| From: | Sam Jordan |
| Date: | 13 Dec 99 at 20:14:48 |
| Subject: | Re: [warpup] several questions about ... |
Am 13-Dez-99 schrieb Mr.Return:
Hi
>As mentioned above I have a few questions :-)
>First: A new ppclibemu is out, but did anybody make it to run anything on it
>? No matter if it was M.a.m.e, ShowPPC, Frogger, FastQuake5 (Yes, I own the
>Clickboom-original !) or whatever, every program crashes instantly or
>refuses to work. I think this is because the ppclibemu does not support
>async 68k-task which are an advantageof the ppc.library (They are often
>used for async io or things like copying the screen as M.a.m.e and Quake
Which doesn't give any advantage over the synchronous approach, but
complicates software development significantly. I have never heard about
asynchronously working software running quicker than other ones. This
async-stuff was born at a time, when synchronously working programs
were impossible to do using PowerUp (because of the monster switch delay),
hence Phase5 pushed the async programming a lot by also suggesting, that
this would result into better performance due to CPU parallelism (which
I never believed in and I never say any example proving this theory).
>Second: About Warp3D and Permedia 2: Is there any speed-difference between
>15/16 bit, dithering on/off ... (Yes, I�m talking about Wipeout2097) ? Does
15/16 bit has no effect, because 15 bit simply implies that one bit
is unused resp. used for transparency. Dithering has hardly any effect.
>the vertical screen-refresh (e.g. 80 or 100hz) affect the P2�s speed ? And
When using double buffering it has an effect, otherwise not.
>what about Mipmapping ?Finally: What happens if there is not enough
Mipmapping has no effect for P2-based boards, because the P2 doesn't
support mipmapping.
>texture-memory, does the program (as I know) simply not run or does Warp3D
>try to offer s.th. like "virtual memory" by copying textures into the
>gfx-cards memory if they are needed ?
In theory Warp3D simply uploads the textures whenever they are needed,
so the game *should* continue to work, but veeeeeeery slow. In practice
the situation is a lot worse, it might happen, that corrupt textures
are displayed.
>Last but not least some words about POP and G4:
>I think that H&P should offer WarpUP and Warp3D for the P5-G4, even if they
>do not licence it. I think so because otherwise they would simply loose
Of course users want this to happen. But you should realize that commercial
companies usually don't work for free or for negative results.
>customers, yes, there will be piracy, but if owners of these cards do not
>have the choice to use WarpUP/3D, they will be lost for all forthcoming
You should realize that even in the case that all those software components
would be written, I don't expect any accelerated 3d application to run
before autumn/winter 2000, including all delays with hardware and all
delays with software development. Quite a long time to wait and during
this time a lot can happen.
>WarpUp/3D-software as well. And if yyyou think that these people are
>willing to pay about 1000$ or more for their Amiga you can imagine that
>they will want to buy some software, too. P5 has a minimum of 1000
A lot of people will say that they have already paid enough for the
hardware, so that they have 'the right to get the software for free,
because otherwise nothing serious would work' or something similar.
I remember that there was another issue in the past, where a huge
amount of people had the moral right to do something illegal.
The enduser price of the operating software would have to be set to a quite
high value to compensate for piracy, and this would again offend people
and maybe convince them not to purchase the software ('...only this one
time...').
>preorders (I heard more) and that is a minimum of 10% of the existing
>PPC-hardware on Amiga. H&P should not give them away. Ok, some people say
If non-PPC owners have preordered such a card, then I can understand it
somehow, but I will never understand how people *with* PPC card enter
such a risk at a time, when their existing PPC board is quite well-
supported, even more as there is meanwhile quite good support for
3d acceleration and applications becoming available more and more.
>claim that P5 is always very late but despite of everything people say:
>They gave us much more than only hot air.
And they are right now going away from AOS. In the past they at least
delivered AOS drivers for all their hardware components and that's
most likely over now. And we know that driver development can take
*a lot* of time, especially if some third party people have to do
the work.
bye