The solution to the kick-ass computer problem

althoff wrote on 8/28/2001, 1:03 PM
OK, problem solved. After testing the same Vegas project on a few different computers, we arrived at this conclusion:

Intel P4 is NOT suitable for Vegas!

The reason is not quite clear, someone suggested that the FPU in a P4 operates slower than an Athlon or a PIII. Also, bit-shifting (common in division and multiplication operations) are much slower on a P4 for some reason.

Here are the test results:

P4-1500 with W2K Server and SCSI system: Problems.
P4-1800 with W2K Pro and IDE drives: Same problems.
2xPIII-1000 with WinXP: No problemo.
AMD Athlon 1000 MHz, still no problemo.

See the pattern? =)

We still have further testing to evaluate, sound card brand and model is one of them. Both P4's have SB Live!, the 2xPIII had an SB PCI128. The AMD still has a Live!, though, and it sounds pretty far-fetched that this should be the only reason.

Anyway, is there anything the Vegas dev team can do about this? I'm not really sure what I'm asking for so I don't really expect an answer, but I still feel like I have been cheated by someone. Probably Intel. Bastahds.

Comments

jboy wrote on 8/28/2001, 2:00 PM
really weird. anybody out there have a well functioning P4 configuration going ??
SonyEPM wrote on 8/28/2001, 2:23 PM
what are the problems?
steveh wrote on 8/28/2001, 3:20 PM
See this post:
http://www.sonicfoundry.com/forums/ShowMessage.asp?MessageID=75453
MacMoney wrote on 8/28/2001, 10:08 PM
I have a P4 1.7Ghz
GeForce MXX400 Video w/64Mb Ram
Intel D850GB 2Ghz Motherboard
1024mb PC800 ram
Adaptec 29160 PCI Ultra 160 SCSI card
2ea Seagate Cheetah 18gb Ultra 160 HDs @15000rpm (recording) Fat 32
1ea Seagate 9gb IDE HD @7200rpm (OS) NTFS
Win2k Pro spk2
MOTU Win2k drivers 1.0
Vegas Video 2.0g 48 trks 7 plugins (as of today)
Acid 3.0b 52 mixed audio tracks, no midi tracks, 11 Waves Rverbs (stress test)
And It Keeps Getting Beter Everytime I Use IT!
Im in love with this set up!
Im checking everyday for WDM drivers for Acid, Vegas and Sound Forge.
George Ware
HPV wrote on 8/29/2001, 10:42 AM
>but I still feel like I have been cheated by someone. Probably Intel. Bastahds.

I can stack 20 audio tracks and playback with about 35% cpu useage. All tracks have the three stock filters. Played back for about two minutes with no glitches. That's with VM also.
I have a P4 1.3 with 128MB and ME. I don't think it is a P4 problem.

Craig H.
althoff wrote on 8/30/2001, 11:47 AM
The problems are unreasonable performance drops on a very high-level computer.

Now it looks like it's a plugin problem. I've tried disabling the plugins one by one, and when nothing changed I tried removing them from the chain, reloading the project and remove the next one.

Lo and behold! I found the villain! His name is...

Waves supertap 2-tap mod.

As soon as I have this plugin somewhere in my chain, CPU usage rockets smack-zabaam right up to 100%, graphics updates in vegas runs at about 0,3 fps and the computer response time is something like 2 or three seconds. Once I remove it everything drops back to normal.

This would seem inexplainable, but my hardware wizard friend still claims it has something to do with floating-point processes and the P4's lack of abitilty. I still don't give a crap as I honestly think that a state-of-the-art CPU with kickass gear shouldn't have performance drops no matter what.

Anyone else been experiencing problems with Waves plugs?
Ben  wrote on 9/2/2001, 2:43 PM
I've also had some serious gapping problems which I, inititally at least, put down to Waves plugs. Was running the Waves Super-2 tap delay and getting serious stuttering. The strange thing is, checking it's CPU usage in Sound Forge, it seems pretty low. I thought it could be specific to that plug-in, but needed a realtime delay, so downloaded the freeware dB-T tempo delay plug-in. Bizarely enough, still get the same problem.

It seems that for some reason delay plugins are causing gapping in Vegas (for me at least). Can run a load of other plugins - including reverb - without any real problems. What might be of interest to you though, Johan, is that I'm not running a Pentium 4 but a Pentium 3 850. Pretty standard hardware with M-Audio Audiophile sound card.

Any ideas anyone?
SonyEPM wrote on 9/4/2001, 9:08 AM
Welcome to the world of P4. We are working on some p4 optimization for our apps and have been in contact with Waves about the same issue. It'll happen, but it is a huge job and will not happen instantly. Kind of like Win16 to Win32- major rewrite of many components.
althoff wrote on 9/4/2001, 7:43 PM
Thanks for replying, Dave. I have all the confidence in you guys!

Interesting that I didn't get any stuttering on the twin PIII computer, though... Maybe all it takes is a little CPU overkill? =)
xgenei wrote on 1/7/2002, 2:50 PM
Definitely concur with the other guy about it not being a P4 problem -- remember the P4 is backwards compatible and there are untold thousands of programs running on it. The simplest explanation is a design bug in the plugin.
I have seen the same kind of problem persist for years in an otherwise successful product (OnTrack PowerDesk) and no help or even an explanation was ever forthcoming. I still use the product constantly (it's a windows explorer replacement) I just know that when my computer slows to a crawl, I have three mouse-clicks to go before the system locks up. I shut down the offending window (it's always a specific iteration of the program), wait a second for the memory to clear, and then work as if nothing had happened.
Restarting PD returns last position -- so no big loss -- but it sometimes happens 2-3 times an hour.
I tried special memory management (RAMPAGE for Win98 SE), tweaking all the PD settings, turning everything automatic off, upgrading to new releases -- still there.

The answer is to know it, replace the offending software if you can, or work around it as I have been able to. It never got worse -- it never got better. But it doesn't crash my system anymore.