iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
PCI/e nForce4 issue?
Sorry if this has been asked a thousand times....I've been searching but no definitive answer.
Is nForce4 still causing an issue with a PCI/e graphics card? I thought I would be safe since my audio card is USB, but perhaps there will still be an issue? At present I have NO video card, using the on-board accelerator, but I wanted to improve graphics slightly (the onboard nVidia 6200 looks pretty bad...blurry text, bad resolution in Sonar). So I figured I would get a new video card (PCI/e), but all this talk about how PCI/e will hog the bus has got me thinking twice....should I stick to what I have for stability?
I've read the articles being referred to in threads, most of them seem to date to feb 2005 though, so perhaps there has been some progress with nForce4 boards and their implementation of the PCI/e architecture now?
thanks in advance!
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
Thomas Campitelli
Max Output Level: -79 dBFS
- Total Posts : 598
- Joined: 2003/12/29 22:13:08
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 14:13:56
(permalink)
Last I heard, problems still exist. I think that the PCIe issues have more to do with hardware design than software. As a result, I think that the only fixes that will be forthcoming involve new hardware.
Your nVidia onboard card should not have problems with text and such. In fact, it should handle Sonar quite gracefully, although it is possible that there's some kind of problem. Before you get a new card, here's a question for you: what kind of monitor do you have? If it is an LCD, then check with the manufacturer to make sure you are running at the correct resolution and refresh rate. LCDs look very bad when you deviate from their intended specs. If you are using a CRT, try borrowing a friend's monitor to see if the same problem exists. I wish you luck.
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 16:58:43
(permalink)
Hi,
thanks for the answer! When you say hardware, I hope that means perhaps future versions of the VIDEO cards will solve the issue...not the motherboard...ho-hum....since I have the motherboard but no video card at present.
I'll look into the settings for the display. It is a LCD (19" Benq), and it looks great when its hooked up to my laptop....with the DAW it looks pretty bad (although pictures look fantastic when displayed from the DAW....?). I'll see what I can dig up....
thanks again!
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 17:14:29
(permalink)
I believe the fix is going to have to be either in the chipset or ........
...just give up and get a firewire audio card.
It doesn't seem possible that altering the design of the video cards will do anything.
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 17:21:07
(permalink)
Update: I messed around with the display settings in the nvidia driver windows, and sure enough I seem to have fixed the display problem. It still won't recognise my monitor properly (Win XP 64 bit....), but the display looks much better now! Thanks for making me pay a bit more attention to what I was doing... 
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 17:22:58
(permalink)
ORIGINAL: Steve_Karl
I believe the fix is going to have to be either in the chipset or ........
...just give up and get a firewire audio card.
It doesn't seem possible that altering the design of the video cards will do anything.
Actually the articles I have been reading state that both PCI AND Firewire audio cards have an issue when using a powerful PCI/e graphics card on the nForce 4 chipset.
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 18:31:31
(permalink)
Good to know ( well not really ... but thanks )
Steve
|
Guest
Max Output Level: -25.5 dBFS
- Total Posts : 4951
- Joined: 2009/08/03 10:50:51
- Status: online
RE: PCI/e nForce4 issue?
2006/03/12 18:47:33
(permalink)
ignorance is bliss i guess .. but i have a PCI/e GeForce4 card and have had no
problems with the USB bus ... and i've got some mighty hungry cards on the
bus too. what problems have been reported with this configuration?
thanks,
jeff
|
mildew
Max Output Level: -84 dBFS
- Total Posts : 338
- Joined: 2004/01/28 23:23:05
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 20:47:39
(permalink)
has anybody tried increasing the pci latency on the nforce4 boards? - is there a pci latency hacker program for them yet? - historically i was always able to fix via boards by increasing the sound card and lowering the video.
m
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 21:24:34
(permalink)
|
Guest
Max Output Level: -25.5 dBFS
- Total Posts : 4951
- Joined: 2009/08/03 10:50:51
- Status: online
RE: PCI/e nForce4 issue?
2006/03/12 21:40:52
(permalink)
ah .. that's why i'm not seeing any problems .. not running Sonar on that system ;-)
jeff
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/12 22:59:49
(permalink)
ORIGINAL: jmarkham
ah .. that's why i'm not seeing any problems .. not running Sonar on that system ;-)
jeff
or any high performance audio?
|
Guest
Max Output Level: -25.5 dBFS
- Total Posts : 4951
- Joined: 2009/08/03 10:50:51
- Status: online
RE: PCI/e nForce4 issue?
2006/03/13 00:32:14
(permalink)
ORIGINAL: Steve_Karl
ORIGINAL: jmarkham
ah .. that's why i'm not seeing any problems .. not running Sonar on that system ;-)
jeff
or any high performance audio?
i'm using it on a Protools HD2 .. and i'm not seeing any performance problems on that .. it's going pretty much to spec. jeff
|
losguy
Max Output Level: -20 dBFS
- Total Posts : 5506
- Joined: 2003/12/18 13:40:44
- Location: The Great White North (MN, USA)
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 01:04:01
(permalink)
ORIGINAL: mildew
has anybody tried increasing the pci latency on the nforce4 boards? - is there a pci latency hacker program for them yet? - historically i was always able to fix via boards by increasing the sound card and lowering the video.
That's the problem, PCI latency is not adjustable with PCIe... it doesn't even really exist except on the "classic" PCI bus.
|
losguy
Max Output Level: -20 dBFS
- Total Posts : 5506
- Joined: 2003/12/18 13:40:44
- Location: The Great White North (MN, USA)
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 01:08:16
(permalink)
ORIGINAL: jmarkham
ah .. that's why i'm not seeing any problems .. not running Sonar on that system ;-)
ORIGINAL: jmarkham
i'm using it on a Protools HD2 .. and i'm not seeing any performance problems on that .. it's going pretty much to spec.
ADK Scott's tests are very challenging, way more than Scott Reams' test. It's to help tease out the various realtime performance issues. IIRC he's reported problems with more apps than merely SONAR. In your case jeff, if you're not seeing problems, I'd say that you've been quite fortunate indeed. Edit: The video card may also be a lesser-hungry variety of PCIe card. That helps the equation (but doesn't solve it 100%).
post edited by losguy - 2006/03/13 01:16:34
|
Guest
Max Output Level: -25.5 dBFS
- Total Posts : 4951
- Joined: 2009/08/03 10:50:51
- Status: online
RE: PCI/e nForce4 issue?
2006/03/13 01:30:46
(permalink)
i suspect there may be two things helping me out then:
1. i've got a fairly cheapo GeForce4 PCI/e card.
2. when they ported PT from Mac to Windows, they supposedly wrote their own
low-level graphics lib .. which suggests ... nothing fancy .. and probably not as bus
intensive as it could be.
in any case, i will consider myself lucky...
jeff
|
paulh
Max Output Level: -90 dBFS
- Total Posts : 44
- Joined: 2006/02/24 18:45:24
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 05:51:47
(permalink)
I hope this message is not too technical, if it is just ask for help: All PC hardware is attached to the PCI/e bus: hard disks, CD-rom, graphics*, sound card, USB, FireWire etc. Think of the PCI/e bus as a multi-lane freeway. Data moves up and down the bus but some graphics cards take more than their fair share of "lanes" or bandwidth. They do this to improve games performance in a very competitive benchmark industry. Unfortunatly other devices suffer and get lower bandwidth, causing like a temporary traffic jam which can result in pops/clicks and sync loss. There are utilities to stop graphics using too much of the bus, you may also find that disabling hardware acceleration helps. Right-click display, Settings | Advanced | Troubleshoot. The new PCI/e bus is worse because its design allows greater bandwidth to critical devices, like graphics. But the problems can be removed by making sound cards a higher priority than the graphics, whether or not that is easy is the real problem as much will depend on the sound/graphics combination you have. A less aggressive graphics driver will help, but that is a paradox as games industry want the fastest graphics on the planet!
post edited by paulh - 2006/03/13 06:11:58
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 06:56:37
(permalink)
ORIGINAL: paulh
I hope this message is not too technical, if it is just ask for help:
All PC hardware is attached to the PCI/e bus: hard disks, CD-rom, graphics*, sound card, USB, FireWire etc. Think of the PCI/e bus as a multi-lane freeway.
Data moves up and down the bus but some graphics cards take more than their fair share of "lanes" or bandwidth. They do this to improve games performance in a very competitive benchmark industry. Unfortunatly other devices suffer and get lower bandwidth, causing like a temporary traffic jam which can result in pops/clicks and sync loss.
There are utilities to stop graphics using too much of the bus, you may also find that disabling hardware acceleration helps. Right-click display, Settings | Advanced | Troubleshoot.
The new PCI/e bus is worse because its design allows greater bandwidth to critical devices, like graphics. But the problems can be removed by making sound cards a higher priority than the graphics, whether or not that is easy is the real problem as much will depend on the sound/graphics combination you have.
A less aggressive graphics driver will help, but that is a paradox as games industry want the fastest graphics on the planet!
Which in my case means I am better off just sticking with the onboard graphics, nForce 6100, rather than adding a PCI/e card. I was (obviuosly wrongfully) thinking I would actually have offloaded some system resources by increasing the graphics computing power - but that was without taking increased bus trafic into account. Learn something everyday....thanks for your time!
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
BruceEnnis
Max Output Level: -58.5 dBFS
- Total Posts : 1665
- Joined: 2003/11/05 16:48:01
- Location: Maryland
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 09:28:13
(permalink)
The problem is noticably poor performance relative to these benchmarks. I've been hearing this for over a year now I posted the results from my AMD 4000+ nForce4 system they are equal to any other system with the same specs, it has no dropouts, no video problems, and I recored 16 tracks at a time whenever I go portable. I've done quick mixes on location with more than 30 tracks my system specs are listed below. SHUTTLE SN25P nVIDIA nFORCE4 CHIPSET AMD ATHLON 64 4000+ 2 GB PC3200 400MHZ CL3 (3-3-3) DDR DIMM ATI Radeon X700 Pro 256MB PCI Express NEC ND-3520A 16X/16 DVD +R/+RW DUAL DOUBLE LAYER MITUSMI 1.44 FLOPPY MAXTOR 250GB SERIAL ATA150 7200RPM 8MB 2) 15" LCD NEC monitors Microsoft Windows XP Professional Logitech Cordless Optical Mouse and Keyboard Digidesign 002 Rack Factory Bundle Furman PL-Plus DM Behringer ULTRAGAIN PRO8 DIGITAL ADA8000 Behringer HA8000 Behringer ULTRA-DI PRO DI800
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 10:00:53
(permalink)
|
ooblecaboodle
Max Output Level: -54 dBFS
- Total Posts : 2102
- Joined: 2004/05/01 21:52:56
- Location: North Wales
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 11:24:26
(permalink)
am I right in assuming that the dual core processors do not show this problem?
|
losguy
Max Output Level: -20 dBFS
- Total Posts : 5506
- Joined: 2003/12/18 13:40:44
- Location: The Great White North (MN, USA)
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 12:05:07
(permalink)
Bruce, I've been thrilled to see the performance that you have been getting with that system configuration. But the nForce4 problem is pernicious, enough so that for someone to benefit from your success with reasonable assurance, they would just about have to duplicate your exact system. For instance, a different choice of video card (or even a hardware/firmware rev of the same card) could easily break the recipe.
At any rate (and to answer caboodle too), yes, generally speaking, plugging in a X2 seems to pretty much solve the nForce4 problem (perhaps better said, "cover up" the problem). That is, until you try to use UAD cards (again, generally speaking).
|
jcschild
Max Output Level: -41 dBFS
- Total Posts : 3409
- Joined: 2003/11/08 00:20:10
- Location: Kentucky y'all
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 12:32:08
(permalink)
HI All,
PCIe motherboard issues, predominately Nforce4, and KT 890. (disco'ed anyway)
single core proc. very poor performance compared to nforce 3/Via KT 800. no or poor UAD ability, powercore ok.
Dual Core: fairly good performance but still not that of Nforce3/KT 800. NO UAD powercore ok.
what video card you use has nothing to do with it, its the fact that the PCIe 16x is given highest priority.
and the PCI bus suffers (this includes, Firewire/usb/Sata/IDE etc)
on the older Intel 915/925 the power of the video card was in direct porportion of your ability to do audio.
present Intel 945/955/975 do not exibit this problem, however lower latency is better on AMD.
also with Present Intel when using UAD/Powercore buffer setting have to be increased with thier use as compared to without.
also we have never never been able to duplicate any ones claims of ANY nforce4 system functioning properly.
despite buying the exact motherboard they claim works. we have tested over 30 boards in the last yr and revisted many of them.
if you are not using a UAD then nforce 4 is fine with dual Core, the ultra is the best one. SLI is the worst.
Scott
ADK
|
losguy
Max Output Level: -20 dBFS
- Total Posts : 5506
- Joined: 2003/12/18 13:40:44
- Location: The Great White North (MN, USA)
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 13:11:14
(permalink)
Scott, thanks for coming in from the trenches and clarifying the latest. For how helpful and committed you are, you are a gentleman AND a great businessman in my book.
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 13:33:48
(permalink)
Scott's message now makes me think that there is NO way I can get my system to work properly. I have no real complaints with it at this point, I don't get dropouts ar pops/clicks etc. I can comfortably track 8 drum inputs at the same time running about 7% CPU and disk. I don't aspire to much greater things, but I did think I had a fairly good system...apparently not. Although as I am not using PCI/e at all perhaps I am OK?
Can someone point me to a recent benchmarking test that I can run to see what is up? (Sonar 5.0.1 SE)
thanks!
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
jcschild
Max Output Level: -41 dBFS
- Total Posts : 3409
- Joined: 2003/11/08 00:20:10
- Location: Kentucky y'all
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 14:13:47
(permalink)
|
iceman2058
Max Output Level: -87 dBFS
- Total Posts : 154
- Joined: 2005/12/04 05:50:21
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 15:20:29
(permalink)
Scott,
thanks for the link. I've downloaded the test, just a question now about the test settings. I can play around with the buffer size with no noticable impact on CPU load when playing back the project. Its the latency adjustment I need to carry out to see the CPU load change. Have I missed something?
For example, I can't use a 9ms latency properly for this project - 12ms makes it stable. 6 or 3 just drop out immediately. Does that seem ballpark about right?
Iceman2058 Vdrums, Edirol, Sonar 6PE (x64) AMD X2 4400+, Gigagyte GA-K8N51GMF-9, WD Raptor 74GB, WD 80GB/Maxtor 250GB, 2GB RAM, XP x64, Edirol UA-101 Music here: Iceman's Drum Page
|
Scott Reams
Max Output Level: -56 dBFS
- Total Posts : 1918
- Joined: 2003/11/06 15:32:28
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 16:45:02
(permalink)
To add some of my own findings to this...
The problem appears to mainly revolve around the PCI bus having a lower priority than the PCIe bus. Issues can appear (but don't always) if using PCI devices (which includes Firewire, since it is a PCI device on these boards). USB, however, is not a PCI device, and doesn't appear to be subject to the issue at all. Firewire devices can skirt the issue through the use of a PCIe Firewire card. UAD1s can work on these boards (even two of them) if the system is configured so that the UAD1s are the only active PCI devices... and through the use of a Microsoft utility that allows you to set specific processor affinity for each UAD1 card. I've got two customers that each have two UAD1 cards on nForce4 motherboards... and are having success.
-Scott
|
ooblecaboodle
Max Output Level: -54 dBFS
- Total Posts : 2102
- Joined: 2004/05/01 21:52:56
- Location: North Wales
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 19:33:05
(permalink)
HI All,
PCIe motherboard issues, predominately Nforce4, and KT 890. (disco'ed anyway)
single core proc. very poor performance compared to nforce 3/Via KT 800. no or poor UAD ability, powercore ok.
Dual Core: fairly good performance but still not that of Nforce3/KT 800. NO UAD powercore ok.
what video card you use has nothing to do with it, its the fact that the PCIe 16x is given highest priority.
and the PCI bus suffers (this includes, Firewire/usb/Sata/IDE etc)
on the older Intel 915/925 the power of the video card was in direct porportion of your ability to do audio.
present Intel 945/955/975 do not exibit this problem, however lower latency is better on AMD.
also with Present Intel when using UAD/Powercore buffer setting have to be increased with thier use as compared to without.
also we have never never been able to duplicate any ones claims of ANY nforce4 system functioning properly.
despite buying the exact motherboard they claim works. we have tested over 30 boards in the last yr and revisted many of them.
if you are not using a UAD then nforce 4 is fine with dual Core, the ultra is the best one. SLI is the worst.
Scott
ADK Thanks for the clarification, Scott. I would just like to take the time to say that I really appreciate all the hard work you put into your research. Good man.
|
Steve_Karl
Max Output Level: -50 dBFS
- Total Posts : 2534
- Joined: 2003/11/06 20:53:26
- Location: Pittsburgh, PA USA
- Status: offline
RE: PCI/e nForce4 issue?
2006/03/13 19:57:51
(permalink)
Both Scotts ! Thank you both!
|