What am I doing wrong? AMD 8120 vs 8320 getting same scores.

[funboy6942]

I replaced my 8120 with a 8320 and Im getting the same scores as I was with my 8120. I oc'd the 8120 to 4.1 and Im ocing the 8320 at 4.4 and my scores on 3dmark 2011 and Heaven are almost the same. Were talking like 10 points different. I thought for sure with a "newer" cpu, and a good overclock of 900mhz over stock, Id see a noticeable difference, but Im not.

What am I doing wrong? Or is it my Mobo, ASUS M5A88 V EVO? I did update it to the latest bios, and its kinda on the old side, as far as computers go at about a year and a half old. But anyway. Im ocing it by bumping up the multiplier, and everything else, like turbo, is set to auto. Though I am ocing my 16 gigs of Kingston Blu 1333 oc'd to 1600mhz.

Anyway, what do you think is giving me the trouble and not allowing me to spank the pants off my old 8120?

TIA

 

[joshhedge]

You failed to mention your GPU and it's potential overclock.

 

[funboy6942]

Its a HD7970 @ 1.1. It will do 1.5 but I just cant justify the 1-2 fps increase.

 

[2is]

You're running a GPU benchmark with the same GPU... Why would you expect a massive improvement by upgrading the CPU???

 

[R0H1T]

You're running a GPU benchmark with the same GPU... Why would you expect a massive improvement by upgrading the CPU???

Yup try PCMark 7 for better results

 

[Gikaseixas]

You're running a GPU benchmark with the same GPU... Why would you expect a massive improvement by upgrading the CPU???

This

OP try benching with Sandra, PCmark and Cinebench

 

[sm625]

What this is telling you is that all potential cpu bottlenecks are covered by the gpu bottleneck. You should try lowering the cpu clock until performance begins to fall off. I bet you could drop 10% off the cpu clock and still maintain 98% of the performance in these benchmarks.

 

[inf64]

@OP

What are the CPU subscores in 3dmark and Heaven benchmarks?Also do a benchmark in some other toy benchmarks like Sisoft Sandra and see if your scores are better than what you got before with 8120. C11.5 is also a solid benchmark and Piledriver does noticeably better there vs Bulldozer.

 

[funboy6942]

Well I already pulled out my 8120 and got it up for sale, plus I went through that "new" hardware bs that windows did forcing me to call in and prove that I didnt just swap out my hdd to another pc or they would deactivate me. So not really wanting to swap out cpus to do that all over again twice.

I never really gave it much thought that I was doing GPU benches while testing out a CPU. Makes good sense. I guess I can rip one of my blu ray discs, again, and see if there is a difference in speed that I can tell when it decodes it.

Ill also try out sandra just to see what it is doing against other computers running a 8120 to see how much faster I am.

Thanks guys for the heads up on this, and I should of been running different tests and not gpu tests Im a dufus.

 

[inf64]

No need to bring back 8120,just post your scores with 8320 @ 4.4Ghz and we will find what Bulldozer scores at 4.1Ghz for you . But I think,like others do, that you hit the GPU bottleneck there. Bulldozer @ 4Ghz is already plenty fast in games, Piledriver @ 4.4Ghz would probably give you a few more fps unless you play Shogun 2 ,SC2 or Skyrim. In those titles CPU is taxed much more and PD would be a lot faster than Bulldozer.

To give you an idea of relative performance of both cores in CPU limited games at fixed clock (4Ghz) here is what hardware.fr found out in their 8350 review(they used OCed GTX680):
http://www.hardware.fr/articles/880-6/bulldozer-vs-piledriver-4-ghz.html

13.5% increase per clock in CPU limited games. In other cases you will see basically what you got in your tests-a GPU bottleneck first.

 

[imported_paulc871]

I do concur