Thanks for your researchSent from my iPhone using TapatalkJB

I did not record the numbers, for I don't want to expend more time and energy to verify what has already been known and understood for years. Hyperthreading does not provide any advantage in FSX whatsoever, and indeed the very opposite is the case. In my opinion something must be amiss in your equipment, the monitoring or your analysis if your data shows otherwise. Mistaken findings can lead people into making wrong hardware decisions, something we all want to bear in mind. I would not want to be responsible for someone buying a hyperthreaded CPU just because they thought it would be beneficial in FSX due to having read this thread.

Let’s agree to disagree on this matter Stephen. ;-)With all respect, FSX can make use of Hyperthreading.Let me show you a picture of 3 screenshots I’ve just put together. Here you can clearly see that when flying fast over high rez photo scenery using a high LOD we clearly get better texture loading by using hyperthreading even thou we also use a lower CPU overclock. You would also notice the faster load time to start the flight.I’m not saying that you should have a CPU with hyperthreading to fly FSX. A regular quad core is normally sufficient. But Hyperthreading can be really useful when it comes to flying fast and low over photo scenery.Merry Christmas

I'm thinking the 17+FPS between AM=13 vs. AM=15 may give people the wrong idea. FPS shouldn't change between the two, so the explanation is fluctuating FPS. Unless you have achieved something with the different mask or changed a setting?

Let’s agree to disagree on this matter Stephen. ;-)With all respect, FSX can make use of Hyperthreading.Let me show you a picture of 3 screenshots I’ve just put together. Here you can clearly see that when flying fast over high rez photo scenery using a high LOD we clearly get better texture loading by using hyperthreading even thou we also use a lower CPU overclock. You would also notice the faster load time to start the flight.I’m not saying that you should have a CPU with hyperthreading to fly FSX. A regular quad core is normally sufficient. But Hyperthreading can be really useful when it comes to flying fast and low over photo scenery.Merry Christmas

Thanks again for sharing your tests Lars, and merry Xmas to you too :)Have you tried disabling cores to test HT? What ultimately led me to believe that HT is not helping in any way is disabling 2 & 3 cores with HT on & off.With only 1 or 2 cores and HT off I get an almost constant 100% CPU utilisation. But enabling HT I never got 100%, even though obviously it looks like FSX could use more execution units. I'll try that again and post some screenshots, maybe check if there's some relevant I/O stats in Process Explorer, or ProcMon

I have tried with disabling cores. Mostly in DX10 where the fibre load is a lot higher then in DX9. Let me put it this way. The main FSX thread, the one that ultimately decides what FPS you get can not benefit from HT. But the texture loaders and the fibres can.So if you have a single core with HTon you will not benefit at all from HT. With a dual core you can benefit a lot from HT if you use DX10 (where the fibres take up 50% thread load) as you can have the Main thread on core#0, nothing on core#1, fibres on core#2 and a t&t loader on core #3. You should be able to benefit from better texture loading in DX9 by putting 2t&t loaders on the last two threads as the few fibres that ends up on the same physical core as the main thread should not limit performance much. I have not looked in to it properly thou.Remember that activating HT does not give you double CPU performance even thou the task manager shows twice the amount of cores.So a dual core with HToff running both threads/cores at 100% does the same work as the same CPU with HTon running the same 2 threads, one on each physical core, at 100% while the 2 remaining threads do nothing. In the first case performance monitor shows 100% load but in the second case performance monitor shows only 50% load. But the same amount of work is carried out in both cases.In FSX with HTon you do not want anything to run on the thread that shares the same physical core as the main thread. That way the main thread does not have to share its physical core with anythingthing else. That empty thread will bring down the total CPU utilization in performance monitor but makes sure the actual FSX performance remains good.

Well, I've been trying, but I can't reproduce your results. I'm 100% CPU limited in both FSMark07 & FSMark11 during the entire test. Both with AG set to Extremely Dense

Would you mind re-running the FSMark07 tests for a full 630seconds instead of the standard 300s? I suspect the PCIe limit will show up during those extra 5,5 minutes.Could you also confirm if you have BP=0, RejectThreshold or none of them in the .cfg?

I'm thinking the 17+FPS between AM=13 vs. AM=15 may give people the wrong idea. FPS shouldn't change between the two, so the explanation is fluctuating FPS. Unless you have achieved something with the different mask or changed a setting?

You shouldn't look too much at the FPS counter in the corner. It's useless for FPS comparisons and that's not what I wanted to show ;-) Look at the ground textures and the speed of the aircraft =)

Edited January 19, 201214 yr by SAAB340

Would you mind re-running the FSMark07 tests for a full 630seconds instead of the standard 300s? I suspect the PCIe limit will show up during those extra 5,5 minutes.Could you also confirm if you have BP=0, RejectThreshold or none of them in the .cfg?

Sure, I'll try that later today. I do have the BP=0 tweak in place, and I allways test in DX9. Should I change something there for the tests?

Sure, I'll try that later today. I do have the BP=0 tweak in place, and I allways test in DX9. Should I change something there for the tests?

Did you have any luck? Try without BP=0 for best result.http://forum.avsim.net/topic/360847-bp0-and-the-pcie-bus-or-why-you-should-have-a-fast-enough-gpu-to-actually-make-use-of-your-cpu/I wish I could edit my first post. In DX9 the fibre load is 30% core load. Not 15-20% as I wrote earlier not realizing I was being limited by the PCIe bus all the time. It's still less than the 50% you get with DX10

Did you have any luck? Try without BP=0 for best result.http://forum.avsim.n...se-of-your-cpu/I wish I could edit my first post. In DX9 the fibre load is 30% core load. Not 15-20% as I wrote earlier not realizing I was being limited by the PCIe bus all the time. It's still less than the 50% you get with DX10

Ok Lars, I'll try without BP=0Thanks for your time and effort, all your posts are great contributions

Did you have any luck? Try without BP=0 for best result.http://forum.avsim.n...se-of-your-cpu/I wish I could edit my first post. In DX9 the fibre load is 30% core load. Not 15-20% as I wrote earlier not realizing I was being limited by the PCIe bus all the time. It's still less than the 50% you get with DX10

This is with Global High settings, I'll try with AG maxed out now

I'll be following this topic with great interest. For now, I agree with Stephen, but am, as always, open-minded.

This is with Global High settings, I'll try with AG maxed out now

Thanks Dario. There is clearly places where you don't have perfect scaling with CPU speed. It's always reasuring when others can reproduce your own findings.I know my settings were not standard at all. Think I just started testing with what I had set at the moment, and once you have a big set of data it's a bit silly to change... If you want to test with the same settings it is what is stipulated for FSXmark11 with these 3 key changes: Resolution 1920x1200x32 (instead of 1680x1050x32) Water set to High 2.x (instead of Mid 2.x) and Traffic set to the "High" setting. (there is actually 2 more but I doubt they will have much impact. Its info text set to continious instead of Single line and Rate of change in the weather set to Medium instead of no change).

I'll be following this topic with great interest. For now, I agree with Stephen, but am, as always, open-minded.

Totally agree. Indeed very interesting.... ^_^

Thanks Dario. There is clearly places where you don't have perfect scaling with CPU speed. It's always reasuring when others can reproduce your own findings.I know my settings were not standard at all. Think I just started testing with what I had set at the moment, and once you have a big set of data it's a bit silly to change... If you want to test with the same settings it is what is stipulated for FSXmark11 with these 3 key changes: Resolution 1920x1200x32 (instead of 1680x1050x32) Water set to High 2.x (instead of Mid 2.x) and Traffic set to the "High" setting. (there is actually 2 more but I doubt they will have much impact. Its info text set to continious instead of Single line and Rate of change in the weather set to Medium instead of no change).

WILCO tonite :) but my max resolution is 1920x1080