search

desson104 / nulldc

Automatically exported from code.google.com/p/nulldc
0 stars 0 forks source link

Task : Hand optimise some ASM / SSE/SSE2 instructions for low clocked processors. Rewrite ARM7 SSE for better performance + Recompiler. #158

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Oh yeah, i went there :D

This can be a tracking issue for when someone actually looks at optimising 
nulldc at code level instead of flagging for SSE2 optimisations in the build 
config, as well as ARM7 improvements.

Original issue reported on code.google.com by danialho...@gmail.com on 30 Jun 2010 at 8:36

GoogleCodeExporter commented 9 years ago 
Yeah well... optimizations are welcomed. The "problem" is finding someone 
willing to go through all that.

Original comment by Mr.PsyMan on 30 Jun 2010 at 8:43

GoogleCodeExporter commented 9 years ago 
Performance improvement on Core 2 and higher

"use the SSE4 dot product instruction to emulate the SH4's dot product 
instruction"

Original comment by danialho...@gmail.com on 30 Jun 2010 at 1:04

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Not gonna happen when I'm around.

Tell someone else who gives a fuck.

Original comment by mudl...@tpg.com.au on 1 Jul 2010 at 11:12

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Its not do or die mud, if it was used it'd maintain a fallback if the function 
was not present.

Original comment by danialho...@gmail.com on 2 Jul 2010 at 5:17

GoogleCodeExporter commented 9 years ago 
Oh hi muddy! Sup? Just take it easy with the personal vendettas, k? :)

Original comment by Mr.PsyMan on 2 Jul 2010 at 7:27

GoogleCodeExporter commented 9 years ago 
The thing is, most of the overhead isn't on the opcode implementations itselfs 
-- so it'l make little difference

Original comment by drkIIRaziel on 6 Jul 2010 at 5:06

GoogleCodeExporter commented 9 years ago 
OMFG....  Would you listen to drkIIRaziel, everyone ???  just listen to him... 
he clearly speaks out his ass, and proves he is an idiot evertime !!   If there 
is very little overhead going into the OPCODE then why not go ahead and disable 
the SSE2.  I mean really, you bunch of dumbasses.... its already been proven 
that its NOT needed and that there is NO need for it.  Once again.... Fuck You 
All.   thank you for reading.

Original comment by test54...@gmail.com on 8 Jul 2010 at 12:44

GoogleCodeExporter commented 9 years ago 
I vote we enable SSE4.

Original comment by danialho...@gmail.com on 8 Jul 2010 at 12:58

GoogleCodeExporter commented 9 years ago 
Hmmm....
People here are strange, they complain even though they get this for free...

Original comment by Koutetsu...@gmail.com on 8 Jul 2010 at 12:23

GoogleCodeExporter commented 9 years ago 
test540va: Actually, i was gonna remove that sse2 option, but then you started 
bitching so i decided not to care. Just FYI :)

Original comment by drkIIRaziel on 8 Jul 2010 at 11:28

GoogleCodeExporter commented 9 years ago 
Actually, when he started bitching I considered forcing SSE2 on the nullDC 
executable too. Then I realized I don't care enough either. xD

Original comment by Mr.PsyMan on 9 Jul 2010 at 12:23

GoogleCodeExporter commented 9 years ago 
drkIIRaziel:   Hahaha, I was just kidding man.  Can't ya take a joke ?  
Everyone here knows this is a wonderful emulation project.  Well bud, I hope I 
didn't cross any lines.  I know if anyone out there can handle this issue it is 
U.  Peace out Bro...

p.s.  i tell all my friends to use Null DC.

Original comment by test54...@gmail.com on 9 Jul 2010 at 1:11

GoogleCodeExporter commented 9 years ago 
LOL!

Original comment by vadimurputi on 9 Jul 2010 at 7:44

GoogleCodeExporter commented 9 years ago 
So, SSE2 for all then?

Original comment by KrossX3 on 11 Jul 2010 at 8:57

GoogleCodeExporter commented 9 years ago 
Well, if it's possible, I'd prefer that we revert back to SSE1.  I have a 
brand-new computer, which supports SSE4 (quad-core), but I'd hate to see people 
with older computers get left out.  

I really do appreciate the work you've all done.  Honest.  It really is quite a 
fabulous piece of work.  If we really feel the need to force SSE2, then do so.  
Perhaps you could add an SSE/SSE2 option?

Original comment by hookonso...@gmail.com on 11 Jul 2010 at 11:30

GoogleCodeExporter commented 9 years ago 
Okay, Psy, I had a change of heart.

Original comment by mudl...@tpg.com.au on 21 Jul 2010 at 11:17

GoogleCodeExporter commented 9 years ago 
A recompiler is the best bet. And Exophase granted me permission. So....

Original comment by mudl...@tpg.com.au on 21 Jul 2010 at 11:21

GoogleCodeExporter commented 9 years ago 
http://code.google.com/p/nulldc/issues/detail?id=169

Original comment by mudl...@tpg.com.au on 23 Jul 2010 at 1:27