T.R | Title | User | Personal Name | Date | Lines |
---|
1250.1 | | WRKSYS::CHALTAS | Someone chanted evening | Mon Feb 24 1997 09:00 | 2 |
| If set to 1, the variable disables all optimizations,
and may result in dramatically slower performance.
|
1250.2 | Please report problems! | WRKSYS::DOTY | Russ Doty, Graphics and Multimedia | Mon Feb 24 1997 10:27 | 27 |
| Please file problem reports against these crashes (and any other
customer problems)!!! An open note to everyone reading this conference
-- we want to know about problems, so that we can fix them. This means
that we need enough information to reproduce the problem in our lab
(unfortunately, "the application crashes frequently during the spinning
phase of a solid model" isn't quite enough for us to reproduce, debug,
and fix the problem). Formal problem reports can be a bit of a
nuisance, but they are the only way to ensure that a problem gets
tracked and fixed.
The black triangles sounds like a triangle clipping bug we have seen,
which is fixed in the next release.
Having said that, try the GSS 4.1 drivers (available at
http://www.service.digital.com/open3d). They contain a number of bug
fixes and may help.
As George mentioned, changing the "solpath" variable can (should!)
dramatically affect performance.
One thing that we have seen in porting pure Intel apps to Alpha is some
differences in handling of IEEE floating point exceptions -- our
porting groups are familiar with the issues.
Thanks,
Russell Doty
PowerStorm 4DT Product Mangler
|
1250.3 | Sounds like a BUG | BGSDEV::COMEFORD | There's coffee in that Nebula ! | Tue Feb 25 1997 16:43 | 22 |
| setting IG_USESOLPATH to 0 should DRAMATICALLY affect performance
UNLESS both the C code and the assembly/optimized are able
to keep the board fully busy (I know of no
paths where this is true, so it SHOULDn't happen).
You said "The application crashes frequently during..."
By crashes what do you mean? Warnings from
Dr. Watson/windbg? System crashes (AKA Blue Screen)?
Incorrect pictures?
It sounds like you have a bug. As stated there is
a fix for something similar to what you are
seeing in the V4.1 version of the driver.
Please try that out first. If that doesn't fix
it please report this as a problem. And report it means
we'll need the application, if we can't reproduce
the problem we most certainly can't fix it...
Thanks,
Keith Comeford
PowerStorm 4d40t/50t/60t project lead
|
1250.4 | Yes, i will send a reproducer | NETRIX::"[email protected]" | alberto galli | Thu Feb 27 1997 12:43 | 10 |
| Yes,
i'm try to arrange the fastest way to send a reproducer and of course
i will open a bug as soon as i receive the reproducer itself (is not
the first time, will be not the last).
Thanks
Alberto
[Posted by WWW Notes gateway]
|