• 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.

**LETS SEE YOUR PILEDRIVER OVERCLOCKS - LET ME START WITH 5GHz+!!**

The FX-83## beat AMD's old flagship CPU by a long way in rendering.

But i would like to see you guys try and beat my CPU + GPU scores, if you can :D

PS: I'm Prime95 stable with a little left in the tank. :p

Watch me get hammered again

LuxCPU.png


LuxCPU-GPU.png


CINE11.png
 
Last edited:
Cooling is AIR i use titan fenrir's with my AMD chips.
Custom brackets to mount heatsink East/West and increase clamping pressure.

Having fun with this old m4a89gtd now.
Running fine at 4.6, will up the clocks and see how she does.

Was doing OK at 4.8 till she shut herself down.
Was priming OK.
 
Last edited:
wolvers said:
Shutting down while stress testing is usually RAM and/or mem controller in my experience.
Click to expand...

Prime95 puts a lot of stress on the Integrated Memory Controller.

I doubt any Software will ever put that much stress on the Memory, so perhaps using that to test for 24/7 clocks is perhaps not ideal.

Its a bit like running Furmark through the GPU for 4 hours, not very clever given that the GPU will never experience that level of stress out side of Furmark and its not healthy, your GPU is not going to thank you for it.

If i'm right AOD has built in stress testing specifically designed for the AMD architecture.
 
Last edited:
wolvers said:
If I run my RAM too fast, it'll shut down while running OCCT. Most likely that's the controller crashing because it'll do it at speeds that I'm pretty sure the RAM can do.
Click to expand...

Rated for 1866Mhz, will be fine and safe at 2133Mhz (1.5v or less) anything much over that and especially at 1.65v it will be under a lot of stress.

IMO Memory Shouldn't really be run at speeds that are that high on any CPU for extended periods or while stress testing, its just for benching.
 
The m4a89gtd works with the FX 8320 but not well.

The VRM's aren't really up to sustined load at high clocks.
For normal use probably fine, but during stress testing they let go after a few minutes.

You have to use low muti's and high FSB.
If you leave the CPU at 200 and use say a 23 multi for 4.6 the multi will bounce around, going down to 14.5 when you put any load on the CPU.

If you keep the multi low and raise the FSB the clock will remain at static at load.
300x15 for 4.5 for example.

At a pinch you could use this board with an FX chip but you would have to live with the limited bios support, the need to use high FSB and low multi and the inability to use it for any extended high load.

The problem with the clock speed jumping around with high multi's could be fixed with a bios update.
I doubt though that ASUS will be forthcoming.

The issue with the VRM's not being able to deal with the huge current draw of the FX chips under high load isn't something that can be fixed.
Saying that, not many if any 8x boards can cope with that sort of load anyway.

The CH 4 suffers the same bios limitations and strange behaviour regarding bouncing multi's as well.
I briefly tested the 8320 in my CH4.

With the voltages required for P95 testing at >4.6 Ghz it also shut itself down.
It just lasted longer before doing so.
 
Last edited:
humbug said:
Prime95 puts a lot of stress on the Integrated Memory Controller.

I doubt any Software will ever put that much stress on the Memory, so perhaps using that to test for 24/7 clocks is perhaps not ideal.

Its a bit like running Furmark through the GPU for 4 hours, not very clever given that the GPU will never experience that level of stress out side of Furmark and its not healthy, your GPU is not going to thank you for it.

If i'm right AOD has built in stress testing specifically designed for the AMD architecture.
Click to expand...

Prime95 has a stress test because those who author the software to discover new prime numbers through heavy computation (the type of thing computers were designed for) wanted people to ensure their computers were stable, otherwise their computers would likely perform a miscalculation and get a false positive which would be a huge waste of time for everyone.

Also GPU's should be able to run things like Furmark even though it's considered an unrealistic load, the problem is GPU's have become so immensely power hungry that the choice was either to limit usage (which they have now done) or halt GPU progress until power draw was able to be brought under control, most high end GPU's today have under designed cooling/VRM's (just good enough to do normal stuff) which is why they would not be stable running Furmark.

If a GPU was built with adequate cooling & VRM's there is no reason they couldn't run Furmark, any failure would be down to inherently faulty components which would have died sooner or later under normal usage anyway.
 
Last edited:
humbug said:
Prime95 puts a lot of stress on the Integrated Memory Controller.

I doubt any Software will ever put that much stress on the Memory, so perhaps using that to test for 24/7 clocks is perhaps not ideal.

Its a bit like running Furmark through the GPU for 4 hours, not very clever given that the GPU will never experience that level of stress out side of Furmark and its not healthy, your GPU is not going to thank you for it.

If i'm right AOD has built in stress testing specifically designed for the AMD architecture.
Click to expand...

It's also pathetic.
I remember it'd stress on that for ages, but doing anything CPU intensive would just crash, ergo unstable.
 
AOD is useless as a stability tester.
It is in the same league of uselessness as the AIDA stress test.

You can pass hours of AOD only to fail P95 or IBT within seconds.
 
Martini1991 said:
That's because they're not designed for AMD architecture.


;)
Click to expand...

What nonsense.

Linpack has been used for years to measure server performance.
AMD were quiet happy with this as a benchmark until very recently.
Basically the embarassing performance of BD made them pull out.

P95 is certainly not Intel specific.
An awful lot of contributers to the gimps project use AMD CPU's
George would have to be a tool not to optimise the code as much as possible for the AMD architure.
 
Well, looks like PCZ wins. :(

My 8350 isn't a great clocker, like the one bro306 has, and hence I doubt it'll do 5Ghz 24/7. 4.8Ghz yes, 4.9Ghz maybe (when I get my new rad) but not 5Ghz. The vcore it would need is too high. In actual fact, it's only marginally better than my 8320! :o

Damn it, I never get lucky in the silicon lottery!! :D
 
You must log in or register to reply here.
Back
Top Bottom