Hi!
6-Май-2005 03:40 [EMAIL PROTECTED] (Eric Auer) wrote to
[email protected]:
EA> shr ecx,1
EA> REP MOVS DWORD PTR [ESI],DWORD PTR [EDI];
EA> BIG_NOP;
EA> adc ecx,ecx
EA> REP MOVS WORD PTR [ESI],WORD PTR [EDI];
EA> BIG_NOP;
Well, you tempt me. :) Ok, if someone specify conditions:
- range of sizes and multiplicity for copied areas,
- which alignment (or missing) is expected for areas,
- how much code may be aligned,
I may prepare some copy routine versions (for example, for multiple-piping
CPUs (like Pentium) declared, that unrolled and aligned loop of MOVs should
be faster, that REP MOVSD).
EA> Where BIG_NOP is a NOP with 67h prefix. This is commented as
EA> being a workaround for 386 / b3 version CPU bugs.
Hm. May you quote description of this bug?
EA> The second REP is taken only 0 or 1 times.
So what?
EA> Some completely unrelated issue: EMM386 and HIMEM are very quiet
EA> at start-up.
I think contrary - they say too much and not nicely formated.
EA> alloc'ing many VCPI pages is. That would be Allocate4KPageFromPoolBlock
EA> (only used for VCPI?) and ExpandAnyPoolBlock / ExpandCurrentPoolBlock,
EA> Allocate16KPageFromPoolBlock (for EMS), AllocateXMSForPool, and their
EA> counterparts Free16KPage and Free4KPage. All very complex code written
EA> by Michael here (thanks!) but maybe you (Arkady) find some interesting
EA> optimization points anyway: Try to limit your search to nested loops
EA> which operate on longer lists.
This will have sense, if Michael will communicate with me. Michael?
-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.
Get your fingers limbered up and give it your best shot. 4 great events, 4
opportunities to win big! Highest score wins.NEC IT Guy Games. Play to
win an NEC 61 plasma display. Visit http://www.necitguy.com/?r
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel
