Couple Turbo Boost questions

[sunsanvil]

I realize this is probably Intel 101 but I'm still a little unclear on a couple of basic points. For conversation sake, I have an i5-3570 which is quoted as nominally 3.4Ghz, max boost of 3.8Ghz (single thread) or 3.6Ghz (multi thread).

First question: In the Windows environment, it even possible to not have more than one core active? I have never been able to log more than 3.65Ghz out of it. I've tried every "single thread" benchmark/utility I can find. Temp is not a factor in that even stone cold it never goes higher than that.

Second question: Despite the notion that achieving spec turbo boost speeds is dependent on power and thermals being "within limits", is there any point at all to "better" cooling and elaborate 10 phase VRM stages (on anything other than overclock platforms)? I've never NOT gotten the full 3.6Ghz multi-core max out of it. Even on torture tests (Prime95 etc), it has never not sustain 3.6Ghz. 30 or 82 degrees, doesn't matter, it has never had to throttle. This is with a crappy stock cooler on a very basic DH77KC mainboard.

 

[gingerrankin]

you can set active cores number in windows. type "msconfig" in cortana hole. Then "start" and "advanced options".

 

[smorizio]

on your second question....cooling and vrm. with cooling the cooler the cpu is over it life time the longer the cpu will last. ie two rigs one with the stock intel cooler running at 50c. and another with eve 212 cooler runs at 30c. both are within intel cpu temp range. the computer with the hotter cpu temp the chip itself can fail sooner. pc parts run better when cold then hot. with vrm;s there made to turn ac voltage from the power supply into dc voltage. the better the vrm the less ripple or noise gets into the mb. to much ripple and a mb chips can cause dc parts to see high and low signals and turn off and on...causing bsod and lock ups. mb with cheap vrm or not a lot on cheaper mb....under gaming loads the vrm can over heat and fail or go into thermal runaway. also one issue with the vrm they have to have good bios and be sized right. a few years ago there were mb that the vrm were blowing up and caught on fire when a high end cpu and gpu was used and over clocked. the vrm were under sized and the bios was not program to strop the amp load for going over the vrm limits. there were youtube videos of people turning the mb on for the first time and it going up in flames.

 

[sunsanvil]

you can set active cores number in windows. type "msconfig" in cortana hole. Then "start" and "advanced options".

Interesting notion but it doesn't produce the result I thought it would: Instead of finally seeing the published single core max 3.8Ghz, setting that to 1 seems to result in the CPU getting locked to one of two states only, 1.6Ghz or 3.4Ghz.

on your second question (...) the cooler the cpu is over it life time the longer the cpu will last (...) the better the vrm the less ripple or noise gets into the mb.

That much I do understand. My question is more about the relevance (if any) to achieving stock (non-oc) turbo boost speeds.

To re-frame my questions a little...

From what I can see, the published max turbo boost speeds (which are always based on single-core usage) are pointless or at very least misleading as it seems only a very special software environment would ever result in only one core loaded for any meaningful time interval. ie, I will never, in any real world scenario, see 3.8Ghz on this 3570. Or am I missing something in my understanding here?

It also seems misleading to constantly attache the "as long as thermal and power limits allow it..." line to everything related to turbo boost since, from what I can discern, nothing special is required to achieve the all-core turbo boost speed anyway (in my case 3.6Ghz). Or am I missing something in my understanding here?