ASUS A8V Deluxe/X2: The PCI Slot Position Doesn't Matter!

[Suntower]

Hi,

Ever since I posted that my X2 system (basically copied from the ADK spec UA used to test 4.0 software) I get about 1 PM a week like so...

Hello,
as you have more or less exactly the same setup as I'm going to build I would like to ask you some questions.

My setup:
Asus A8v
Matrox p750 (three screens)
2*UAD
RME9652

Which pci-slots did you put the uad's in and which one did you take for the 9652? What pci latency did you use for the uad's and for the rme? Do you use powerdesk?

Thank you very much for your help!

So I'm posting this -here- hopefully to save people having to PM me for the same bit of info. OK, here's the answer:

It doesn't seem to make a difference.


That said, my layout is:
1. RME
2. PoCo
3 UAD-1
4. UAD-1
5. <empty>

If this differs from what I wrote you individually, that's only because I moved them around once, simply to make it easier to get my fingers around various wires (SHIT, why can't case makers design a case where you don't have to shred your fingees to ribbons running cables!)

[Epilog]
So far, after 60 days, my ONLY beef is this:
I have had a few lockups which I, at first thought were bad memory chips, but I am now convinced that it is an RME device driver problem since it almost always happens when I have Wavelab -and- SX or Samplitude open at the same time and the 'multi-client' code is working. I have e-mailed an event log entry to RME and am awaiting reply.

Other than that? Works like a dream.

<HTH>

---JC

 

[Kontrakatze]

This was my pm and my apologies if I stepped on your nerves.

About the topic:
If you are right, this seems to be one of the rare motherboards where it does not matter. By the time now I had several motherboards (a7v 266, a7v333, just to name two) and it always mattered. Pops, crackles and performance degration were only the least severe problems I encountered, blue screens the bigger ones.

Ok, most of the time it is hard to distinguish what the problem is related to, but almost all the time swapping cards solved the problem.

 

[Akis]

JC, when you say lock-ups, you mean complete system freezes? If yes, I'd say it's your Powercore: I found out that apart from the PCI mkII, the PCI Classic causes system freezes, as well.

BTW, yes, all slots seem to work fine with UAD-1.

 

[Suntower]

This was my pm and my apologies if I stepped on your nerves.

Not a bit of it. I've just answered the same question 6 times now, so I thought it might warrant general display.

About the topic:
If you are right, this seems to be one of the rare motherboards where it does not matter. By the time now I had several motherboards (a7v 266, a7v333, just to name two) and it always mattered. Pops, crackles and performance degration were only the least severe problems I encountered, blue screens the bigger ones.

I hear ya. Maybe the PC world is getting easier. This was the first PC I had built in a couple of years and I was -amazed- at how much easier it had gotten. Everything pretty much worked right out of the box which was, in itself, a shocker.

Cheers,

---JC

 

[Vocalpoint]

Suntower,

The only thing I can add is that I put my audio interface (Delta 1010) in Slot 4...since that one (at least in my setup) is all by it's lonesome with no IRQ sharing at all. I did move my UAD-1's (2) into a couple of different spots and it made no difference at all.

I haven't had one lockup, crash or issue of any kind. This is a killer combo.

Cheers!

 

[Suntower]

Correct. But it only seems to happen when I have WL open (so the multi-client thing is going on with the RME.)

I'm likin' the PoCo less and less as time goes by. It was great for it's time, but now that the Plate 140 and the Roland Chorus are out, the only thing I use is Tubifex ---- and even though it sounds good it is one seriously NOISY mutha. I think if Amplitube2 turns out to be good I'll probably sell it.

Thanks for the heads up.

---JC

JC, when you say lock-ups, you mean complete system freezes? If yes, I'd say it's your Powercore: I found out that apart from the PCI mkII, the PCI Classic causes system freezes, as well.

 

[Akis]

About the topic:
If you are right, this seems to be one of the rare motherboards where it does not matter. By the time now I had several motherboards (a7v 266, a7v333, just to name two) and it always mattered. Pops, crackles and performance degration were only the least severe problems I encountered, blue screens the bigger ones.

I hear ya. Maybe the PC world is getting easier. This was the first PC I had built in a couple of years and I was -amazed- at how much easier it had gotten. Everything pretty much worked right out of the box which was, in itself, a shocker.

Nah, it's just the first time UA posted a specific setup that works after extensive testing. I used to think that developers would try their products on as many mobo/CPU combos as possible, but UA and TC proved me wrong. At least they do try various combos now - I read they're testing another troublesome setup at the time, so respect to them!

Dammit, everybody should check their products in both AMD/Intel and in both nForce/VIA. :evil:

 

[Agent Cooper]

JC,
just for the sake of discussion, my personal experience is a bit different.

With the RME in slot 1 I couldn't get Thonex 2 to play at 128 buffers correctly, even 256 didn't always work without crackles.
RME in slot 4 means 128 is perfectly ok.
I didn't try slot 2 then anymore.
The UADs don't seem to matter though, I never had any problems witn them anyway, on different systems.
The PoCo will remain in slot 5 if possible for airflow reasons.

For the record (preliminary specs):
X2 4200, Zalman 7000Cu
A8V del Bios 1013
2GB RAM MDT PC3200 (2 sticks)
4 HDs 7200/8MB Maxtor & Samsung
Burners Plextor & Benq
PSU Enermax 465W
Cards (from AGP on down):
MSI TI4200/64 AGP 8x (with fan)
3* UAD-1
RME Digiface PCI card (small) in slot 4, essential !!!
PoCo Element (long PCI card)
Case : Silentmaxx SP11 big

Cheers, Cooper

 

[cAPSLOCK]

To further hijack Suntower's thread...

I have an A8V and Digiface and 2 UAD1 cards and a strange problem that started happening back when I owned an ABIT AV8 but has continued with the ASUS board.

Frequently the RME card makes the sound of it's buffers being cycled. Like it temporarily just stops getting data so it just plays over the same thing in its buffer for a short period.This will happen with ANY application and with ANY flavor of driver (ASIO, MME, or DirectX drivers all do it).

It sounds like this:
http://noisevault.com/temp/help.mp3

Now, it is not as bad as I made it sound in this example, which I was able to create by selecting regions in Soundforge with the mouse while music was playing.

It happens at both low and high buffer settings.

If it happens while I am recording something then it obviously will hose the recorded audio.

I have come to think it MIGHT be related to the chipset both these motherboards share. RME swears it is not their fault and do not seem willing to investigate the problem.

I ran into one other person in the RME forums with the exact same problem. He solved it by replacing his motherboard with an NForce4 chipset.

Obviously this may be a solution, but I have thrown so much time and money at this that I am tired of the trial/error process.

I am wondering, has ANYONE else out there heard this?

Oh, and I have tried LOTS of settings and changes including:

1. Change card slot locations... every permutation.
2. Update all drivers.
3. Downgrade all drivers.
4. Remove all plugins
5. Remove all uneeded hardware.
6. Defrag dskchk etc.
7. Replace power supply
8. Change PCI latency timings in every direction and combination imaginable.
9. Reinstall OS from scratch.
10. INSTALL ENTIRELY NEW MOTHERBOARD
11. Reinstall OS AGAIN on new system.
12. Try other hard drives
13. Try other hard drive CONTROLLERS (FOUR in total!) including 2 SATA and 2 PATA

cAPS

 

[Agent Cooper]

cAPSLOCK,
maybe stating the obvious, but this is the exact noise I'm getting when I change a DSP plugin while SX is running, obviously PDC related.
Digiface here, too, but also with other RME cards.
On my old P4 Northwood/ I 845PE system as well as on my new X2 / KT 800 rig.
I've learned to live with it, I just have to stop SX when changing plugins.
Cheers, Cooper

 

[Animix]

Hi,

Ever since I posted that my X2 system (basically copied from the ADK spec UA used to test 4.0 software) I get about 1 PM a week like so...

Hello,
as you have more or less exactly the same setup as I'm going to build I would like to ask you some questions.

My setup:
Asus A8v
Matrox p750 (three screens)
2*UAD
RME9652

Which pci-slots did you put the uad's in and which one did you take for the 9652? What pci latency did you use for the uad's and for the rme? Do you use powerdesk?

Thank you very much for your help!

So I'm posting this -here- hopefully to save people having to PM me for the same bit of info. OK, here's the answer:

It doesn't seem to make a difference.


That said, my layout is:
1. RME
2. PoCo
3 UAD-1
4. UAD-1
5. <empty>

If this differs from what I wrote you individually, that's only because I moved them around once, simply to make it easier to get my fingers around various wires (SHIT, why can't case makers design a case where you don't have to shred your fingees to ribbons running cables!)

[Epilog]
So far, after 60 days, my ONLY beef is this:
I have had a few lockups which I, at first thought were bad memory chips, but I am now convinced that it is an RME device driver problem since it almost always happens when I have Wavelab -and- SX or Samplitude open at the same time and the 'multi-client' code is working. I have e-mailed an event log entry to RME and am awaiting reply.

Other than that? Works like a dream.

<HTH>

---JC

You have your RME card in PCI #1. That PCI slot is hard wired to the AGP. Have you checked you IRQ's. I don't care how *elegant* ACPI purports to be with IRQ sharing, I have found that using a PCI card in a slot that is hard wired and shares with the AGP is trouble looking for a place to happen.

 

[dt]

My experience with the A8V is a little different so far - I can only get 75-85% out of my 2 UAD-1s without getting dropouts. Increasing latency only helps a little (i.e. to go from 75% to 85%, maybe 90%). In general however, 85% seems to be the limit. What are you guys seeing? Any suggestions?

A8V deluxe Bios 1013
2G Mushkin RAM
Enermax 550 (I think) power supply
3 SATA drives
1 IDE for system
DVDRW
Matrox P650 AGP

I'll have to check slot assignments when I get back in town - of course slot 1 is empty, but looking at the IRQ sharing for this board, every slot shares with something critical - SATA on 4(?), Firewire, USB and AGP - no way to avoid any of those - I need the extra SATA ports (Promise Raid, in non-raid config) for video, firewire for video capture, and of course thanks to iLok keys, USB is no longer optional.

I think I now have my Digiface in slot 2 (will try 4, but conflicting with SATA seems like a bad idea), UAD-1s in 4 and 5 I believe (need to change that as 5 also shares with the AGP slot). I had read this thread before building the system and concluded that UAD-1s location wasn't going to make a difference, so the dropouts are troubling. I could never get these two cards past 90% on the last system either, and assumed one of the more recent updates had something to do with it. I'm wondering now though.

With the new X2 rig, I had to removed 3-4 Cambridge EQs from my UAD-1 load test that ran fine on the last PC. Doesn't run without dropouts at the same instance count here. Sigh...I was hoping to push 96% or so with a faster PC, not lose 5-10%. Hopefully I've just spent too little time testing slot configs to get it sorted properly.

Also, any advantage to enabling ACPI 2.0? I have it disabled currently (ACPI install, but I assume that defaults to a 1.x version of table support).

Other than that the system runs great - just not so stellar with UAD-1s. Thonex 2 test runs at 128ms without pops/clicks.

 

[Animix]

My experience with the A8V is a little different so far - I can only get 75-85% out of my 2 UAD-1s without getting dropouts. Increasing latency only helps a little (i.e. to go from 75% to 85%, maybe 90%). In general however, 85% seems to be the limit. What are you guys seeing? Any suggestions?

A8V deluxe Bios 1013
2G Mushkin RAM
Enermax 550 (I think) power supply
3 SATA drives
1 IDE for system
DVDRW
Matrox P650 AGP

I'll have to check slot assignments when I get back in town - of course slot 1 is empty, but looking at the IRQ sharing for this board, every slot shares with something critical - SATA on 4(?), Firewire, USB and AGP - no way to avoid any of those - I need the extra SATA ports (Promise Raid, in non-raid config) for video, firewire for video capture, and of course thanks to iLok keys, USB is no longer optional.

I think I now have my Digiface in slot 2 (will try 4, but conflicting with SATA seems like a bad idea), UAD-1s in 4 and 5 I believe (need to change that as 5 also shares with the AGP slot). I had read this thread before building the system and concluded that UAD-1s location wasn't going to make a difference, so the dropouts are troubling. I could never get these two cards past 90% on the last system either, and assumed one of the more recent updates had something to do with it. I'm wondering now though.

With the new X2 rig, I had to removed 3-4 Cambridge EQs from my UAD-1 load test that ran fine on the last PC. Doesn't run without dropouts at the same instance count here. Sigh...I was hoping to push 96% or so with a faster PC, not lose 5-10%. Hopefully I've just spent too little time testing slot configs to get it sorted properly.

Also, any advantage to enabling ACPI 2.0? I have it disabled currently (ACPI install, but I assume that defaults to a 1.x version of table support).

Other than that the system runs great - just not so stellar with UAD-1s. Thonex 2 test runs at 128ms without pops/clicks.

I strongly suggest getting a 13 slot Magma chassis, putting the host card in slot #4 (if it's an ASUS A8V-Deluxe mobo) and forgetting about your problems. You may need to do a little shuffling of the cards within the Magma, but it's so friendly (at least mine is) that I'm running 3 x RME audio cards and 4 x UAD-1's and life is good.

 

[Akis]

dt, my 2 UAD-1's are in slots 3 & 5 on the A8V Deluxe (with an X2 4400) and I can use them @ up to 95% without any problems; if I pass that limit, they'll work fine for a couple of minutes and then the UAD-1 meter will jump to something over 100% and crackles will start occuring. Turning 1 UAD-1 plug off/on will solve this for another couple of minutes and so on. I talked to UA Brett Patrick about it and he said that's normal, just like what happens to a native CPU at these loads. However, I think that under 95% UAD-1 load you shouldn't get these problems. :?

 

[dt]

Akis - I will try that slot combo - good to know that config works.

DJ - A Magma is in the longterm plans as I will also have to run a Declink Blackmagic HD card in this machine. For sure that sounds like the most reliable solution.

Do you guys still have firewire and/or the Promise SATA controller enabled? I could test without them, but I still have to make it work with these enabled anyway.

Thanks and regards,
Dedric

 

[Animix]

Akis - I will try that slot combo - good to know that config works.

DJ - A Magma is in the longterm plans as I will also have to run a Declink Blackmagic HD card in this machine. For sure that sounds like the most reliable solution.

Do you guys still have firewire and/or the Promise SATA controller enabled? I could test without them, but I still have to make it work with these enabled anyway.

Thanks and regards,
Dedric

Hi Dedric,

I'm not using the FW or Promise controller so they are disabled.I'm currently getting good performance from IDE drives. when I decide to go SATA, I'm going to get a Silicon Image SATA controller and use it in the PCI slot that is sharing with the (disabled) Promise controller.

Regards,

 

[csl]

I've just set up a 4400 X2/A8V system, and I'm having problems keeping it stable. I'm pretty sure it's my Echo Mia though, as I can't stop the system from crashing with it in, and the slot doesn't seem to make a difference. I don't know if there are any specific settings I needed to set within the BIOS for it to work. Any other A8V/Mia users here? Now I can't even get the thing to power on.

Assuming I get it working again without any damage, I think I might just try and reinstall Windows again and reset the BIOS to its default settings. If the problem persists with the Mia I guess I'll have to think about an alternative card. Bloody computers.

 

[csl]

Oh, but it was fine with 2 x UAD-1s though when it was up.

Edit: Oh, I see the Echo Mia is in fact incompatible with the Via chipset. Great.

http://www.planetz.com/forums/viewtopic ... forum=19&9
http://www.hardwareanalysis.com/content ... 5548/?o=20
http://northernsounds.com/forum/showthr ... 022&page=1



Looks like an excuse to buy a Fireface...

 

[Jeraz]

JC-

Are you sure that it doesn't matter what slot that you put it in?

LOL.

I just downloaded the manual, as my copy of this board ( they had a few left at http://www.spacecentersystems.com ) is due to arrive via UPS on Friday.

I see that:

—Slot 1 is shared with firewire, slot 5 and AGP

If I were to put my UAD-1 in this slot, I would start by overiding my video card latency, prehaps resetting it to 64, and by disabling my firewire in the bios, since I don't use it. Then I would keep PCI slot 5 clear.

—Slot 2 is shared with Gigabit LAN.

I would probably put my soundcard in this slot, and when I am running demanding apps, would load a hardware profile that disabled all network services.

—Slot 3 shares with RAID and SATA.

If I am running IDE, I am loving this slot, but if I am running RAID or SATA, I would rather steer clear.

—Slot 4 is free in clear for interrupt D.

Any PCI Hog would be comfortable here, I bet.

—Of course, Slot 5 shares with 1 & Company


Point is, depending on what you are running and when (concurrency can matter in some cases), judicial selection of PCI slot may matter after all?

 

[Akis]

In my experience, the UAD-1 slot placement doesn't matter with this mobo.