New PC

[Neilg]

you've got 12 cores and only 12 threads, is hyperthreading turned off? that would explain the difference between yours and the top result.
You shouldn't be surprised about the quadro.

[glorybound]

I didn't do anything, so where do I look? Is it a sytem thing or a MAX thing?

you've got 12 cores and only 12 threads, is hyperthreading turned off? that would explain the difference between yours and the top result.
You shouldn't be surprised about the quadro.

[glorybound]

or, is it a BIOS thing?

[danio]

Hyper threading is determined by a Bios setting. Take a look in there...should be a simple yes/no toggle.

Another way to check if it's on is too look at the resource monitor tab in windows' task manager. How many CPU graphs are shown?

[glorybound]

there are 12... should I be seeing 24?

[Neilg]

yeah. that should give you an extra boost of 30% or so.

[glorybound]

brilliant!

[danio]

Funny HT would come disabled by default.

[glorybound]

I thought so.... I should have known, so I can image most wouldn't have had a clue to enable it.

[alexyork]

That is a very interesting benchmark there. Look how similar the mildly overclocked single-CPU 3930k system is, and would cost less or the same to build. I'm running mine at 4ghz, with 32GB RAM. The new 5960X will be 8-core and should reach 4ghz as well, with decent cooling. So you'll get most likely a better score for less or the same money with that, and with a much simpler consumer system the individual components are far easier and cheaper to replace if needed.

I suspect this is because your xeons are previous generation, not current. As someone else posted earlier, the actual architecture makes a difference when comparing these chips. It's not just the clock speed or number of cores.

It looks like you got a decent system for the price but I still wonder about the drawbacks and whether or not it's worth it for such a mild improvement over a single-processor consumer system. And I actually wonder how a properly (4ghz) overclocked 3930k or 4930k would compare, at the same price-point... I think I will run a benchmark when I'm back in the office.

[joconnell]

I'm the same - if I can get the same overall grunt but in a single cpu machine I'd rather that for all the programs that have multithreaded limitations.

[Vizioen]

I'm the same - if I can get the same overall grunt but in a single cpu machine I'd rather that for all the programs that have multithreaded limitations.

Oh god, I just noticed by running the same cinebench test that IT at my job really doesn't do a good job when it comes down to setting up their computers. I've been working more than a year on a computer that also has hyperthreading disabled...1.44x slower, that could've saved me A LOT of time.

[alexyork]

Oh god, I just noticed by running the same cinebench test that IT at my job really doesn't do a good job when it comes down to setting up their computers. I've been working more than a year on a computer that also has hyperthreading disabled...1.44x slower, that could've saved me A LOT of time.

ouch.........

[Vizioen]

ouch.........

I asked them, turns out they do it on purpose because of this:
-Hyperthreading is only interesting when a few cores are used
-when they are all in use, the computer works faster when it's disabled
-for daily use (not rendering) it's better this is disabled

[alexyork]

I asked them, turns out they do it on purpose because of this:
-Hyperthreading is only interesting when a few cores are used
-when they are all in use, the computer works faster when it's disabled
-for daily use (not rendering) it's better this is disabled

hah... oh dear. just run a simple benchmark render comparing with and without HT. will solve the issue within 5 mins... one would hope, anyway!