Let’s talk about it next week.
On the fly native code generation for specific hardware is a capability that I haven’t tapped yet.
.Chris.
> On Feb 12, 2017, at 2:49 PM, Scott Le Grand <varelse2005.gmail.com> wrote:
>
> Make it run transparently across all GPUs and CPUs and you have a winner.
> Depending on your viewpoint, the future is GPUs or looks just like them.
>
> If only Nvidia hadn't turned evil...
>
> On Feb 12, 2017 1:29 PM, "Christian Schafmeister" <drschafmeister.gmail.com>
> wrote:
>
> This is why I’m developing CANDO.
> It doesn’t have any of these problems that led me to give up on Python a
> decade ago.
> CANDO generates compiled code, does automatic memory management when you
> want it,
> interoperates with C++ and C and has a full featured chemistry library
> built into it.
>
> Python was a trap for scientific computing.
>
>
>
>> On Feb 12, 2017, at 2:23 PM, Scott Le Grand <varelse2005.gmail.com> wrote:
>>
>> And of course place it in a Docker container with the precise set of conda
>> dependencies because that's f****** awesome.
>>
>> This is why we no longer can have nice things in an age of insane
>> computational firepower that we insist on reducing to single-threaded
>> interpreted code. 2003 is calling and it wants to offer you a sweet deal
>> on Pentium 4 CPUs.
>>
>> On Feb 12, 2017 1:03 PM, "Jason Swails" <jason.swails.gmail.com> wrote:
>>
>>>
>>>
>>>
>>>
>>> --
>>> Jason M. Swails
>>>> On Feb 12, 2017, at 11:12 AM, Scott Le Grand <varelse2005.gmail.com>
>>> wrote:
>>>>
>>>> Python's GIL hates America. We should build a wall around Python and
>>> make
>>>> Guido Van Rossum pay for it.
>>>
>>> Python is a language spec. It has no GIL. Nor does several of its
>>> implementations. Just use one of those ;)
>>> _______________________________________________
>>> AMBER-Developers mailing list
>>> AMBER-Developers.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>>>
>> _______________________________________________
>> AMBER-Developers mailing list
>> AMBER-Developers.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Sun Feb 12 2017 - 12:00:03 PST
