What causes different CPU performance by motherboards

Mathias_8

Commendable
Apr 8, 2016
58
0
1,640
What causes different CPU performance in motherboard reviews in benchmarks like Cinebench 15 anyway? Does this all come down to variations in boost clock? Core utilisation somehow?
 
Solution


There will always be better and worse programmed motherboard firmware, and more or less devices (which take processing time to check), so you will see some minor, nearly insignificant differences. Most of the time these variations in capability and performance are well under the statistically significant amount (usually less than 1% in tests that vary by up to 5%)
A million different things, but usually it's just random variability and reviewers failing to run the suite a statistically significant number of times. Occasionally you will have a motherboard with funky firmware that slows down the system (especially audio related stuff), but usually it's not a number significant enough to consider in a purchase.
 

Mathias_8

Commendable
Apr 8, 2016
58
0
1,640
But if it's consistent? Say, a motherboard always comes in at a certain place in Cinebench and never at a top spot, but is better in other tests. And with others it's the other way around. It looks like that with the Asus Z170I and MSI Z170I. It's win and lose in some for both, but MSI always comes ahead in Cinebenc (as far as I know). And high performance boards are usually consistent.
 


I would suggest you take a course in statistics. Consistent results do not mean accurate results, especially if there's only one trial done
 

Mathias_8

Commendable
Apr 8, 2016
58
0
1,640
I meant consistent across different trials of course, what else? That's all I'm talking about. Indeed I'm saying that there's not much consistency within a single run of different benchmarks.
It doesn't answer the question anyway. These tests exist, and also relative statistical performance.Whenever one asks a question like this and about the technical reasons, all I get are universal and technically indifferent relativizations. I guess the reviewers are just a bunch of weirdos then who don't know what really matters and is going on? I didn't get that memo.

However they do fail to provide some proper analysis of the results, for the most part... And I daresay that would not be too hard: clockspeed monitoring pehaps? Well, depends on what is the point of the reviews and the reason for dfferent performance anyway. Usually it's said it's "phases" and vrm cooling... So what does it do? At stock speeds, mind you. Simple, general question.

I'm not asking for a purchase advice or even anyone's opinion on a given result, dammit.
 


There will always be better and worse programmed motherboard firmware, and more or less devices (which take processing time to check), so you will see some minor, nearly insignificant differences. Most of the time these variations in capability and performance are well under the statistically significant amount (usually less than 1% in tests that vary by up to 5%)
 
Solution

Tradesman1

Legenda in Aeternum
Generally as said above it often a BIOS thing, can also be the power phase design. They might run a hair more voltage through given traces, ofr have seen mobos that advertised as standard 100 BCLK but come out of the box at slightly higher 100.04 or whatever which gives them a boost, there a number of things that could play in