home *** CD-ROM | disk | FTP | other *** search
- > > When I first got my Falcon, I ran several tests which didn't show anything
- > > close to that speed. A program I found on my hard drive today claims to
- > > reach 6.0 Mbyte/s (read) in 320x240x2 (monochrome) mode and 3.8 Mbyte/s in
- > > 640x480x256 or 320x240xTC, which is close to what I remembered.
- > > (This is all on a VGA monitor.)
- >
- > I've tested it too, with a program called MEMSPEED.TTP. And in 640x480 2
- > cols, it says 8 MB/s.
-
- That's the program I used yesterday and it said 6.0 read, 6.9 write, IIRC.
- Really strange!
-
- Could we be using different versions of the program?
- I'll check what date and size mine has and email you with the details.
-
- > > Still, they're talking about an accelerator bord that manages to copy
- > > 4 Mbyte/s to ChipRAM (256 colour mode, lowres) and that is apparently very
- > > slow for such as board.
- >
- > They're partly right. But a guy I'm working with at uni, tells me that
- > the normal figure for ChipRAM is about 5 MB/s.
-
- But this was in the 256 colour mode that they're going to be using for
- BAD MOOD. The Falcon don't do very well in its equivalent TC mode.
-
- Anyway, they'll hopefully soon be flocking here and then we can ask the
- c2p folks directly.
-
- > > >From this and some other data, I'd say that the Amiga can write at least
- > > 6 Mbyte/s to ChipRAM.
- > > At least as much as the Falcon in other words.
- >
- > Then how come that the 256 640x480 mode is useless on the Amiga1200 ie.
- > slow ?
-
- I'm sure that our maximum _total_ bandwidth is quite a bit higher than theirs.
- The processor isn't hurt as much by higher graphics resolutions, but the 16
- bit bus makes for low CPU<->RAM bandwidth in the first case.
-
- Perhaps there really is a reason for the 16 bit processor interface?
-
- > Anyway, there's a 3D graphic card on the way for the Amiga (with
- > Zorro-bus), which should indeed be capeable enough.
-
- I've been dreaming about having one of those new 3D PCI cards on the AB040's
- local bus...
-
- > > The lack of a 256 byte data cache shouldn't matter much for BAD MOOD, but
- > > in general the '030 is slightly faster as you say.
- >
- > Is that because the 68030 does all the texture-mapping ?
- >
- > Hence too much data to be within the cache ?
-
- Too much and too spread out. There isn't much repeated accessing in time for
- the cache to still have the data around.
- Almost everything but the texure data is kept in registers all the time.
-
- > > I think it's quite reasonable to interpret 'similar' as 'within 30% or so'.
- >
- > Yeah ok, maybe you're right :-)
-
- We can't go and make them upset this early, now can we? ;-)
-
- --
- Chalmers University | Why are these | e-mail: rand@cd.chalmers.se
- of Technology | .signatures | johan@rand.thn.htu.se
- | so hard to do | WWW/ftp: rand.thn.htu.se
- Gothenburg, Sweden | well? | (MGIFv5, QLem, BAD MOOD)
-
-