• 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.

FIVR rumoured to return in Icelake

Soldato
Joined
22 Aug 2008
Posts
8,338
http://www.hardwareluxx.com/index.p...tage-regulator-for-skylake-and-kaby-lake.html

By integrating the voltage converter Intel wanted to simplify the power supply on the motherboard. Before FIVR, motherboards had to suppply five different voltages to the CPU package: Vcore, vgpu, VCCSA, VCCIO and PLL. In addition, this step was supposed to give Intel more control over the power supply of CPUs, which also created efficiencies in regards to power consumption.

Now, FIVR seems to be history for the upcoming "Skylake" and "Kaby Lake" processors, as Intel plans to release the two upcoming CPU generations without it. However, "Ice Lake" as the generation after next is codenamed, will then one again use a FIVR. A source not only provided us with the news, but also gave a reason: the thermal design power of the overall CPU package was too high. This seems to have had a negative impact on the overall system, Intel has to save the extra heat elsewhere.

Good idea? Bad? There will have been two shrinks by then so density will be possibly making it even harder to cool than Haswell.
 
So the point of this is to reduce heat on the chip? I would have thought an internal voltage regulator would have produced more heat, not less
 
So the point of this is to reduce heat on the chip? I would have thought an internal voltage regulator would have produced more heat, not less

Well if you look at Haswell, it certainly does seem to produce more heat due to this. Also, when you use adaptive voltage the VCore spikes above what you've set.

I'm glad that they're dropping it with Skylake. Unless they can improve it to the point where it works in a sane manner then I don't think they should bring it back.
 
Seems a strange one - I've been playing about with a bit of electronics lately using ICs of various types i.e. USB DACs and quite often in the datasheet they recommend bypassing the internal voltage regulators using external regulators for high performance operation due to better characteristics (especially current handling) and less thermal considerations on the package.
 
With it internally I would have thought they could have much tighter control over very fine voltage adjustments and gating.
 
I'm sure they think it;'s a good idea if they're doing it and they of course know more than any of us... but I was happy to see it go for Skylake so not overjoyed to see it return. Also the reason seems very back to front to me for desktops though I guess if you're looking at total system power not package power as talked about then it makes some sense.
 
Another thing to consider, if they are going to keep swapping and changing this then there will be no choice but for a new motherboard for each and every new chip, rather than being able to slot a couple of chips into the same boards.
 
Who mentioned Skylake? OP was referencing Icelake.

I was referring to Icelake not Skylake. I understand it'll be a different socket again from Skylake and Kaby Lake. Its too early to say for sure though so who knows....
 
Last edited:
Back
Top Bottom