Path: chuka.playstation.co.uk!news From: Andrew Partington Newsgroups: scee.yaroze.freetalk.english Subject: Re: Performance Date: Fri, 27 Jun 2003 05:20:51 +0000 Organization: PlayStation Net Yaroze (SCEE) Lines: 32 Message-ID: References: <01c33bef$d92f8ae0$8500a8c0@portable-pal> NNTP-Posting-Host: public2-ward1-6-cust140.oldh.broadband.ntl.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030313 X-Accept-Language: en-us, en In-Reply-To: <01c33bef$d92f8ae0$8500a8c0@portable-pal> Hi Pal, I tried reducing my OT length and size as you suggested, no luck i'm afraid. It wasn't a massive OT anyway, but making it smaller is always helpful. The only time I do a DrawSync() is in my vsync callback routine, before GsSwapDispBuff() and GsDrawOT(). I'm pretty sure that's the only place where drawing needs to be synchronised with anything, at least in this case. Even then, I do a DrawSync(1) which according to the manual forces drawing to terminate whether its finished or not. It could be that i'm syncing in the wrong place, but it seems that is where it should be. Looks like it could be the dynamic TMD route, unless SCE want to release an updated libps.a with the scalable/rotatable version of the GsBG handler :) Cheers, Andy P pal wrote: > Just a thought, sorry if it's obvious: make sure that your ot's length is > as small as you need. This increases insertion dramatically. Also, just in > case, double check that drawing and logic are properly handled in parallel > - it's easy (well, it was easy for me at least ;) ) to put some DrawSync > where it doesn't belong and nothing's gonna tell you until The Big Slowdown > happens. >