Re: amber-developers: Altivec wrapper for PMEMD

From: Mengjuei Hsieh <mengjueh.uci.edu>
Date: Wed, 16 Jun 2004 14:04:45 -0700

Dear Dr. Duke,

Yes, you are right, I should use something like "ALTIVEC" for the flag.
It's
also a good flag if some people want to use LinuxPPC with G4/G5. And I
attached the updated version of files, too. (If anyone wants to try my
modification, please, be sure to backup the original files.)

By the way, I haven't thought about submitting the changes, I will
probably
wait for the engineers doing AMBER optimization from Apple Computer to
catch
up after their developer conference in the end of this month. They
probably
have better idea to play with FFT.

Best,
-- 
Mengjuei Hsieh, Luo group, Molecular Biology and Biochemistry, University
of
California Irvine. Tel: 49562, Address: 3144 Natural Science I building,
UCI, Irvine CA 92697-3900. Group Homepage: http://rayl0.bio.uci.edu/
On 6/15/04 12:47 PM, Robert Duke at rduke.email.unc.edu wrote:
> Mengjuei -
> Thanks much for working on this apple/pmemd stuff.  I have no apple
> resources whatsoever, so I am going to stay out of the fray; please just
> thoroughly test before checkin.  I would also recommend that maybe we
make
> the conditional something like APPLE_VEC (or something else that
effectively
> indicates the scope of the usage) instead of vecLib.
> Regards - Bob Duke
> ----- Original Message -----
> From: "Mengjuei Hsieh" <mengjueh.uci.edu>
>> Hi,
>> Although there are few people playing AMBER on Macintosh here, this is
just
>> for your information.
>> I attached a wrapper for vrsqrtf (1/r^.5) in altivec (available in
Apple
>> G4/G5 machine) on MacOS X. You can call vrsqrtd_wrapper(input array,
output
>> array, number of item in the array) from PMEMD. This wrapper accepts
double
>> precision input and output while altivec is still thinking in single
>> precision, so beware of those accuracy issues.
>> And the patch file for PMEMD as a minimal modification (just an
added-in
>> modification, it won't delete any line or modify any loop in the code).
>>[skipped]





Received on Wed Apr 05 2006 - 23:50:01 PDT
Custom Search