Can disabling HyperThreading increase single core performance?

[Recon-UK]

My XEON is 20c colder running with HT off, now really i only need HT for my video work and some editing.

Not sure if it is related or not but i gained around 4FPS average to the GTA V benchmark.

This bench is with HT on below - see spoiler.

4FPS is not noticeable because my system performs well above 60FPS.

Spoiler

Has anyone got any further testing on this?

 

[Mr.Scott]

Just disabling HT will not increase single thread performance. What it WILL do is decrease temps so you can OC higher, which in turn will increase performance.

 

[newtekie1]

Yes, disabling HT will give a small boost to single threaded performance. The cores are no longer divided and sharing work, so the one thread going to each core gets the complete attention of that core, so the single thread performance will slightly improve.

 

[Recon-UK]

Yes, disabling HT will give a small boost to single threaded performance. The cores are no longer divided and sharing work, so the one thread going to each core gets the complete attention of that core, so the single thread performance will slightly improve.

This makes a lot of sense, in all honesty i want to keep HT off now due to the gains from it being off, better per core and runs colder, i can easily crank it a lot higher now without HT, i mean i reached 4.46ghz rock steady in benchmarks with HT on but hitting 90c is no laughing joke.

 

[FR@NK]

I was having performance issues with single threaded programs until I disabled HT.

 

[Recon-UK]

Ahh well interesting.

Is The Intel i7 5930K Faster In Gaming With Hyper Threading OFF ?

 

[BiggieShady]

Ahh well interesting.

Is The Intel i7 5930K Faster In Gaming With Hyper Threading OFF ?

I believe that result is connected with SLI setup ... I think there is no such difference with a single card.

 

[FYFI13]

I believe that result is connected with SLI setup ... I think there is no such difference with a single card.

I'd say so. About a yeat ago i tested every single game i own with HT on and off on my i7 4790k + GTX970 and saw from very small to no difference at all. Arma 3 and Butterfield 1 Beta clearly taking advantage from HT and so will upcoming titles.

 

[CAPSLOCKSTUCK]

X5670....stock 2.93 ghz........running at 4.2 ghz

H/T on and off running the cpuz benchmark, my cpu is 5 deg lower under load with h/t off. Single core performance is identical 1474 v's 1475.


on

off

 

[Bill_Bright]

now really i only need HT for my video work and some editing.

Your operating system will take advantage of HT. And the faster the OS can complete its own tasks, the more resources there are available for your other programs.

 

[Kanan]

I took inspiration from here and tested it with the GTA 5 bench:

HTT on:
Frames, Time (ms), Min, Max, Avg
19732, 186281, 49, 271, 105.926

HTT off:
Frames, Time (ms), Min, Max, Avg
19952, 186594, 54, 271, 106.927

Same as the youtube video, min fps is lower with HTT on. Also the loading times seemed lower to me without HTT - and I reduced the voltage to 1.29V now. Let's see how low I can get the CPU running on voltage now.

 

[BiggieShady]

I took inspiration from here and tested it with the GTA 5 bench

Interesting ... so there is measurable difference even with single GPU.
Good choice with GTA5, cpu heavy and uses multi threaded rendering in dx11 ... that gives most noticeable difference between two modes.

 

[P4-630]

I still need to find a way to enable HT on my processor

 

[Recon-UK]

I still need to find a way to enable HT on my processor

LOL if only intel did an AMD like when you could unlock those beautiful Phenom II's

 

[Ferrum Master]

I took inspiration from here and tested it with the GTA 5 bench:

Could you put up both txt files?

LOL if only intel did an AMD like when you could unlock those beautiful Phenom II's

Intel did actually offer to unlock HT for certain models in the past with an upgrade card for 50$.

 

[CAPSLOCKSTUCK]

GTA V benchmark with Xeon X5670 @4.3ghz and Sapphire HD 7970 OC Boost H/T on and off


On

Off

 

[Ferrum Master]

Pretty much within margin of error. GTA5 bench differs on each time, it has variations.

And... try running with the process priority on high... then the results won't differ with or without HT. It is just the Windows Kernel making fun... not the HT.

 

[Recon-UK]

Thanks guys for testing this out, i am very happy that people took this seriously, HT is definitely gaining ground in newer games though that is evident by BF1.

My CPU is trashing 4000 series i5's because it has more threads enabled, the 4000 series are pegged at 100% usage and Battlefield forums is full with people mentioning it, so definitely HT has it's uses.

 

[Kanan]

Could you put up both txt files?

Here you go:

With HTT:
https://drive.google.com/file/d/0B4yY0sTUte68Qkg4RTVkcHVKcnM/view?usp=sharing
https://drive.google.com/file/d/0B4yY0sTUte68R1FFZG8wN0hqSm8/view?usp=sharing
Without HTT:
https://drive.google.com/file/d/0B4yY0sTUte68X1lSMGZIcXNuS1E/view?usp=sharing
https://drive.google.com/file/d/0B4yY0sTUte68UXZCSTBhMDYxTGc/view?usp=sharing