Message-Id: Comments: Authenticated sender is From: "Salvador Eduardo Tropea (SET)" Organization: INTI To: DJ Delorie , djgpp-workers AT delorie DOT com Date: Wed, 28 Apr 1999 11:16:33 +0000 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: Re: [dma AT hpesdma DOT fc DOT hp DOT com: Performance Observation] In-reply-to: <199904281552.LAA00704@envy.delorie.com> References: <199904281342 DOT OAA12455 AT melchard DOT ee DOT ucl DOT ac DOT uk> (cmatraki AT ee DOT ucl DOT ac DOT uk) X-mailer: Pegasus Mail for Windows (v2.54) Reply-To: djgpp-workers AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk DJ Delorie wrote: > > I'm not sure whether 32-byte alignment is the best option, but the waste > > 8-byte alignment is my vote. Note that a global alignment is wrong, > because ctor/dtor tables will get corrupted if they're padded. > > > Another related question is, what is the guaranteed alignment from malloc? > > malloc returns pointers that are guaranteed to be sufficiently aligned > for any use. Of course, on the ix86, any alignment is "sufficient". > However, djgpp's current malloc aligns to 8-byte boundaries. I vote for 16 bytes. Why? 128 bits data path is 16 bytes and some seconds ago I taked a look into a Win32 program (almost sure is a VC++ one) and I saw upto 14 NOPs, so it looks like 16 bytes is what they use for code. Of course you can see the bloat in the code, it is full of NOPs. Also: NOP is not the only one used, but is very easy to see with an Hexa editor. SET ------------------------------------ 0 -------------------------------- Visit my home page: http://welcome.to/SetSoft or http://www.geocities.com/SiliconValley/Vista/6552/ Salvador Eduardo Tropea (SET). (Electronics Engineer) Alternative e-mail: set-soft AT usa DOT net set AT computer DOT org ICQ: 2951574 Address: Curapaligue 2124, Caseros, 3 de Febrero Buenos Aires, (1678), ARGENTINA TE: +(5411) 4759 0013