Asus CHV Formula overclocking

[The Stilt]

For air, water (+chiller) and mild phase a chip with average leak & VID seems to be best.
There are quite massive differences between the power consumption (= heat) between the keepers & leakers.
The least leaky chip (happens to have the highest VID too...) I got has over 10% lower power draw than the one which has the highest leakage (lowest VID...).

The chip with lower leakage does better on air with 4CU/8C configuration, while the high leaker beats it hands down with only one core enabled. The "keeper" also runs 6c cooler at the same clocks, even the VDD is 0.1V higher.

[PolRoger]

For air, water (+chiller) and mild phase a chip with average leak & VID seems to be best.
There are quite massive differences between the power consumption (= heat) between the keepers & leakers.
The least leaky chip (happens to have the highest VID too...) I got has over 10% lower power draw than the one which has the highest leakage (lowest VID...).

The chip with lower leakage does better on air with 4CU/8C configuration, while the high leaker beats it hands down with only one core enabled. The "keeper" also runs 6c cooler at the same clocks, even the VDD is 0.1V higher.

Thanks for the comment.

From what you've said then I gather Dumo's 1.200 VID chip might well be too HOT (power hungry) when attempting to run on air at ~4.6GHz(+) 8-cores as compared to my 1.287 VID sample with a similar o.c. and a higher VDD...

From your experience and from the samples that you've tested... what range or ~ number would you describe as the ideal average leak/VID? A chip(s) that would hopefully offer the best balance of clocks to lowish voltages to low/manageable temps?

[Blacklash]

I plan on running 4.6 myself. I can boot and use 5.0GHz fine, however I'm not even going to try to stress test that on air cooling.





http://valid.canardpc.com/show_oc.php?id=2064730

[The Stilt]

Thanks for the comment.

From what you've said then I gather Dumo's 1.200 VID chip might well be too HOT (power hungry) when attempting to run on air at ~4.6GHz(+) 8-cores as compared to my 1.287 VID sample with a similar o.c. and a higher VDD...

From your experience and from the samples that you've tested... what range or ~ number would you describe as the ideal average leak/VID? A chip(s) that would hopefully offer the best balance of clocks to lowish voltages to low/manageable temps?

I got FX-8150 chips with 1.2625V, 1.2875V, 1.3000V, 1.3125V, 1.3250V and 1.3375V VIDs.
The best chips on air have 1.3000V+ VID, the lower ones are simply too hot to be run on air.
The 1.3375V VID chip does ~4740MHz Prime95 LargeFFT stable (4CU/8C) with 1.40V on poor UEX120 + Ultra Kaze 3krpm fan.
After that the third (0-3) compute unit starts failing and does not scale with further voltage.

The chips with higher leakage do only 4550MHz (4CU/8C LargeFFT) or so with 1.35V.
The temperature raises way too high on these and the third CU starts failing at ~70c.

When the power draw / heat is limited by disabling the compute units and leaving only one core active, the lower VID chips do ~5400MHz / 1.49V (light tests) while the low leakers do only 5200MHz or so with the same voltage.

If I would purchase a single FX cpu to be run on air or water, I would want a chip with atleast 1.3000V VID.

The things will probably change as the CPUs get more mature (in terms of manufacturing process) thou.