There's no such thing as No AM. hehe

 

Remember that no AM on a four core HT Off is 15=1111, four straight cores is what P3D wants mostly, it's going to work great. It's not the best you can get but it's close. So an equivalent for HT On would be perhaps 85=01,01,01,01. You should be able to go from HT Off to HT On and see more throughput capacity throughout the system. If you don't get that then you got some weird problems, maybe system or maybe some kind of software running doesn't like HT and places double the threads, could be anything. The Ultimate fps is more dependent on the CPU frequency so you'll not be able to change that much. What you get is more consistent fps and better handling on big systems with lots of SimConnect traffic. So some don't notice much difference because they are not using it all.

Sorry, I thought the OP meant zero AM effect :AM=14.

 

I have spent few more hours testing.

My conclusions didn't change.

I am perfectly aware that I am loosing the HT benefit on my whole PC but as I said I use my computer for P3D and daily work only (so no need for HT) and above all I am not convinced at all that P3D takes any benefit of double virtual threads (HT).

It seems to me like FSX was written for 2 cores and P3D optimized for 4 but without the expected benefit of HT, which in my case has even a negative effect.

That's right, only the resources called upon by the sim have improved performance with HT enabled.

 

There's no getting away from the fact the PC runs better overall with HT enabled, and that it has no adverse effect on P3D.

 

Any loss of performance to the sim when HT is enabled is not down to the sim, that works the same, assuming the equivalent AM is provided. The problems are of another sort.

 

It's a shame to turn off HT because of hyperbole.

 

I am not convinced at all that P3D takes any benefit of double virtual threads (HT).

 

P3D makes one fixed thread per unmasked LP, each has a specific job. These are placed in the pattern of ones in the mask. The affinity is passed on to the system (jobscheduler) and that follows the same pattern. Threads spawned by these jobs count >50 wih P3D, and these threads all exist together on those LPs, placed by the jobscheduler.

 

So they cannot intimately gain access to the function of HT.

 

What does make a difference is that two threads one per LP on an HT core finish ahead of the same two threads on the same core without HT.

 

Ultimately there's going to be hundreds of threads of the system resources invoked by that activity and these will run with the HT function improvements.