Interesting comparision of performance with 1,2,3,4 cores

[Duke]

Thought this was quite interesting. I didn't realise how much things have moved on and how many applications/games now gain performance with multi cores.

http://www.tomshardware.co.uk/multi-core-cpu,review-31567.html

As far as games go, we see a huge 60% performance jump from going single-core to dual-core, and a further 25% leap from dual- to triple-core. Quad cores offer no benefits in the sampling of games we tested. While more games might change the landscape a little, we think the triple-core Phenom II X3s are looking good as a low-cost gaming option. It's also important to note here that as you start shifting to higher resolutions and adding visual detail, the picture gets a lot murkier as graphics muscle becomes the prevalent determinant of frame rates.

 

[Fenris]

Very nice article, will read the whole thing when I've got a bit more time.

I have to say, coming from a Pentium 4 3.06Ghz in my Dell to my current Phenom II X3 720, I have been blown away by the performance differance in games.

I wasn't expecting the processor to have that much of an impact on gaming.

 

[Jono8]

But that is all at 1024x768 with low details. Once you ramp that up it's going to be all on the graphics card although i do hold the notion that cpu's really help the minimum fps. I went from an x2 4400 to a core [email protected] and it really made a huge difference to minimum frames.

 

[PaulyD]

But that is all at 1024x768 with low details

Reason they would run that low res and graphics is to take the graphics card out of the equation and test the CPU purely

 

[Jono8]

Reason they would run that low res and graphics is to take the graphics card out of the equation and test the CPU purely

well yeh, but if we are talking about playing games in the real world the cpu isnt going to make much difference. I don't see the point in that benchmark to be honest except for to just test the cpu.

 

[PaulyD]

But they just showed that the CPU does make a difference and the test was purely to test the CPU and not the graphics card

obviously as the graphics and resolution goes up it will be bound to graphics power, but CPU will also be a factor

 

[mmj_uk]

In CPU limited conditions i7 has been show to be about 30% faster clock for clock than anything previous, so even if all the cores do not get fully used (yet) it's certainly still advantageous under dual or tri-core conditions

 

[Audigex]

I think the idea was basically just to show that most processes, especially games and even those which are multi-threaded, have already finished with process #1 by the time process #3 has been started, so there's not as much use for 4.

Think of it as a 4 lane motorway, unless it's REALLY busy, only the first 3 lanes get used.

Okay, so the game performance doesn't actually relate to real world too much (although in an integrated graphics situation it could), the point is just to show the speed differences with more cores. They aren't actually expecting most people to use it in a substantive way, it's academic for those interested in the ability of different programs to scale to the architechture.

To me, it re-inforces the idea which most already have that quad isn't currently much use other than while video encoding, but that as newer programs are developed or modified to use multi-core there's a clear difference.

Shame they didn't use a newer version of photoshop though - I wasn't expecting CS3 to be properly multi-threaded anyway.

 

[ecat]

Interesting article. I remember when the world + dog laughed at the idea of tri-core but for current day gaming it looks a compelling choice.