|
|
Hi folks,
I've just read that end of 2017 Intel plans to release a special Xeon
Phi variant codenamed "Knights Mill" (not to be confused with "Knights
Hill"), aimed specifically at deep learning applications.
In case you happen to come across this CPU, keep your fingers off it. It
will probably sacrifice a lot of double-precision floating-point
performance (which is what POV-Ray needs) for improved single- and
half-precision arithmetics (which in POV-Ray are of lesser importance
and entirely irrelevant, respectively).
Post a reply to this message
|
|
|
|
On 11/09/16 14:17, clipka wrote:
> Hi folks,
>
> I've just read that end of 2017 Intel plans to release a special Xeon
> Phi variant codenamed "Knights Mill" (not to be confused with "Knights
> Hill"), aimed specifically at deep learning applications.
>
> In case you happen to come across this CPU, keep your fingers off it. It
> will probably sacrifice a lot of double-precision floating-point
> performance (which is what POV-Ray needs) for improved single- and
> half-precision arithmetics (which in POV-Ray are of lesser importance
> and entirely irrelevant, respectively).
>
Yup. Heard about it, but it's not a problem; I won't be able to afford it.
John
--
It's not about bravery.
It's about doing what I need to do to win
Post a reply to this message
|
|