Thema: RE: Hi! Datum: 08.10.96 12:17:12 From: wilki1k@nectech.co.uk (Wilkins, Keith) To: Chrissalo@aol.com (Chrissalo) Chris, > If you don't mind I would like to set up another vectrex page with > 'my' version of your emulator inclusive all sources. Why not release DVE 1.1, I've no objections and it would mean there was still a single source release and a single emulator (less confusion), you can edit the readme with the changes you've done and put your name on the author list and make other changes ?? If you wan't to release your own version and have your own site I really don't mind. Let me know what you think. If you wan't you can take over DVE development, thats one of the reasons I released the code. I have a family (2 year old daughter) and I was spending too much time on this. Things I wanted to add: Proper joystick support (approx 2% speed hit) I have the joystick code I wrote for Stella if you want it, its not bad, fully autocalibrating, none of this move to top left & press a button shit. Light pen support via mouse. 3D Imager support (Hard, maybe use read/green glasses) Modeling of integrator inaccuracies > I don't think your vesa routines were all that efficient (sorry, but I gotto > say that...)... I guess it didn't even run with all cards, since you assumed > ... much stuff deleted Ouch! I guess nobody likes critisism. To be honest I have no experience writing VESA stuff, I just cobbled it together from some bits and pieces I found on x2ftp.oulu.fi so I'm not that suprised about its efficiency. Towards the end of doing DVE my enthusiasm was flagging... > (Ugh, if that sounds bad... sorry... I really like your emulator... it is just > that grafic routines...) Don't worry about it. > These are some benchmarks: > Your version: about: 47% > Banked Framebuffer: 53% > Linear Framebuffer version: 59% Nice improvement for not so much effort. What you really want is source profiling to find where the real time is being chewed up, I never managed to make the profiler work properly. > Probably the next thing I'll try to do is a 'better' support for overlays. > I first thought about an editor... but since most overlays are available I did it in the quickest way. If you use the overlays then you've got to mess around scaling them down to the screen size and also aligning them with the beam, good luck... Why not keep both systems ?? > at a certain ftp site... I think I'll settle for a picture load routine > first 16 colors for the beam, 16-240 for each 16 colors of > overlay colors and 241-255 for 16 different border colors. At the moment it is mapped into 16 beam colours with 16 intensities of each colour. Hence the limit on 16 overlay colours. Carful you don't lose the intensity, tit makes a difference on some games. > I recently had a look at emu.exe, that atari vector games emulator !!! > It is VERY fast... any clue how further optimiziation of your emulator may be > accomplished? What section is the most power hungry? (See the profiler comment) I did have some specific ideas. One of the reasons its so slow is that every emulator loop checks the counter & serial port status as they operate some of the vector hardware independantly of the CPU. What I considered doing was: IF (timer1_count > limit) AND (timer2_count > limit) AND (serial port inactive) { Use CPU model with no peripheral checking and no loop so: cycle_count+=instuction cycle count timer1_count+=instruction cycle count timer2_count+=instruction cycle count } else { Use current single cycle CPU model checking peripherals every cycle } limit is set to max instruction cycle count + 1 So when the screen is inactive then we use a dumb CPU model that counts cycles but doesn't execute the FETCH, WAIT, WAIT.... WAIT, EXECUTE model round the loop, this should make things execute about 5x faster when the hardware is not in use. I ran out of enthusiasm for doing this change as its fairly major. Good luck. Keith ----------------------- Headers -------------------------------- From <@nectech.co.uk:wilki1k@nectech.co.uk> Tue Oct 8 07:16:51 1996 Return-Path: <@nectech.co.uk:wilki1k@nectech.co.uk> Received: from nectech.nectech.co.uk (nectech.nectech.co.uk [194.129.183.2]) by emin01.mail.aol.com (8.6.12/8.6.12) with SMTP id HAA21950 for ; Tue, 8 Oct 1996 07:16:49 -0400 Received: from [193.116.199.251] by nectech.nectech.co.uk id aa00723; 8 Oct 96 11:59 BST Received: from parodius by lego.nectech.co.uk (4.1/25-eef) id AA12185; Tue, 8 Oct 96 12:08:36 BST Received: by parodius with Microsoft Mail id <325A453E@parodius>; Tue, 08 Oct 96 12:12:46 GMT From: "Wilkins, Keith" To: Chrissalo Subject: RE: Hi! Date: Tue, 08 Oct 96 12:14:00 GMT Message-Id: <325A453E@parodius> Encoding: 106 TEXT X-Mailer: Microsoft Mail V3.0