The great dual vs quad vs hex gaming benchmarkapalooza

[AdamK47]

With the lack of reviews with potent hardware comparing the three in gaming, I decided to take it upon myself to explore it.

My specs are as follows:

Intel Core i7 980X @ 4200MHz
Asus P6X58D Premium @ 21 x 200MHz
12GB Corsair XMS3 @ 2000 DDR
Three GTX 580 in Tri-SLI @ 850/2200
120GB Vertex 3 SSD
Two 3TB Deskstar 7K3000 in RAID-0 - 6TB
1TB SpinPoint F1
Samsung 22X SH-S223L DVD-RW
Asus Xonar Essence STX sound card
Antec Twelve Hundred case
Thermalright Silver Arrow
Corsair AX1200 power supply

The 980X is a 12MB L3 cache six core CPU built upon a 32nm process. To enable dual and quad core all I had to do was make the setting in the BIOS. No other settings where changed. Note: The enabled cores have access to all 12MB of the shared L3 cache.

Dual vs Quad vs Hex

The nVidia drivers used are version 275.50. SLI is enabled and the global driver settings are at default except for Texture Filtering Quality, which is set to High Quality instead of the default Quality setting.

The resolution is kept at 1920x1080 with everything in each game as maxed in order to have everything at parallel. Each game is updated to the latest version. Some of the results were suprising and some not so much. I won't go into any analysis. I'll leave that up to discussion. Feel free to ask me about any setting.

First up, the synthetics, followed by actual game benchmarks.




3DMark 11

Dual:

Quad:

Hex:

Unigine Heaven 2.5

Dual:

Quad:

Hex:

 

[AdamK47]

Alien vs. Predator (2010) - Stand alone benchmark

Dual:

Quad:

Hex:

Batman AA

Dual:

Quad:

Hex:

Battleforge

Dual:

Quad:

Hex:

 

[AdamK47]

Civilization 5 - Units benchmark run for 200 seconds

Dual:

Quad:

Hex:

Company of Heroes

Dual:

Quad:

Hex:

Call of Juares - Stand alone DX10 benchmark

Dual:

Quad:

Hex:

 

[AdamK47]

STALKER Call of Pripyat - Stand alone DX11 benchmark

Dual:

Quad:

Hex:

Crysis 2

Dual:

Quad:

Hex:

Doom 3 - timedemo demo1 (results from the second run)

Dual:

Quad:

Hex:

 

[AdamK47]

Far Cry 2 - Ranch Small (3 runs)

Dual:

Quad:

Hex:

Just Cause 2 - The Dark Tower

Dual:

Quad:

Hex:

Lost Planet 2 - Benchmark A

Dual:

Quad:

Hex:

 

[AdamK47]

Mafia 2

Dual:

Quad:

Hex:

Metro 2033

Dual:

Quad:

Hex:

Resident Evil 5 - Variable Benchmark

Dual:

Quad:

Hex:

 

[AdamK47]

Shogun 2 - DirectX 11 1080P benchmark

Dual:

Quad:

Hex:

Grand Theft Auto 4

Dual:

Quad:

Hex:

 

[ElFenix]

awesomesauce


looks like 6 cores isn't much of a step yet.


i wonder how this would change with HT off?

 

[AdamK47]

I'll leave HT on/off for someone else to explore. Doing this comparison took me a bit of time. It was fun to do though.

 

[ElFenix]

I'll leave HT on/off for someone else to explore. Doing this comparison took me a bit of time. It was fun to do though.

if we had reps i'd give you some

 

[magomago]

good work, but could you not have made a few charts? that would be so much easier to visualize and understand the information

 

[Borealis7]

great stuff, adam. looks like quadcore with HT for gaming is the sweetspot.

 

[podspi]

Awesome! Looks like you really don't "need" anything more than a dual-core at the moment. Quad seems to help a bit with the min framerates, though. And hex-core appears to be downright useless... for gaming anyway.

 

[nyker96]

thanks for the data, looks like quads are worth it, but hex might not make as much difference in games yet.

 

[BTRY B 529th FA BN]

Very nice post, Thanks!

 

[d4a2n0k]

Good stuff, thanks.

 

[TakeNoPrisoners]

Awesome! Looks like you really don't "need" anything more than a dual-core at the moment. Quad seems to help a bit with the min framerates, though. And hex-core appears to be downright useless... for gaming anyway.

Do keep in mind that is a dual core with 12mb of L3 and clocked at 4.1GHZ.

 

[jones377]

Great thread. Do you think you could do some in game benchmarks with CoH? Preferably on a 4vs4 map. The canned benchmark for it is kinda useless.... But it still showed a boost in min FPS with increasing cores.

 

[Markfw]

Do keep in mind that is a dual core with 12mb of L3 and clocked at 4.1GHZ.

Also, it has HT enabled, so its already almost a quad core.

And the quad test is 6-8 threads and the hex is like 10-12.

Now if it was done with HT off, I bet we could see much more interesting numbers.

 

[betasub]

Now if it was done with HT off, I bet we could see much more interesting numbers.

This is a fair point, as a lot of gamers either have a CPU without HT, or intentionally switch off HT in BIOS.

 

[Phynaz]

Man, lot's of good info. Thanks!

 

[Zap]

Nice.

BTW isn't it time to update your sig?

 

[formulav8]

Good job! A Tri-Quad core is more than enough. Even a Dual Core with HT was plenty for most of your tests.

So really, a high-end dualcore or lower-mid end quad is the best balance of cost/performance.

 

[Blue_Max]

Best. Thread. Ever.

But I agree with others about Hyperthreading... back in the 1 vs. 2 core debates, those Pentium4's with HT did almost as well as a dual-core in many cases. If this means the games are benefiting from 4 vs. 8 vs. 12 "cores", that should give AMD's 6-core CPU's a benefit. (Yet the 2500k beats them?)

 

[Mr. Pedantic]

Why could you not just have posted a bunch of graphs?