By: Exophase (exophase.delete@this.gmail.com), March 6, 2015 11:45 pm
Room: Moderated Discussions
David Kanter (dkanter.delete@this.realworldtech.com) on March 7, 2015 12:37 am wrote:
> It's a totally different thing for Intel, AMD, ARM, or whoever, to claim that AES acceleration
> via dedicated hardware will impact general workloads in a meaningful way.
>
Do they? I don't think they do. Both instruction sets have some amount of niche functionality that isn't useful in most workloads, particularly in the SIMD parts.
The problem here is Geekbench letting it effect the integer score. I kind of wonder if that was even a deliberate decision on their part, and that they didn't instead just use a library that ended up using those instructions. Either way, it really should be changed.
(that's not to say that the crypto instructions are completely irrelevant to mobile, though)
> It's a totally different thing for Intel, AMD, ARM, or whoever, to claim that AES acceleration
> via dedicated hardware will impact general workloads in a meaningful way.
>
Do they? I don't think they do. Both instruction sets have some amount of niche functionality that isn't useful in most workloads, particularly in the SIMD parts.
The problem here is Geekbench letting it effect the integer score. I kind of wonder if that was even a deliberate decision on their part, and that they didn't instead just use a library that ended up using those instructions. Either way, it really should be changed.
(that's not to say that the crypto instructions are completely irrelevant to mobile, though)