Anyone else dislike FIVR on the desktop?

[jana519]

Is it just me, or is voltage regulation something that should be left on the motherboard for desktop CPUs? It's not enough silicone to change the form factor or cost, and it doesn't bring down peak power usage either. It definitely does increase the CPU temperature and affect higher clock speeds. Seems like a great idea for mobile, but for desktop not so much.

 

[ShintaiDK]

You will most likely be able to say the exact opposite when Skylake hits and you run 512bit AVX3.2 tests.

 

[jana519]

You will most likely be able to say the exact opposite when Skylake hits and you run 512bit AVX3.2 tests.

I'm not too familiar with AVX tbh. Does it have some power requirements that would benefit from a FIVR?

 

[ShintaiDK]

I'm not too familiar with AVX tbh. Does it have some power requirements that would benefit from a FIVR?

I think FIVR gets its main flak from people running AVX based benchmark applications, where Haswell performs roughly twice vs Ivy Bridge. With 256bit AVX Haswell increases the vcore with 100mv due to 256bit cache width vs 128bit on Ivy Bridge.

 

[Fjodor2001]

So it's primarily the AVX aspects that determine if FIVR is a good thing or not. Huh? 😵

 

[PPB]

You will most likely be able to say the exact opposite when Skylake hits and you run 512bit AVX3.2 tests.

AVX2 gains are laughable for the massive increase in pipeline width it is needed. But i guess your post was born more out of wishful thinking than the facts at hand.

 

[Khato]

AVX2 gains are laughable for the massive increase in pipeline width it is needed. But i guess your post was born more out of wishful thinking than the facts at hand.

They're not laughable in synthetic benchmarks/stress tests like those frequently used to test stability. No, in that case you can see pretty much the theoretical performance gain and corresponding increase in power consumption.

Yes, the FIVR does result in a marginal increase in CPU power consumption under load scenarios. And as such if you're determined to have AVX stability the minor increase in power consumption will result in a reduction in maximum overclock under air/water due to thermal limits. But soon as you ignore AVX then it's a non-issue.

Oh, and overclocking is the only area negatively impacted by the FIVR that I'm aware of. Hardly a reason to dislike it considering all of the positive aspects.

 

[Deders]

AVX stability is great, when you have applications that can actually make use of it

 

[aigomorla]

id rather have the voltage regulators on the board.
The reason is, im fairly certain, you can get better quality voltage regulation at the higher end enthusiast lvl boards, then what is given inside the cpu.

Oh, and overclocking is the only area negatively impacted by the FIVR that I'm aware of. Hardly a reason to dislike it considering all of the positive aspects.

^ this makes me and others like me a very sad panda.. 🙁

If your a budget or mid tier system builder, then yea, its hardily a reason.
However if your the enthusiast level system builder, who looks at $300+ boards... its a sad panda...

 

[bgt]

FIVR means less loss/waiste and easier cooling so why criticise it??

 

[Homeles]

id rather have the voltage regulators on the board.
The reason is, im fairly certain, you can get better quality voltage regulation at the higher end enthusiast lvl boards, then what is given inside the cpu.

Depends how you define quality.