By: David Kanter (dkanter.delete@this.realworldtech.com), November 16, 2012 8:59 am
Room: Moderated Discussions
Eric Bron (eric.bron.delete@this.zvisuel.privatefortest.com) on November 16, 2012 4:57 am wrote:
> Thanks for a really nice read.
>
> At page 4 you write that "while the divider [...] unit remain at 128-bits" this
> looks a bit odd since it is 256-bit wide in Ivy Bridge (the throughput was doubled
> from SNB to IVB for 256-bit SQRTPS and DIVPS instructions for example)
>
> Are you really sure it will be regressed to 128-bit again as in Sandy Bridge?
I think it was only 128-bit in the first place (I mean the internal datapath here).
DK
> Thanks for a really nice read.
>
> At page 4 you write that "while the divider [...] unit remain at 128-bits" this
> looks a bit odd since it is 256-bit wide in Ivy Bridge (the throughput was doubled
> from SNB to IVB for 256-bit SQRTPS and DIVPS instructions for example)
>
> Are you really sure it will be regressed to 128-bit again as in Sandy Bridge?
I think it was only 128-bit in the first place (I mean the internal datapath here).
DK
