YATR (Yet Another Test Request)

[fzammetti]

If anyone can test this out for me on an iPaq 36xx series device, I'd be very appreciative:



Note that this new version (although it doesn't say it on the menu screen!) pops up about seven windows before the menu appears giving various information. If you could relay that information to me it might be helpful if there is a problem.

I've tested it successfully on the following devices (Best Buys must hate me, I was there for about 30 minutes beaming it and testing it on various devices, told them I was deciding on my next device and wanted to compare speed with a benchmark program!):

Casio E-125
HP 568
Toshiba e740

So, I know that it BASICALLY works (meaning probably not perfectly, but very close). In fact, the only defect I noticed on any of them (on the 568 and 740) was on the Max FPS Test, it looks like the walls that shift to the left don't show up. I'll have to figure that out at some point, but I consider it minor for the moment.

As an interesting side-note, the plasma test was considerably slower on the e740 by about 10 FPS than even my stock E-125! We already knew XScale had some performance issues, here's another example!

I know it does not work on the 38xx series iPaq's, I gotta go look at the code specifically for that and see what I screwed up. But I'm hoping it works on 36xx's now (of course, this is about the fifth version, I'm fighting depression here!).

Thanks in advance!

[Jasper]

Your probably not going to like this, but on my iPaq 3660 the

following are the results:

1: GX_isBuffered = False
2: cbxPitch = 640
2: cbyPitch = -2
5: Final cbxPitch = 640
6: Final cbyPitch = -2
7: Buffer Size = 204800
8: Allocating Buffer

Invoking each of the menu options produces no
effect, until the stylus is then touched onto the screen again after selecting a menu option. A display is then seen overlayed on the menu (there appears to be no screen clear!), only for the following items. No others produce any output.

Lines
---------------------
Lots of coloured lines on screen.

Polygons (filled)
---------------------
Lots of coloured filled triangles.

Region functions
------------------------
A black square about 100x80 pixels in the bottom half

of the display.

Sprite manipulations
-------------------------------
Four face in black squares in the top third of the
display and a texture about 100x100 pixels in the
middle of the display.

The Exit function works ok.

I have reset my PPC a couple of times and unloaded
all other apps running but still does the same thing.

Maybe someone else with a 36xx could confirm the
same problems, before you get too depressed.

Let me know if you need further information.

[fzammetti]

I just put v5 up, I actually think I might have it this time (yeah, I know, I say that EVERY time!).

Same url:



The reason I feel kinda confident this time around is that I completely rewrote the buffer-flipping routine (I blit it now rather than a straight memcpy, got the idea from a random newsgroup post). I flipped some switches so that I would be using double-buffering in the emulator (usually it's doing it itself I think because GXIsDisplayDRAMBuffer returns true in the emulator) and it worked flawlessly.

If anyone tries this, if you could try and IM me (fzammett), I'm really anxious to know how it goes...