Better scores in Cinebench: Disable Hyperthreading

[Burninator]

Cinebench is really highly tied to the CPU, it seems. I had been noticing that when it ran, it would but my CPU up to 100%, that is, 100% of 1 of the CPU threads.

So I thought, why can't I disable Hyperthreading, and only have 2 threads for my 2 cores? It ought to give me a better score! So I disabled Hyperthreading in the BIOS, and I was able to increase my OpenGL score from about 20-21 with Hyperthreading to 29.68 without it. (Note that my CPU is clocked at 3.41 GHz, which will affect the score.)

My CPU is still pegging at 100% of one of the cores during rendering, but at least it's more representational of my graphic performance.

A question then is this: Is the CPU pegging at 100% because of the Cinebench renderer? Or is it a Mac OS X deficiency in graphic drivers causing the CPU to be overused?

Another question is this: How much perfomance gain/loss do I get from not having vs having hyperthreading enabled? I did note a reduction in Cinebench CPU score from around 2.8 to around 2.1 points.

*wanders off to ponder whether to leave HT off or turn it back on...*

 

[karacho]

Cinebench is really highly tied to the CPU, it seems. I had been noticing that when it ran, it would but my CPU up to 100%, that is, 100% of 1 of the CPU threads.

So I thought, why can't I disable Hyperthreading, and only have 2 threads for my 2 cores? It ought to give me a better score! So I disabled Hyperthreading in the BIOS, and I was able to increase my OpenGL score from about 20-21 with Hyperthreading to 29.68 without it. (Note that my CPU is clocked at 3.41 GHz, which will affect the score.)

My CPU is still pegging at 100% of one of the cores during rendering, but at least it's more representational of my graphic performance.

A question then is this: Is the CPU pegging at 100% because of the Cinebench renderer? Or is it a Mac OS X deficiency in graphic drivers causing the CPU to be overused?

Another question is this: How much perfomance gain/loss do I get from not having vs having hyperthreading enabled? I did note a reduction in Cinebench CPU score from around 2.8 to around 2.1 points.

*wanders off to ponder whether to leave HT off or turn it back on...*

I can't agree to this. I'd tested this and disabled HT in my Bios too and ran cinebench.

• OpenGL w HT -> 23,46
  OpenGL w/o HT -> 21,70
  CPU w HT -> 6,24
  CPU w/o HT -> 3,50

I'll turn it back to enabled

 

[Burninator]

I can't agree to this. I'd tested this and disabled HT in my Bios too and ran cinebench.

• OpenGL w HT -> 23,46
  OpenGL w/o HT -> 21,70
  CPU w HT -> 6,24
  CPU w/o HT -> 3,50

I'll turn it back to enabled

Your CPU must act differently without HT than mine... =/

 

[karacho]

I can't agree to this. I'd tested this and disabled HT in my Bios too and ran cinebench.

• OpenGL w HT -> 23,46
  OpenGL w/o HT -> 21,70
  CPU w HT -> 6,24
  CPU w/o HT -> 3,50

I'll turn it back to enabled

Your CPU must act differently without HT than mine... =/

Probably yes. Although all four cores were running at 100%