• 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+!!**

Requoted from earlier in the thread re power consumption of PD ref 1090T

nkata said:
@ 4.6Ghz the FX-8350 uses 60 Watts more than a Phenom II x6 at 3.8Ghz

Results from testing at the plug, using prime 95 in each case on all cores, eight for piledriver, six for 1090T. 1.392V and 55C core temp on PD, 1.376V and 51C on 1090T

Exact same system, 290W for 1090T and 350W for 8350.

Idling,browsing, office both were comparable at 130-140 ish.

Full prime running on all cores plus heaven 2.5 bench running on 6950 gave 420W for 1090T and 480W for 8350.

ASUS M5A99X very high LLC virtually no vdroop.
Click to expand...

I have just rerun the tests at 4.675GHz and the system draw with prime on eight workers is 360W, much the same as previous. There is obviously a 6950 in 2D, the chipset and the memory, fans, drives etc. to factor in.

Idle is 130W with only one or two cores having any load and that being < 5%.

I have a delta over idle of 230W for the PD system, 29W/core which sort of ties into the 1090T/PD difference of 30 - 35W/core total at full load.

I would estimate a maximum CPU load of between 30 and 35W per core, the lower figure is more likely as most cores are unused at idle also the system draw will increase under load. Averaging gives 8 x 32.5W ~ 260W at 4.67GHz on eight cores. Cooling requirement should be based on this figure.

This allows for an equivalent processor draw of one full core at idle or 30W my system taking 100W approx on top. (230W delta + 30W idle + 100W system) = 360W

Vcore is 1.39V however there are several voltages within the processor so it is not easy to estimate from that.

More testing would mean disabling modules and noting the differences, the cache would make this not ultra accurate though.
 
Last edited:
mmj_uk said:
Noticed an interesting thread over at Anandtech:
http://forums.anandtech.com/showthread.php?t=2289809

They seem to think AMD are misrepresenting FX8350 by claiming it's 125W, makes sense to me given how much hassle I've found it to keep cool under any sort of stress, even with a DH-14.
Click to expand...

Does this guy think the TDP is the makers representation of power-consumption?

If he does, and it seems he does, someone needs to tell him he is wrong.

also, do Intel chips have the Memory controller built on to the CPU's DIE or is it on the Motherboard?
 
Last edited:
Intel have the IMC on CPU.
AMD were the first to do it, but Intels was better.

Although TDP isn't going to be massive miles different away from power consumption, that's where the logic comes into it, won't be scientific, but it's a ball park figure.
 
Martini1991 said:
Intel have the IMC on board.
AMD were the first to do it, but Intels was better.

Although TDP isn't going to be massive miles different away from power consumption, that's where the logic comes into it, won't be scientific, but it's a ball park figure.
Click to expand...

I didn't ask whose IMC was better :rolleyes: so why you felt the need to post you subjective opinon i will never know, well, actually i do but the less said about that the better.
 
It's not a subjective opinion, it's factual.
Not sure the need to get so defensive at AMD's IMC.

And I'm not sure what you're trying to imply.

Seems you just love the confrontation. Grow up.
 
Last edited:
Can any of you do >100 Gflops for ten runs ?
The AMD 100 Gflop club

Some of those 5Ghz+ OC'd chips should be able to do it.
 
nkata said:
I think the fact that it was intel oriented in the first place led the developers to produce an amd version.

My first effort at 4.9GHz, needs work, memory size set between standard and high. Will take some doing to maintain > 100

Click to expand...

You have a lot of throttling going on there.... :)

You should be getting consistent Gflops. like this...

ibt.png
 
Looks like your socket temps go over 70C, which I think is when throttling occurs.

It also looks like your OC with fsb increased is giving you higher gflops, so I think you can all guess what I'll be doing this weekend! :D
 
The throttling is strange.
It is not the MB reducing the CPU clock.
Clock speed remains constant.

Has to be error correction in the CPU or some internal CPU throttling adding wait states rather than clock speed reduction.

Below is a run i did at 4.5Ghz.
That is 95 Gflops so up the clock and 100 will be a walk in the park.
Well it isn't

Testing continues

IBT-4.5.png
 
You must log in or register to reply here.
Back
Top Bottom