1  |   Order by index

Jul 7, 2001, 4:56am

Okay, so I forgot to bring these up at techtalk. There are some pretty major issues that definitely need to be addressed

1) Collision detection, in particular, climbing logic. Either it lags the holy livin shit out of me with a ton of number crunching
and the higher priority processes drown it out or it's just slow by nature. Takes about 30 seconds to climb a staircase... So, I fly
a lot.

2) Can we get a Linux client in here? OGL only of course, until some wank takes the time and ports the D3d API (And is promptly
murdered for bringing M$ software into the Linux world) I can't see anything that would truly stop it, with the exception of RW
using Windows-specific calls or maybe some Windows DLLs that I don't know about AW using.

3) In the help files, midis can be looped. In practice, THEY CAN'T (not with the documented code anyway)

4) Create texture (and probably color too) to a named object doesn't work, help files suggest that it can. In reality, it appears
that a complex astart/adone command is required.

5) Probably a dozen other things that I can't think of right now, I'll post the rest as I think of them.

BTW: Regarding the bug in #1, my system specs are
AMD Duron 700 running stable and ice cold at 1200MHz
256mb PC150 at 100
SB Live Value
RAID hard drives (In other words, I assure you, the hardware is not causing the slowdown with the possible exception of el crappo
cache on CPU)

I've tested it on the following video systems, all are 32mb unless noted
Quadro II MXR (Same chip as GeForce II MX with slighly different voltage) solo
Quadro II MXR with Jaton Trident PCI secondar
GeForce II MX solo
GeForce II MX with Jaton Trident PCI
GeForce II GTS 64mb
TNT2 m64 16mb
Jaton Trident PCI (no 3d) running in AW3.1's hardware emulation
TNT2 Ultra
TNT2 Vanta
TNT2 Vanta LT 16mb

Okay, so I'm an nVidia whore, so sue me :)

I tested under the following software conditions:
Normal processes (including Genome crunching)
Sterile (Nothing but Winderz and AW)

ALL configurations have yeilded the same results. I also have another system that I seem to recall having the same problem, but it's
IDE controllers are dead so I'll check that out when I get a replacement board for it.

Jul 7, 2001, 4:38pm

[View Quote] Roland claimed when he released 3.0 that this was as fast as he could get it
to go, i get the same problem...


|
| 2) Can we get a Linux client in here? OGL only of course, until some wank
takes the time and ports the D3d API (And is promptly
| murdered for bringing M$ software into the Linux world) I can't see
anything that would truly stop it, with the exception of RW
| using Windows-specific calls or maybe some Windows DLLs that I don't know
about AW using.

AW 3.2 will support OGL and Software rendering


|
| 3) In the help files, midis can be looped. In practice, THEY CAN'T (not
with the documented code anyway)

AW always crashes my computer when it plays midis, plus the fact they sound
bad, so i dont listen to them...
AW 3.2 will go back to using windows media player to play midis, as opposed
to it using DirectSound like it is now, maybe this will fix it, but i dont
know

|
| 4) Create texture (and probably color too) to a named object doesn't work,
help files suggest that it can. In reality, it appears
| that a complex astart/adone command is required.

it is only applying the texture and color when the object that contains the
action is loaded, when the objects with the names are loaded they are not
getting it because there is nothing triggering it, the adone action works
only because it is constantly re-applying the textures or colors


|
| 5) Probably a dozen other things that I can't think of right now, I'll
post the rest as I think of them.
|
| BTW: Regarding the bug in #1, my system specs are
| AMD Duron 700 running stable and ice cold at 1200MHz
| 256mb PC150 at 100
| SB Live Value
| RAID hard drives (In other words, I assure you, the hardware is not
causing the slowdown with the possible exception of el crappo
| cache on CPU)
|
| I've tested it on the following video systems, all are 32mb unless noted
| Quadro II MXR (Same chip as GeForce II MX with slighly different voltage)
solo
| Quadro II MXR with Jaton Trident PCI secondar
| GeForce II MX solo
| GeForce II MX with Jaton Trident PCI
| GeForce II GTS 64mb
| TNT2 m64 16mb
| Jaton Trident PCI (no 3d) running in AW3.1's hardware emulation
| TNT2 Ultra
| TNT2 Vanta
| TNT2 Vanta LT 16mb
|
| Okay, so I'm an nVidia whore, so sue me :)
|
| I tested under the following software conditions:
| Normal processes (including Genome crunching)
| Sterile (Nothing but Winderz and AW)
|
| ALL configurations have yeilded the same results. I also have another
system that I seem to recall having the same problem, but it's
| IDE controllers are dead so I'll check that out when I get a replacement
board for it.
|
|

Jul 7, 2001, 5:07pm

[View Quote] Time to go back to 2.2 collision detection then, cuz this ain't makin the grade. Flat=BORING.
> |
> | 3) In the help files, midis can be looped. In practice, THEY CAN'T (not
> with the documented code anyway)
>
> AW always crashes my computer when it plays midis, plus the fact they sound
> bad, so i dont listen to them...
> AW 3.2 will go back to using windows media player to play midis, as opposed
> to it using DirectSound like it is now, maybe this will fix it, but i dont
> know
>
> |
> | 4) Create texture (and probably color too) to a named object doesn't work,
> help files suggest that it can. In reality, it appears
> | that a complex astart/adone command is required.
>
> it is only applying the texture and color when the object that contains the
> action is loaded, when the objects with the names are loaded they are not
> getting it because there is nothing triggering it, the adone action works
> only because it is constantly re-applying the textures or colors
>

These two, I was mostly concerned about the misinformation that the helpfiles produced.


Yet another one that bothers me, though it doesn't deal with the software.

On Alpha's object path, the sprites all seem to be in really badly compressed JPEGs. This causes the mask to make any transparency
behind it completely transparent and the texture not to extend to the edge of the mask, therefore giving a really ugly visual
effect.

I assume that the glass over glass triangle crap (for lack of a better term) was discussed before I arrived, because there was a
rather stunning example of it right in front of Roland's avatar.

Jul 8, 2001, 10:10pm

[View Quote] > I assume that the glass over glass triangle crap (for lack of a better
term) was discussed before I arrived, because there was a
> rather stunning example of it right in front of Roland's avatar.
>
>

1  |   Order by index