By: Michael S (already5chosen.delete@this.yahoo.com), January 11, 2015 5:44 am
Room: Moderated Discussions
Aaron Spink (aaronspink.delete@this.notearthlink.net) on January 10, 2015 7:21 pm wrote:
> juanrga (nospam.delete@this.juanrga.com) on January 10, 2015 6:32 pm wrote:
> > Any GPU that I know fits the manycore definition.
> >
>
> How? Most GPUs have the same if not less cores than something like a E5-2699v3 which no one would
> consider manycore. For instance something like the GM204 is basically a 16 core design.
>
Do you count Maxwell SMM as single core?
I think, it's more reasonable to count it as quad-core module. It would be more consistent with what we consider "a core" in traditional or semi-traditional (Bulldozer) CPUs.
> > I am claiming that KNL is manycore because KNL fits the manycore
> > definition and is aimed to compete against GPGPUs.
> >
> KNL doesn't have enough cores to even be considered a real manycore.
What is your threshold? Mine is 50.
> Not to
> mention that KNL is using multi-issue OOE cores.
The integer part is most likely OoO, but I am not expecting VPU ALUs to be OoO.
As far as I am concerned, the most intriguing uArch question about KNL is if VPU loads can be issued ahead of preceding VPU ALU and VPU store operations and if yes then to which degree.
> if anything, from programming
> model to functionality, KNL is just a high core count multi-core.
>
>
> juanrga (nospam.delete@this.juanrga.com) on January 10, 2015 6:32 pm wrote:
> > Any GPU that I know fits the manycore definition.
> >
>
> How? Most GPUs have the same if not less cores than something like a E5-2699v3 which no one would
> consider manycore. For instance something like the GM204 is basically a 16 core design.
>
Do you count Maxwell SMM as single core?
I think, it's more reasonable to count it as quad-core module. It would be more consistent with what we consider "a core" in traditional or semi-traditional (Bulldozer) CPUs.
> > I am claiming that KNL is manycore because KNL fits the manycore
> > definition and is aimed to compete against GPGPUs.
> >
> KNL doesn't have enough cores to even be considered a real manycore.
What is your threshold? Mine is 50.
> Not to
> mention that KNL is using multi-issue OOE cores.
The integer part is most likely OoO, but I am not expecting VPU ALUs to be OoO.
As far as I am concerned, the most intriguing uArch question about KNL is if VPU loads can be issued ahead of preceding VPU ALU and VPU store operations and if yes then to which degree.
> if anything, from programming
> model to functionality, KNL is just a high core count multi-core.
>
>
