780i vs 790i

[gbrantner]

Last weekend, built a frisnd's system at my house, so we had these two systems, side by side. The first (mine):

780i, Q6600@3.2GHz (1600x8) synched & linked with 4GB PC800, 2 x 512MB 8800GTS @ 675/1940 - Vista 32-bit

The second (his):
790i, E8500@4.1GHz (1726x9.5) synched & linked (1/1) with 2GB DDR3 PC1800, 2 x 512MB 8800GTS @650/1940 - Vista 32-bit

In 3DMark'06 testing, the CPU scores for the E8500 trailed the Q6600 by well over 1000 points. However, at the speeds reported above the 8800 GTS shader model 2 scores were approx:

Q6600/PC800 system: 5600
E8500/PC1800 system: 8500

The shader model 3 scores were approx:

Q6600/PC800 system: 5800
E8500/PC1800 system: 9300

The final 3DMark'06 scores were approx:

Q6600/PC800 system: 15000
E8500/PC1800 system: 18700

Yes, the price of DDR3 is simply insulting...rediculous. But if you can afford it there is much more to this story. Obviously the additional bandwidth is assisting the video cards far more than I would ever have imagined.

Had I not seen it, I would have denied it. Now...I just want a 790i board and 4GB of DDR3. Unfortunately, the wife's school costs come first, so I'll just dream about it.

 

[heyheybooboo]

Switch microprocessors and run your tests again this weekend

 

[sauria]

That's a lot == can you do the cpu swap and report?

 

[myocardia]

Swapping the CPU's would be useless. Those differences are caused mostly, if not wholly, by the video cards. SM2.0 and SM3.0 are video card-related scores, not CPU-related scores.

 

[VirtualLarry]

Sounds like DDR3 does offer performance advantages. Now if the price would only come down to something reasonable, then it would be time to switch.

 

[ThaJollyMan]

Is there a point in running 4gb with 32-bit OS?

 

[Sheninat0r]

Yes there is; to put it simply, 3.2GB > 2GB.

 

[gbrantner]

Originally posted by: myocardia
Swapping the CPU's would be useless. Those differences are caused mostly, if not wholly, by the video cards. SM2.0 and SM3.0 are video card-related scores, not CPU-related scores.

Of course they are, but they can be affected by other, related systems. CPU, system memory, etc, are all involved in graphics preprocessing. Since the video cards can perform their part of the rendering proces much faster than the CPU+ part of the job, a stout increase in the speed that preprocessing data reaches the video cards should produce an increase in overall graphics performance.

In short, every system affects every other system to some degree. I was also reminded recently that since the 790i board has true PCIe 2.0 and the 780i does not, this could account for some of the difference and would likely directly affect the video card scores only. If anyone can shed some knowledgable light on that subject, I'm all ears.

Thanks to everyone for your input. If my bud comes back to visit with PC in tow (he lives 60 miles away) I'll look into the CPU swap 'n' test. That should prove interesting indeed! Might swap video cards as well to establish things from that perspective.