• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

piledriver - Official AMD FX4300,FX6300 and FX8350 review thread

I said i would do some tests in Heaven to see if there was any multi GPU improvement.
As this test is in the main GPU limited there shouldn't really be much difference between different CPU's,
As long as the CPU has enough grunt to keep GPU's fed the scores should be very similar.

I used a pair of 680's with mild overclocks. core+120 mem+480

The FX 8150 and FX 8320 ran at the same clocks 4.6 with the same memory speed and in the same MB. Sabretooth 990FX
I just swapped the CPU's around.

For comparison i used an Intel 3820 also running at 4.6

FX8150
fx8150-heaven.png



FX8320
fx8320-heaven.png



i3820
i73820-heaven.png


The 8320 shows improvement over the 8150 but is still holding the GPU's back some what.
 
I said i would do some tests in Heaven to see if there was any multi GPU improvement.
As this test is in the main GPU limited there shouldn't really be much difference between different CPU's,
As long as the CPU has enough grunt to keep GPU's fed the scores should be very similar.

I used a pair of 680's with mild overclocks. core+120 mem+480

The FX 8150 and FX 8320 ran at the same clocks 4.6 with the same memory speed and in the same MB. Sabretooth 990FX
I just swapped the CPU's around.

For comparison i used an Intel 3820 also running at 4.6

FX8150
fx8150-heaven.png



FX8320
fx8320-heaven.png



i3820
i73820-heaven.png


The 8320 shows improvement over the 8150 but is still holding the GPU's back some what.

Don't suppose U've got an i5 you can do the test on?
 
Well that's me confused, I'm sure a Phenom II X6 gets much higher results.
And the AVX extension was meant to double results, I get about 126 at 4.8 on a 2500k, and I was getting 84 with my clocked Thuban without AVX.

Never expected that sort of result, my bad :p
My first facepalm moment of the month.

a 2600k @ 5.2 scores around 66 / 70

Shouldn't do, with AVX I get 126ish on my 2500k.
 
Last edited:
Wouldn't quite say it should be compared with an i5 3330.
Something like an i5 3570k/2500k (2500k is like a 1 year and 8 months older than the FX8320 too)
But I see where you're coming price wise.
At the price of the FX8320, you'd have trouble justifying getting a locked i5.
 
It looks like my core temps are reading OK under load, but not at idle as everyone says. They've not gone over 58c yet, but I have seen 66c on the CPU temp which I think is the socket?

Currently stress testing 4.8ghz. What's your PLL voltage (also called VDDA)? I've got that at just under 2.7v and it's helped me get a higher clock without loads more core.

Edit; Woot! 20 run IBT stable @4.8ghz. :)

http://img24.imageshack.us/img24/2489/48ghz240x2020runibt.png

I don't know if IBT stresses these CPU's that much, 45GFlops is chronically low you'd be best using Prime95 imo.
 
Price is all that matters comparing CPUs.

If you're deciding what to buy right this second then yes.
But there's different factors, the age of the 2500k is definitely a huge factor, given people will have spent money in AM3/AM3+ stuff during that time, all adds up to total cost.
Obviously, I'm not a fan of comparing 130 quid chips to 250 quid chips (The 3820)
Much rather see an i5 result compared to the 8320.
 
Last edited:
The average FPS is only 15% lower for the 8320, despite the fact the 3820 is clocked at 4.4GHz. That's not a bad show for a chip that is half its price.
 
That comparison should be done with an i5-3330. Good effort though, it's a lot of hassle doing all of that.

No not a i5-3330 :(

Thought a 3820 is a valid enough test.
The difference in price between a 3820 and an FX 8320 isn't great once you facter in £900 for GPU's a £150 PSU etc.
Don't think somone buying multi top end GPU's would be counting pennies to that degree.

I have some 3570k results somewhere maybe you would consider that a bit more of a valid compatison,
They were run with an older driver though and the scores if run with the latest available driver 310 would be a little higher.

Found a screenie
Same GPU's as the above were used at the same clocks
3570k
i53570-heaven.png
 
Last edited:
Suppose if PCZ wouldn't mind, he could emulate an i5.
Run in dual channel and remove HT, obviously it won't be the same completely as an i5, but close enough as he can for socket 2011.

I could if i wanted use a real I5.
Still have a 2500K and an ASUS p67 pro around somewhere.

Got a 3770K on a P77 as well that is if anything actually faster at Heaven than the 3820.
Guess thats too pricey for comparison.

Strange isn't it how we come down to using middle of the road Intels to test against AMD's finest
 
Last edited:
Good point, it's possibly not a worthwhile comparison at all. If you've got the cash for two high end GPUs it's likely you'll have the cash for a 3770k.

8320 should be compared with the 3330 along with something like an HD7870 or GTX660. What you've shown is good to know, but to be expected really.
 
the only comparison that matters is how much both are

intel is better all round option unless you have a am3 plus board

all rest is just hear say
 
A CPU by itself does not a PC make.
If your going to build a complete PC a few quid extra on the CPU is only a small percentage of the total cost.

Where it is a major consideration is upgrading the CPU only.
Folks with Phenoms and compatable MB's fall into this camp.

The MB's where these FX chips work best are also not budget items.
Because of the heavy current draw the cheaper boards don't work well at all.
Certainly not well if someone wants a big overclock.

So these budget chips end up in CHV's and Sabertooths etc where the combined cost is close to that of the Intel K chips and a P77 board.
 
My IBT run... i can run it at 4.2Ghz, even 4.3Ghz but for that i need to turn the CPU power line up to 130%, i just don't think it drawing that much power 24/7 is healthy.

ibt.png


This is an AMD CPU running on software designed around Intel CPU's, the numbers in that result are pretty meaningless to me, but it is good for short run stress testing, the temps i get in that are about the same as a 2 hour P-95 run.
 
Last edited:
Back
Top Bottom