Core i7 doesn't produce good results in gaming

[kmote]

what is ther to re-analyse?

 

[zmatt]

Super Pi was written in 1995. Back then a lot of the code optimizations weren't around yet. I doubt you could get such uneven scores with super pi.

 

[Apokalipse]

you could get "optimisations" such as

if (CPU.ID == Intel)
{
//run this codepath
}
else
{
//run this deliberately slower codepath
}

I don't really know Assembly, so I wouldn't know how exactly it might look there. But other people do and have identified codepaths where Intel's compilers deliberately made non-Intel CPU's run unnecessarily slower codepaths, even if the source code has no such thing.

 

[zmatt]

CPU-Z was also written at a university. I doubt that they intentionally did that. Could it just be that Intel chips are inherently better at it?

 

[Apokalipse]

CPU-Z was also written at a university. I doubt that they intentionally did that. Could it just be that Intel chips are inherently better at it?

Even if they were, I really don't think there's twice the performance difference.

 

[zmatt]

Well the difference goes over into other things aswell. If Intel only did that much better in Super PI then I would be inclined to agree with you.

 

[Apokalipse]

Well the difference goes over into other things aswell. If Intel only did that much better in Super PI then I would be inclined to agree with you.

If Intel's compilers make slower codepaths for non-Intel processors (which there is evidence for), superpi wouldn't be the only program to be Intel biased.

 

[kmote]

Do we even know that super pi is compiled on the intel compiler?