>So your telling me that a video card with 2 megs of ram or whatever is
>getting 2 megs of data per clock cycle, or maybe around 90 times a second
by
>using windows?
No, I'm telling you that the amount of data being sent to the card by the
driver is slightly too much for the input buffer on the card. If a card has
2Mb of RAM it doesn't automatically follow that all of this is the input
buffer. The monitor can hold up the card, but it is not fair to blame the
monitor because it is the drivers which are at fault. There is no other
logical explanation, especially when this stuttering can be observed on a P3
600MHz system. You can't seriously sit there and tell me that a system with
that much CPU power is struggling to decode the MP3! My hard drive occupies
8% of CPU power, and the CD-ROM occupies 18% (44x system - IDE) but as I am
now plwying the file out of a RAM drive anyway, that's now irrelebant. No
ISA cards BTW - I just pulled them out.
>The monitor does not accept frames, they accept lines. IE
>refresh rates are how many times the monitor redraws a line per second,
>starting from top down.
LOL!!!!
If your monitor with a refresh rate of 80Hz (check the monitor manual, it
may be less) has as many lines down as mine does, it will only have drawn
the top portion in one second! The refresh rate (try looking under VSync!)
is the speed at which all the lines can be rendered, and the flyback system
reset ready to start another frame. 80 lines per second indeed!
> MP3 decoding can take up MASSIVE ammounts of CPU power, older
versions
>of winamp took up more than 30%. Newer versions around around 12-15%.
Hmmm..... 30% (WinAmp) + 40% (hard drive under stress - an MP3 at
5Mbytes/sec?!?!?) = 70%
That leaves 30% free to run the sound card and graphics card - more than
ample. So why stutter unless..... wait for it.... the graphics card drivers
are crap!
> In Half Life, I had stuttering noises, this was cured by A.
Lowering
>resolution, why? Not because too much data was going to the video card and
>it had to wait to fill buffers, but because too much info was going to the
>CPU! That was on my P200, now with my P2-450, I can run the same res, with
>the same vid card, same drivers, same OS...and NO STUTTERING!!!
That is slightly different because Half-Life is a very processor intensive
program. An MP3 can be decoded by a 486 in DOS - I don't think the same
could be said of running Half-Life. I have also seen Half-Life, Quake 2 and
Carmageddon stutter on a P2-450 because the graphics card was at fault -
replacing the drivers fixed it and all games now run very smoothly.
> Have you ever looked at the bus scematic for BX chipsets? PCI bus is
>on a different bus, the AGP is pretty much on the PCI bus. ISA is on a
>different one. Have you ever looked at benchmarks across CPU's?
Have you ever designed a CPU? Sorry, perhaps that was unfair. Ok, it was
only Z80 equivelent, but it worked! (Pity I never got to build it on
anything other than software).
If card A sends a time-critical request to the CPU, but the CPU can't do
anything because it is waiting for card B to say "finished" and release the
interrupt disable flag, what happens? Card A will miss out on data. It's
really a simple concept, I don't see why you're having such trouble with it.
>Does the
>graphics card scale to the cpu? Or does the CPU scale to the graphics card?
That depends on whether the drivers tell the CPU to ignore interrupts from
other parts of the system, effectively locking up the CPU. A *good* driver
will ont do this.
>Let me ask you this. If you have a P150 and a GeForce256, will the
computer
>lock up?
That depends what you are doing.
> HOW CAN IT!?!
Quite easily if you decide to run somelike like Alien vs Predator on it!
>***************The CPU has to send so much data to the graphics card, that
>it has no time to get data to the sound card, hence the
>stuttering*************.
In Windows at 640x480 resolution? Hmmm..... the 480 doesn't seem to bothered
about this with a Trident ISA card. I guess the GeForce256 must be crap
then. On the other hand, it might be that the damn thing just needs
replacement drivers......
> If you look at benchmarks, Framerates scale to the power of the CPU up
to
>a point. That point is the polygon or triangle output of a video card, a
>CPU finally is powerfull enough to send enough data to overwhelm a video
>card...BUT it does not start stuttering it only tops out at frame rates. I
>would say the GEForce256 will top out with a K7-800 or so.
Which is completely irrelevant because we're talking about playing MP3s in
Windows, not watching 3D graphical representations of audio renderes using
as many mip-mapped texels as you can fit into 32Mb!
> Why can a 386 run Unix as a server? I have one running a small file
>server :). Because it is not runing a GUI AND it is not running any
>background stuff.
Yeah. No GUI - no drivers needed. Eliminates the whole problem!
>Why can ur dos machine decode MP3's? Because its not
>running windows wich takes up 50% of system, resources.
Which doesn't explain why it wont do it in 640x480x256 mode..... oh hang on,
it does if I install VESA2 drivers for DOS.... I suppose those VESA2 drivers
must have a CPU secretly stored somewhere I hadn't noticed before because
the processor must now be able to work faster.
> Shawn
>PS. Who the hell has a 32k video card? Considering vid cards have upwards
>of 64 megs of ram now. What fills that much ram? Not windows or winamp
>decoding!?!
You missed the point Shawn. 32k was the buffer size, not the RAM. Anyone
with a PCI Guillimot video card (including the Voodoo Banshee cards) - they
all have 32k buffers, as do quite a few others I am told.
Don't you think this is maybe a little OT for this list by now? Dont you
think we ought to move this debate to private mail before the MD fans all
run for the hills screaming? ;o)
Magic -> Now having read more graphics card specs than he really wanted to.
--
"Creativity is more a birthright than an acquisition, and the power of sound
is wisdom and understanding applied to the power of vibration."
Location : Portsmouth, England, UK
Homepage : http://www.mattnet.freeserve.co.uk (under construction)
EMail : [EMAIL PROTECTED]
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
