i7 930 @4.0GHz: HyperThreading vs No HyperThreading

isaak · Sep 4, 2010, 03:43 AM

Because I couldn't find many direct comparisons online (please don't bombard me with any links that I might have obviously missed

) here is my small comparison of an i7 930 clocked at 4.0GHz with and without HT.

Tests run were: A render on "high" settings in Revit Architecture 2010, 3dMark Vantage and Cinebench 10.

See System Specs Tab for more info.

Here are the settings used for my OC [with HT/without HT]:

Code:

 Ai Overclock Tuner.....................[Manual]
CPU Ratio Setting......................[21.0]
Intel(r) SpeedStep(tm) Tech............[Disabled]
Xtreme Phase Full Power Mode...........[Enabled]
BCLK Frequency.........................[191]
PCIE Frequency.........................[100]
DRAM Frequency.........................[DDR3-1531MHz]
UCLK Frequency.........................[3063MHz]
QPI Link Data Rate.....................[Auto]

CPU Voltage Control....................[Manual]
CPU Voltage............................[1.35000/1.3000]
CPU PLL Voltage........................[1.80]
QPI/DRAM Core Voltage..................[1.33125/1.29375]
IOH Voltage............................[1.14]
IOH PCIE Voltage.......................[1.50]
ICH Voltage............................[1.20]
ICH PCIE Voltage.......................[1.50]
DRAM Bus Voltage.......................[1.64]
DRAM DATA REF Voltage on CHA...........[Auto]
DRAM CTRL REF Voltage on CHA...........[Auto]
DRAM DATA REF Voltage on CHB...........[Auto]
DRAM CTRL REF Voltage on CHB...........[Auto]
DRAM DATA REF Voltage on CHC...........[Auto]
DRAM CTRL REF Voltage on CHC...........[Auto]

Load-Line Calibration..................[Enabled]
CPU Differential Amplitude.............[800mV]
CPU Clock Skew.........................[Delay 300ps]
CPU Spread Spectrum....................[Disabled]
IOH Clock Skew.........................[Auto]
PCIE Spread Spectrum...................[Disabled]

C1E Support............................[Disabled]
Hardware Prefetcher....................[Enabled]
Adjacent Cache Line Prefetch...........[Enabled]
Intel(r) Virtualization Tech...........[Disabled]
CPU TM Function........................[Enabled]
Execute Disable Bit....................[Enabled]
Intel(r) HT Technology.................[Enabled]
Active Processor Cores.................[All]
A20M...................................[Disabled]
Intel(r) SpeedStep(tm) Tech............[Disabled]
Intel(r) C-STATE Tech..................[Disabled]

And here are the results:

Render + Cinebench + Min and Max Temps if you can read them

Vantage + Temps

NB - results on the left side of the pictures are 4.0GHz WITH HT, on the right is 4.0GHz WITHOUT HT

Hope this can help someone.

Takaharu · Sep 4, 2010, 11:55 AM

Can you separate the 4 pictures? I can barely read them at full size; they just adjust to the width of the screen.
As far as I can tell, HT gives it a little extra performance with a temperature that's only 1 degree higher.

isaak · Sep 4, 2010, 12:49 PM

Does ctrl + scroll work? Usually I would fix it right away, but I've gone over my internet cap and my speed has been slowed to a standstill...

With HT the temperature went up around 5-10*C. What I don't understand is the render times and Cinebench scores. They're both render tests, yet with HT the Revit render was slower despite its higher Cinebench score.

BlueMak · Sep 5, 2010, 12:41 PM

It's very difficult to read them. Why did you compress them? Please upload better screenshots.

Which is faster with or without HT?

isaak · Sep 5, 2010, 12:54 PM

Sorry about the original pics.... I've got some new ones uploading as I type. It may take a while lol

I would've liked to include some game FPS benchmarks but I happened to be banned from them during school term time.

EDIT: scratch that for now, it hadn't finished after an hour. sorry about this

mkk · Sep 5, 2010, 03:15 PM

Interesting enough.

The difference in max temperature alone is a pretty good sign there's more being done in well threaded applications.

IvanV · Sep 5, 2010, 04:07 PM

Those who have difficulty viewing the pictures, just right click on them and choose "Open Image" or your browser's equivalent option.

BlueMak · Sep 5, 2010, 06:47 PM

Quote:

Originally Posted by IvanV

Those who have difficulty viewing the pictures, just right click on them and choose "Open Image" or your browser's equivalent option.

This one works.

Well, it looks like there is a drop of about 30 percent or more, disabling HT. No reason to do so then. I mean, if you are going to, you might as well buy an AMD cpu and save some money.
lol

I kid I kid, you know I like you AMD.

Matth · Sep 5, 2010, 11:54 PM

Interesting, with just HT, the gain from it in a multithread app seems to be about 30%.

Now if you move from threads = cores and no HT to threads = 2x cores and HT, you gain that 30%.

The fly in the ointment, is that at anything short of total saturation, it's unclear if the scheduler will place threads on a core by core basis, or produce contention by putting 2 threads on the same core, hyperthreading - if this is the case, then you lose performance with hyperthreading enabled and only enough threads to fill all the cores.

There are some suggestions that the Windows 7 scheduler is intelligent enough to load 1 thread per core and not the second thread per core unless all cores are loaded, but it's hard to find solid proof or statements.

isaak · Sep 6, 2010, 07:21 AM

Generally when browsing between benchmarks I found that the HT enabled one loaded a fair bit quicker and felt a lot more responsive, not to mention the improved performances in Vantage and Cinebench.

But yeah, as you guys are pointing out, the difference with HT seems to be around ~30% more. I guess thats what I set out to find out

Takaharu · Sep 6, 2010, 08:22 AM

Quote:

Originally Posted by IvanV

Those who have difficulty viewing the pictures, just right click on them and choose "Open Image" or your browser's equivalent option.

Good call - much clearer and about 50x faster to open.

Sep 4, 2010, 11:55 AM	#2
Takaharu I can fart in 7 languages Join Date: Aug 2009 Location: England, UK Posts: 1,847 Rep Power: 110 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Can you separate the 4 pictures? I can barely read them at full size; they just adjust to the width of the screen. As far as I can tell, HT gives it a little extra performance with a temperature that's only 1 degree higher. __________________ I don't get paid to know the answer, therefore I'm far more likely to give you a straight and honest answer. Mods Rig, Box Mods Rig, Folding details

Sep 4, 2010, 12:49 PM	Thread Starter #3
isaak HardwareHeaven Senior Member Join Date: Oct 2008 Location: Australia Posts: 2,286 Rep Power: 112 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Does ctrl + scroll work? Usually I would fix it right away, but I've gone over my internet cap and my speed has been slowed to a standstill... With HT the temperature went up around 5-10*C. What I don't understand is the render times and Cinebench scores. They're both render tests, yet with HT the Revit render was slower despite its higher Cinebench score. __________________ The first step... shall be to lose the way. - Galway Kinnell

Sep 5, 2010, 12:41 PM	#4
BlueMak HH's Asteroids' Dominator Join Date: Nov 2004 Location: Hellas->Macedonia->Thessaloniki Posts: 13,465 Rep Power: 349 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading It's very difficult to read them. Why did you compress them? Please upload better screenshots. Which is faster with or without HT? __________________ The people who are regarded as moral luminaries are those who forego ordinary pleasures themselves and find compensation in interfering with the pleasures of others(Bertrand Russell)"You go into Afghanistan, you got guys who slap women around for five years because they didn't wear a veil,You know, guys like that ain't got no manhood left anyway. So it's a hell of a lot of fun to shoot them." - Lt. Gen. James N. Mattis This is slavery, not to speak one's thought. [Euripides-The Phoenician Women (c.411-409 B.C.)] http://www.macedonia.info/FALLACIESANDFACTS.htm *Sic semper tyrannis.*

Sep 5, 2010, 12:54 PM	Thread Starter #5
isaak HardwareHeaven Senior Member Join Date: Oct 2008 Location: Australia Posts: 2,286 Rep Power: 112 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Sorry about the original pics.... I've got some new ones uploading as I type. It may take a while lol I would've liked to include some game FPS benchmarks but I happened to be banned from them during school term time. EDIT: scratch that for now, it hadn't finished after an hour. sorry about this __________________ The first step... shall be to lose the way. - Galway Kinnell Last edited by isaak; Sep 5, 2010 at 01:35 PM.

Sep 5, 2010, 03:15 PM	#6
mkk Cthulhu/Dagon 2012 Join Date: Oct 2003 Location: Gefle, Sweden Posts: 4,490 Rep Power: 132 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Interesting enough. The difference in max temperature alone is a pretty good sign there's more being done in well threaded applications.

Sep 5, 2010, 04:07 PM	#7
IvanV HH Assassin Guild Member Join Date: Dec 2004 Posts: 5,914 Rep Power: 333 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Those who have difficulty viewing the pictures, just right click on them and choose "Open Image" or your browser's equivalent option. __________________ If anyone has Portal 2 and hasn't played the co-op and wants to do me a favour, let me know (PM me or whatever).

Sep 5, 2010, 11:54 PM	#9
Matth Flash Banner Hater Join Date: Jun 2002 Location: UK Posts: 3,426 Rep Power: 93 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Interesting, with just HT, the gain from it in a multithread app seems to be about 30%. Now if you move from threads = cores and no HT to threads = 2x cores and HT, you gain that 30%. The fly in the ointment, is that at anything short of total saturation, it's unclear if the scheduler will place threads on a core by core basis, or produce contention by putting 2 threads on the same core, hyperthreading - if this is the case, then you lose performance with hyperthreading enabled and only enough threads to fill all the cores. There are some suggestions that the Windows 7 scheduler is intelligent enough to load 1 thread per core and not the second thread per core unless all cores are loaded, but it's hard to find solid proof or statements. __________________ Mary had a little lamb, Her father shot it dead Now Mary takes her lamb to school, Between two crusts of bread

Sep 6, 2010, 07:21 AM	Thread Starter #10
isaak HardwareHeaven Senior Member Join Date: Oct 2008 Location: Australia Posts: 2,286 Rep Power: 112 System Specs	Re: i7 930 @4.0GHz: HyperThreading vs No HyperThreading Generally when browsing between benchmarks I found that the HT enabled one loaded a fair bit quicker and felt a lot more responsive, not to mention the improved performances in Vantage and Cinebench. But yeah, as you guys are pointing out, the difference with HT seems to be around ~30% more. I guess thats what I set out to find out __________________ The first step... shall be to lose the way. - Galway Kinnell