user.cfg trick lower CPU usage doesn't work - hurts FPS on Intel E-Cores

I'm sure some of you are aware that a lot of youtubers/redditors have been going around talking about a magic cure for "high CPU usage".

You put a 'user.cfg' file with the following words, into your battlefield steam folder:

Thread.ProcessorCount 8
Thread.MaxProcessorCount 8
Thread.MinFreeProcessorCount 0
Thread.JobThreadPriority 0
GstRender.Thread.MaxProcessorCount 16

You adjust to 6/6/12 if you have a 6 core processor, or 4/4/8 for a quad core.

And then the youtuber loads into a completely different map from the one he was on when he shows the 'high cpu usage', and like magic, in this other map, he's getting lower CPU usage! Miracle!

Be advised that this doesn't work.

It does nothing. The game already knows how to use your CPU cores. Telling it you have 6 cores and 12 threads doesn't do anything if you only have 6 cores and 12 threads. If you have 8 cores and 16 threads it knows how to use those.

It's a placebo effect. It's not harmful if you have a straight shooter 1c/2t x whatever CPU, like most AMD CPU's and any Intel CPU from 11th gen and earlier.

What it will do however, is block any CPU cores/threads that are on the CPU that are after the number you list. This is a problem if for whatever reason, you tell it to only use 6c/12t on a 8c/16t CPU.

More significantly, using 8/8/16 like the redditors & youtubers tell you to use, because you "only have 8 P-Cores", is a massive problem for 12th Gen, 13th Gen, 14th Gen and Core Ultra Intel CPUs.

BF6 knows how to use E-Cores. It works them perfectly well. I have a 14900k & 5080. The settings aren't that important except to know I keep them the same. When you tell the game to use "8 processors and 16 threads" (or 6/12 for a 14600k or similar) you're telling it to not use any of the E-Cores. Sure, you might get "reduced CPU usage", but it's doing it by not using the processor, and killing your FPS. The P-Cores don't increase their workload (or don't increase enough) to make up for the lack of E-Cores.

I did a benchmark, using a special benchmark map in Portal (it's called "Bench De Mark" if you want to find it, then stand on the tyre on the right of the map as you spawn in, and let the bots shoot and throw nades and blow up cars, it's not a perfect situation but it's as good as we get now). 2 minutes captured with CapFrameX for statistics.

Without user.cfg I get:

• Average FPS - 150.8
• 1% Low - 109.3
• 0.1% Low - 96.9

Then using the 8/8/16 user.cfg:

• Average FPS - 146.4 (4.4 worse)
• 1% Low - 87.7 (21.6 worse)
• 0.1% Low - 77.6 (19.3 worse)

As you can see, this is massively inferior. During the user.cfg 8/8/16 run it completely stopped the e-cores getting any load on them from the game. If you weren't looking closely into it (eg, an Afterburner overlay only) you might think it's good, because you have similar FPS, and significantly reduced CPU usage. Except you don't realise that it kills your 1% and 0.1% lows, which will cause dip and stuttering.

And I also did a test where I set the user.cfg to match the 14900k fully. So 24/24/32.

• Average FPS - 153.5
• 1% Low - 104
• 0.1% Low - 84.8

It's very similar to the "without user.cfg" run. Some small differences that I think are just the nature of using this test map with lots of bots and grenades etc (in the 0.1% low mostly, so maybe a bunch of nades went off at the same time). If I did more and more runs they'd probably average out to be identical.

The thread/processor user.cfg trick is a waste of time, or actively hurts your performance. Don't use it.