1 three more times to make sure there were no errors, my worst result was 26 seconds, still quite fast.
00:25.62
00:25.91
00:26.24

I have ran a geekbench benchmark and getting the same results as others, so thankfully the performance issues aren't down to the hardware. Will have to try a fresh install of fm. 

[Brother Ben]

What version of MacOS are you both running?

[Bainsy]

1 hour ago, Brother Ben said:

What version of MacOS are you both running?

15.1.1 as of 2 days ago, I believe my original benchmarks were done on 15.1 

[yohanes1985]

2 hours ago, Brother Ben said:

What version of MacOS are you both running?

macOS Sequoia 15.1.1 (24B2091)
Fresh installed FM24 from steam, no logo kits etc...

[Brother Ben]

Plot thickens.  Maybe its the graphics addons?  I'd be shocked if it slowed it down that much though

[Bainsy]

Fresh install and still getting the same results as previous.  Only other difference to @yohanes1985 is my FM is from EPIC instead of steam, unsure how that could possibly make that big of a difference though. 

[jwchriste]

Have either of you disabled the epic/steam overlays? I know you can in steam. Not sure how epic works. Maybe just having them enabled, even though they might not be on/visible, makes a difference? Just trying to think what could be different between epic & steam versions once you launch the game. Maybe epic is collecting data or some other data processing that steam isn't while the game is simulating?

[JAwtunes]

Is caching switched on in preferences? Also windowed Vs full screen. Not sure how much these things affect processing times.

[gizmo11x]

On 21/11/2024 at 11:13, caz said:

That would be my initial reaction as well....though it does seem a number of simulation and CPU bound titles also benefits from the 3d cache. Stellaris is the most used example in reviews/benchmarks.

But I went back to the previous resutls, and from there is does not look like the 7800x3d is significantly better than the non-x3d chips from same gen Ryzen CPUs....so seems to be more about number of cores and speed per core.

From what I've heard/read the non X3D 9000 series CPUs have a memory controller bottleneck in games. The large, fast L3 cache on the 9800X3D basically bypasses that bottleneck so simulation times are improved as well.
This wasn't the case with the 7000 series so the x3d chips didn't have much advantage (or were slower) compared to non-x3d in Stellaris and other simulation time benchmarks.

Based on this I suspect the 9800x3d would be pretty fast in FM.