You are not logged in.
- Topics: Active | Unanswered
Pages: 1
#1 2013-11-23 11:19:57
- louis_riviera
- Member
- Registered: 2013-09-23
- Posts: 61
FastMove
Hello!
Here is an MMX, SSE, SSE2, SSE3 optimized Move function but AFAIK it don't support x64.
Maybe possible to include in SynCommons.pas 
Offline
#3 2013-12-11 15:57:10
- louis_riviera
- Member
- Registered: 2013-09-23
- Posts: 61
Re: FastMove
Actually is not the same. Your supports SSE2 MAX! That one supports from MMX - SSE3. Now cpus support SSE 4.1, 4.2 etc
So not the same.
Last edited by louis_riviera (2013-12-11 16:03:43)
Offline
#4 2013-12-11 17:50:32
Re: FastMove
I doubt there will be a huge difference between SSE2/SSE3/SSE4 in this case.
There is already almost no speed difference between FPU move and SSE2 move!
Perhaps for some border-case tests, not for real world applications.
Most of the move() operation is about a very small amount of data, where the already existing version is OK (but not for x64).
Main bottleneck will be the CPU cache, not the instruction set used.
And it will also depend on the CPU model it runs on...
Offline
#5 2013-12-13 12:24:00
- louis_riviera
- Member
- Registered: 2013-09-23
- Posts: 61
Re: FastMove
There is no need for syncommons optimized move. ScaleMM2 has added this for SSE3.
https://code.google.com/p/scalemm/sourc … n117&r=117
Last edited by louis_riviera (2013-12-13 12:24:20)
Offline
#6 2013-12-13 16:01:36
Offline
#7 2013-12-14 10:27:08
- louis_riviera
- Member
- Registered: 2013-09-23
- Posts: 61
Re: FastMove
What is the license of SynCommons? I would like to extract all the low level RTL replacements and add it in to a more light unit. With your license ofcourse.
Offline
#8 2013-12-14 17:54:04
Re: FastMove
See http://synopse.info/forum/viewtopic.php?pid=4467#p4467
This is also clearly stated in the framework documentation.
Enjoy!
Offline
Pages: 1