PDA

View Full Version : Buffer drop out on RME 9652



Alex
12-02-2009, 05:34 PM
My problem is buffers dropping at a certain time on my amd x2 RME cards set at 256ms 40ch running at 60% I drop out at 1.5 hrs toggle on off it clears for 1.5 hrs. Switched to new Intel 9450 no parts used for the x2 machine RME cards set to 64ms load is 40% drop out at 30min , set cards to 32ms load at 72% drop out in 10min.

RME said its not them I load new drivers from there site I have no irq conflicts, no dpc latency issues and no onboard sound running.

Im Stumped

Alex

905shmick
12-02-2009, 06:54 PM
Try setting SAC to use the "Force RealTime Priority Class" option. You can also try "Force Single CPU".

gdougherty
12-02-2009, 07:21 PM
My problem is buffers dropping at a certain time on my amd x2 RME cards set at 256ms 40ch running at 60% I drop out at 1.5 hrs toggle on off it clears for 1.5 hrs. Switched to new Intel 9450 no parts used for the x2 machine RME cards set to 64ms load is 40% drop out at 30min , set cards to 32ms load at 72% drop out in 10min.

RME said its not them I load new drivers from there site I have no irq conflicts, no dpc latency issues and no onboard sound running.

Im Stumped

Alex

Are you talking about slipped output buffers?

Alex
12-03-2009, 07:10 AM
Yes Im talking about slipped buffers it goes to garbled audio in all outputs, I am running forced realtime. When this happens the cpu meter drops about 20% in SAC and in the task manager. This is happening on 2 completly different Pc's. I am going to pull 1 card run it to the max and see if it happens then swap out cards and try agian. I changed out systems thinking this was a old tired AMD issue.

Alex

Bob L
12-03-2009, 07:52 AM
Alex,

Are you using the ASIO drivers? If not... try that.

You may also want to try forcing single cpu mode in the options menu... see if that helps.

Though... with your apparent repeating times before garble... it does suggest that something in the system is set to come on and that process is stepping on the cards.

Bob L

Alex
12-03-2009, 08:10 AM
Hey BOb this is problem I have been having for sometime now. It is reoccurring in 2 seperate system setups 2 different versions of XP. The system I have the 9652 cards in ran all weekend on a Raydat card 32ch in 8out with no problems. I changed out the card put in the 52's installed new drivers and I have the same old problem that the AMD sys had?

Alex

And I am running ASIO

Bob L
12-03-2009, 09:11 AM
Well... then it sounds like possibly a card problem... perhaps a component failure on one of the cards when it heats up.

Bob L

KUI
12-03-2009, 10:03 AM
Sounds like the AMD cool and quiet driver regulating processor speed according to load. uninstall that driver if you installed it from your mobo driver disk .
I did extensive test over the summer with AMD Dual and Quad cores. After some months of frustrating buffer dropping I switched to Intel
Core2Duo processors and the issue disappered. Also make sure that you do not have C1e enabled in the bios. This lets the mobo control processor speed.
The AMD's kinda worked but in the end Ii deemed them to not be as stable as their intel counter part.

kui

Naturally Digital
12-03-2009, 11:21 AM
Hey BOb this is problem I have been having for sometime now. It is reoccurring in 2 seperate system setups 2 different versions of XP. The system I have the 9652 cards in ran all weekend on a Raydat card 32ch in 8out with no problems. I changed out the card put in the 52's installed new drivers and I have the same old problem that the AMD sys had?Which 9652's are these? HDSP or Hammerfall? Make sure you're running the correct driver and if it's HDSP with latest driver, make sure to have the correct firmware. If it's the Hammerfall, you have to get the old driver from the 'driver archive' section of the rme site.

Having said that, it certainly sounds like something is kicking in and stealing cycles from the audio driver.

Alex
12-04-2009, 03:53 PM
Ok I have tried it all RME still says it not the cards but how can they have they same problems in 2 different systems. I pulled the cards put 1 in ran it hard got it to slip 67 output buffers pulled it ran the other card no drops. Reinstalled them both got 267 slipped buffers running 2 @ 128 after 2hrs and it corrected them but I had about 1sec delay between the in and outs. I think I will just get a RAYDAT and shelf the card that slips buffers.

Also RME thinks its a MB issue I asked how 1 is asus w/x2 and the other is evga w/quad 9450? no comparison

And yes they both have new firmware and drivers

Alex

Iain Westland
12-04-2009, 04:20 PM
Its the amd, I had no probs with my single core then upped to dual (7750BE)and voila, same as you.

force single, normal or high priority, turn off everything in the bios you dont use.

its also predominantly SAC that has this prob, SAW does not display it to the same extent.

iain

Iain Westland
12-04-2009, 04:23 PM
Think the software is too quick for the hardware:D:D

1-0 Bob

iain

Warren
12-04-2009, 04:26 PM
Hi Alex:

I have had simular issues with an RME HDSP 9652 (1)
Things I found that can step all over your good day are:
Running Anti-Virus programs or extra non required programs in the background, Including basic Microsoft services. www.blackviper.com is a great resource. If you don't need a service to run, while running your DAW or SAC Live rig don't. Sometimes video cards hangup process time on your CPU.
Microsoft by default runs a lot of options (Services) for use by all people, that we doing recording and Live sound don't always need.

I feel its more software related then hardware.
Hope you get to the problem.

All the best!
Chris

Alex
12-04-2009, 04:51 PM
I changed to the intel for more power and coundnt belive the problem would follow to a different system. To me it still seems like a card issue, the newer 1 is dropping buffers and the older 1 is stable. I have checked it with dpc latency checker and I have no issues.

Alex

Naturally Digital
12-04-2009, 06:51 PM
I changed to the intel for more power and coundnt belive the problem would follow to a different system. To me it still seems like a card issue, the newer 1 is dropping buffers and the older 1 is stable. I have checked it with dpc latency checker and I have no issues. I know this feels like being grilled at an inquiry and no one means to insult your intelligence but can you share as much about the setups as possible? Especially what are the video cards and what are the exact model numbers on the mbs?

Do you have any other cards installed in the machines?

Have you tried all the pci slots on the mb's? If so, did you get different results when you moved it around?

Alex
12-04-2009, 07:54 PM
I dont feel grilled, its just a pain to deal with.

I have used all the PCI's, each board has 2. On the amd w/ asus M4A78-EM board it had dual monitors on board sound is off and it has dual gig lan. The EVGA nForce 790i Ultra SLI board has no video and Im using an EVGA GeForce 9400 GT PCIe 2.0, sound is off and it has dual gig lan. They are so different I cant see anything that can be causing the buffer slip. Like I wrote earlier one card never slips on its own the other will. The AMD sys was running a stripped down version of XP and the Intel is running a full XP Pro sp2

Alex

gdougherty
12-05-2009, 10:14 AM
I dont feel grilled, its just a pain to deal with.

I have used all the PCI's, each board has 2. On the amd w/ asus M4A78-EM board it had dual monitors on board sound is off and it has dual gig lan. The EVGA nForce 790i Ultra SLI board has no video and Im using an EVGA GeForce 9400 GT PCIe 2.0, sound is off and it has dual gig lan. They are so different I cant see anything that can be causing the buffer slip. Like I wrote earlier one card never slips on its own the other will. The AMD sys was running a stripped down version of XP and the Intel is running a full XP Pro sp2

Alex

And you've run only one card at a time? If one card is perfectly fine in one slot and the other causes problems in the same slot with nothing changed otherwise, then I would certainly suspect the card, no matter what RME said before. If that's where you've gotten to, then I'd call back to RME and go through things again.