Page 1 of 1
Posted: Fri Apr 16, 2004 8:47 am
by braincell
I followed a tip and removied ACPI. I was warned that I would have to install my drivers again but even after I did this my ethernet card and cw cards would not work. Don't try this unless you have an image of your C: drive. This was the first time I tried Microsoft system restore and it failed. I had to install Windows XP over again. Now I am waiting for the new SFP 4.0 before I go further.
Posted: Fri Apr 16, 2004 8:58 am
by Counterparts
Hi Braincell
Doing this post-OS install can sometimes lead to system instabilities ... it's best to do this when you're installing the OS.
If you have a modern MOBO which supports APIC and you're installing XP, then it's best to have ACPI enabled, rather than installing the OS as a 'Standard PC'.
Royston
Posted: Fri Apr 16, 2004 9:15 am
by hubird
isn't ACPI standard?

Posted: Fri Apr 16, 2004 1:38 pm
by braincell
Thanks, and what about hyperthreading? I'm having some stability issues but I can try some other things such as switching the midi from serial to usb (someone mentioned that could help) I have both as an option on my interface. I can live with Cubase bombing once a day but once an hour is too much.
Posted: Mon Apr 19, 2004 4:02 am
by Counterparts
braincell wrote:
Thanks, and what about hyperthreading?
Haven't tried it myself and do not know the ins-and-outs of using HT with [insert your system configuration here]. It has been covered in a few threads in this forum - from what I've read, it works for some and others have a nightmare with it. Turn it off and see!
I'm having some stability issues but I can try some other things such as switching the midi from serial to usb (someone mentioned that could help) I have both as an option on my interface. I can live with Cubase bombing once a day but once an hour is too much.
I only use the MIDI interface on the Pulsar card, so can't comment here, really. USB-1 isn't too intensive, so that's proabably OK for MIDI. I'd disable USB-2 in the BIOS if you have no need of this interface - as I've mentioned before, it generates the Mother of all Interrupt Shit Storms. I don't think USB-2 has been properly ironed-out yet either.
Royston
Posted: Tue Apr 20, 2004 6:35 am
by braincell
I could try turning off USB 2. I might need it for something in the future but probably not all the time.
Posted: Wed Apr 21, 2004 3:36 am
by King of Snake
For some reason I need to install as "standard PC" otherwise I get issues with midi recording too early in Cubase 5.
It doesn't seem to affect my system negatively otherwise.
Posted: Wed Apr 21, 2004 12:53 pm
by Sunshine
IMO, which istallation method (ACPI or Standard PC) should be chosen depends on what kind of configuration you have. Problems whith PCI based cards did very often in the past have their cause in an unlucky arrangement of cards. To assign those problem cards the right IRQ has most often been the key to better performances. However an interrupt request does only occur when data is being requested by the CPU. But depending on the situation such requests might interfere whith other running processes, whith each interrupt request facing a different situation from...... different "background processes", through general routines of the "OS" itself, up to the users unforeseeable and unpredictable steering of the pc.
The APIC controller, which is the Interrupt Controller, recieves its information from the PIC (programmable Interrupt controller). There are two PICs availible, which give us a total of 16 IRQs (8 each). In case ACPI is activated, IRQ9 is used as well and no longer free. But on the other hand you'll get 8 additional IRQs! With modern PCs you have USB Controllers and other MB conponents that also require and IRQ. So at the end no IRQ is left free.
Whith the APIC Controller enabled...
During the boot-process the Bios transmits the routines to the OS. When an IRQ request is being send to the I/O APIC controller, the controller has access to a certain table which contains all the requests. The OS is capable of assigning all of those requests an urgency level. So according to how the OS estimates the urgancy of those IRQ requests, they will be executed. The advantage of APIC is, while the request is being send to the APIC controller, the next request is being worked on, without needing any request confirmation from the CPU. So theoretically this should lead to a faster execution of those requests and a higher performace.
Whithout the APIC Controller
The soundcard itself does send and IR request to the PIC Controller, but with the PIC Controller not knowing wich device is sending the IR request. In order to send the IRQ request to the CPU, an ISR (Interrupt Service Routine) has to be started. But the problem is... when various IRQ request are being send at the same time the PIC has to priorities which of them are being send first.
The advantage of Standard PC is, you can choose by yourself which IRQ your preferred devices should use (by assigning them the right number). Those IRQs always differed in bahaviour. There seemed to be strong IRQs and weaker ones. IRQ 5,10 and 11 have always performed better imo than all the other IRQs. Not too long ago ie. 2 Scopes only achieved their maximun dsp-efficiancy when they were set to IRQ 5 and IRQ 10, whith the premise of being in standard pc mode....
Anyway in both cases I think it's advisable to deactivate as many components as possible. With the deactivation of not needed peripherals further IRQs can be gained. Therefore you need to go to your bios and deactivate those unneeded components.
One other thing that I've encountered, is when you crank up the monitors you can even hear interactions between those components (hd noise, some strange klicking and other strange noises). This raises the question of which of those IRQs do really work independantly. According to my findings.....if the soundcards are set to IRQ 5 and 10 or IRQ 5 and 11 no so such strange sounds can be heard. So personally I'm of the opinion that those IRQs can change the sonics of your mix. However there's no proof of my findings and still I keep looking for a different technical explanation....
Anyway, the overall point is... as long as you don't get any negative effects with either ACPI or STANDARD PC, none of those modes should be written off too soon, just because there's the other way. Still there are people who do benefit from having the choice to decide. Personally I'm of the opinion that things run more fluently when being in standard PC mode. Nonetheless, I try to stay completely neutral, so no offence all you ACPI dudes -))
and regards,
to all you Creamware warriors -))
<font size=-1>[ This Message was edited by: Sunshine on 2004-04-22 18:08 ]</font>
Posted: Wed Apr 21, 2004 1:00 pm
by valis
nice post

Posted: Wed Apr 21, 2004 1:53 pm
by arela
thanks Sunshine, a beam at the point
Posted: Wed Apr 21, 2004 2:26 pm
by wsippel
Maybe it's a stupid question, but does WindowsXP support true APIC (not ACPI)? And what about ACPI 2, my Tyan Manual still suggests to switch off ACPI 2 in BIOS when you use WindowsXP?
Posted: Wed Apr 21, 2004 4:52 pm
by valis
I beleive that WinXP does support APIC 2.0 for IRQ & Power management and hence doesn't stack everything on IRQ 9 or 11 anymore although it still does grab one of them for the ACPI system itself.
As for ACPI 2.0 and Tyan's support, Tyan has been known to pioneer with their motherboards which means there are often 3-4 revisions and its possible there was some errata in v1 or v2 of your motherboard that caused problems. Otherwise I'm not really sure and probably can't answer this properly
