dsp capacity limit reached or, Scope Pro busted?
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
A capacity limit of your DSP cards has been reached. The specific cause is: Not enough DSP memory: cannot connect PC Master 32k Delay.Tap1 and DLEXTM0.RI
This always appears to happen at the point where DSP #9 of #14 is reached in the DSP load window. I'm thinking that my Scope Pro card is busted.
Nooooooooooooooooooooooooooooo!!!!!
I've done full reinstalls in 3.1 and 4. And I've moved the card to another machine to test. Same result.
Is there a way to either disable an individual DSP on a card? Or is there a way to repair bad hardware?
This always appears to happen at the point where DSP #9 of #14 is reached in the DSP load window. I'm thinking that my Scope Pro card is busted.
Nooooooooooooooooooooooooooooo!!!!!
I've done full reinstalls in 3.1 and 4. And I've moved the card to another machine to test. Same result.
Is there a way to either disable an individual DSP on a card? Or is there a way to repair bad hardware?
to repair a board in a way that one or more DSPs are exchanged is close to impossible for us regular folks.
I normally don't shriek away from any solder job (even with inappropriate tools if nothing else is available - and I HAVE soldered on a Pulsar board already), but the DSPs is too much (and too tight), no - thanks.
to disable a DSP might be more promising, but it depends on circuit details (of which I have no idea at all)
at least there are ICs that you can switch into a transparent mode with a 'chip select' line (dunno if this applies to Sharcs), but it was a regular method for memory upgrades (many) years ago by cutting off the chip select and solder larger Rams on top of the existing ones...
anyway, even if your error message points in the direction of hardware failure, it is in NO WAY an evidence that it tells the truth.
it could simply be a coincidence (what do you use to test ? do non-delay related loads of the DSP come up with the same message ?)
messed up (driver-)software can easily do the same - my Pulsars are currently hosted by a dual CPU system and it's very unstable, but only with more complex devices.
Lot's of hardware related errors, too.
I don't worry because I anticipated trouble (tnx to PlanetZ) and I'll move the boards again, but I just had to see it myself...
good luck, Tom
I normally don't shriek away from any solder job (even with inappropriate tools if nothing else is available - and I HAVE soldered on a Pulsar board already), but the DSPs is too much (and too tight), no - thanks.
to disable a DSP might be more promising, but it depends on circuit details (of which I have no idea at all)
at least there are ICs that you can switch into a transparent mode with a 'chip select' line (dunno if this applies to Sharcs), but it was a regular method for memory upgrades (many) years ago by cutting off the chip select and solder larger Rams on top of the existing ones...

anyway, even if your error message points in the direction of hardware failure, it is in NO WAY an evidence that it tells the truth.
it could simply be a coincidence (what do you use to test ? do non-delay related loads of the DSP come up with the same message ?)
messed up (driver-)software can easily do the same - my Pulsars are currently hosted by a dual CPU system and it's very unstable, but only with more complex devices.
Lot's of hardware related errors, too.
I don't worry because I anticipated trouble (tnx to PlanetZ) and I'll move the boards again, but I just had to see it myself...

good luck, Tom
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
Ah, well thanks for the suggestions. It is unfortunate to have to toss a $1500 board though. Kind of depressing. I'll likely be selling all my plugins here for peanuts, and then move on to the next big thing.
It will be cool if Creamware does ever put out the 200MHz chip. I would consider buying again at that point. But as it is now, I can't see buying the same thing again and getting stuck in the same place. I'm going to miss the Scope environment. But I guess that's life in the world of boutique sound cards...
<font size=-1>[ This Message was edited by: ScofieldKid on 2005-10-08 19:14 ]</font>
It will be cool if Creamware does ever put out the 200MHz chip. I would consider buying again at that point. But as it is now, I can't see buying the same thing again and getting stuck in the same place. I'm going to miss the Scope environment. But I guess that's life in the world of boutique sound cards...
<font size=-1>[ This Message was edited by: ScofieldKid on 2005-10-08 19:14 ]</font>
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
- next to nothing
- Posts: 2521
- Joined: Mon Jul 29, 2002 4:00 pm
- Location: Bergen, Norway
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
I admit I was in doubt if it would be useful at all to answer those 2 questions...On 2005-10-08 19:11, ScofieldKid wrote:
Ah, well thanks for the suggestions. It is unfortunate to have to toss a $1500 board though. Kind of depressing. ...
I wrote 7 lines about why 'quick and dirty' repairs for cheapo don't apply here (it's no secret that you can send the board to CWA anytime).
8 lines were dedicated to the fact that it's not likely to be broken.
Of course there is always a chance, but 99% of all hardware faults reported here turned out to be combinations of bios/mobo/software/installation errors.
if the latter cannot be sorted out by yourself (with assistence from here), then and only then it's time for a proper hardware service (as Gary suggests)
Either they do have the professional tools for circuit repair, or they charge less for exchanging the board and reuse parts.
In any such case CWA has been reported to be very helpful - no need for such a depressing answer

cheers, Tom
Minimax also has delay lines (afaik)On 2005-10-08 22:37, ScofieldKid wrote:
...Then duplicated using just MINIMAX...
given the fact that those delay modules are identical in MV and MM (both are CWA devices), and are messed somehow, you would receive the buffer error even though the hardware is happy.
fill up the DSPs with completely FX free devices to see if #9 is strange
cheers, tom
-
- Posts: 777
- Joined: Sat May 25, 2002 4:00 pm
- Location: The Great White North
- Contact:
DSP memory does not equal the processor itself. Generally you'll run out of DSP memory long before you run out of DSP horsepower, and it can be caused even by tons of connections.
Post a picture of your setup. I really really really doubt this is a failed processor, I suspect it is just maxed out DSP memory. You would get nasty errors and possibly crashes if there was a failed hardware component.
I suspect your options are:
1) use fewer devices
2) buy a second board!!!

Post a picture of your setup. I really really really doubt this is a failed processor, I suspect it is just maxed out DSP memory. You would get nasty errors and possibly crashes if there was a failed hardware component.
I suspect your options are:
1) use fewer devices
2) buy a second board!!!

Actually I ran my Creamware cards in a dual Xeon for several years and the ONLY issue I'm aware of is that vorb would BSOD everytime on use. I've confirmed that using Vorb in my single cpu machine (a second PC where all my scope cards now reside) works fine. I never had to fiddle with priority modes for Logic as many do, since it only uses a single cpu for its engine, and my UI was never sluggish under load (not even in Scope). Now prior to SFP 3.01c or whatever it was I had *major* problems (it wouldn't even run on an NT OS that supports duals until 2.04g BETA).On 2005-10-08 15:12, astroman wrote:
...messed up (driver-)software can easily do the same - my Pulsars are currently hosted by a dual CPU system and it's very unstable, but only with more complex devices.
Lot's of hardware related errors, too.
I don't worry because I anticipated trouble (tnx to PlanetZ) and I'll move the boards again, but I just had to see it myself...
good luck, Tom
Now if you have an AMD box your experiences may certainly be different than mine. When I built my box I was patiently waiting for a stable MP/MPX chipset for audio use, and it never appeared so I caved in and went the pricier route. Nforce3 seems to be the most stable AMD rig, and I'm unsure of its dual core support (it certainly doesn't support MP/MPX). I'll admit I've not read much about dual core AMD successes or failures but I wouldn't be surprised if there are problems.
-
- Posts: 777
- Joined: Sat May 25, 2002 4:00 pm
- Location: The Great White North
- Contact:
Stardust, it's only one card -- no S/TDM cables. 
Also PCI / system memory have nothing to do with the (small amount of) onboard DSP memory. IIRC, the PCI bus max-outs cause different error messages, no?
I've got a list of all the error messages I've ever seen, somewhere around here, but I can't seem to find it...

Also PCI / system memory have nothing to do with the (small amount of) onboard DSP memory. IIRC, the PCI bus max-outs cause different error messages, no?
I've got a list of all the error messages I've ever seen, somewhere around here, but I can't seem to find it...
no, it's that cheapo dual P3/1G HP graphic workstation mentioned in OT.On 2005-10-09 05:43, valis wrote:
...Now if you have an AMD box your experiences may certainly be different than mine...
Regarding PCI capacity it's not different from a BX and probably inferior to I815, so it's no big deal to move the cards.
There are several devices that just hang or crash the machine - though I haven't ruled out midi yet.
cheers, tom
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
-
- Posts: 307
- Joined: Thu Nov 13, 2003 4:00 pm
- Location: Oregon
- Contact:
Well, I'm really trying to a) verify the behavior, and b) understand the logic.
Anyone else with a Scope Pro can try this? I think I get at most 7 MasterVerbPro's, and then MVP #8 is what kills it. But in the DSP Load window, it's always the point where DSP #9 gets put in use.
<font size=-1>[ This Message was edited by: ScofieldKid on 2005-10-10 07:59 ]</font>
Anyone else with a Scope Pro can try this? I think I get at most 7 MasterVerbPro's, and then MVP #8 is what kills it. But in the DSP Load window, it's always the point where DSP #9 gets put in use.
<font size=-1>[ This Message was edited by: ScofieldKid on 2005-10-10 07:59 ]</font>
there is no need to verify anything - you can do it yourself.
Load a Minimax, set it to max # of voices, this should fill a significant amount of your dsps.
if it didn't fill dsp #9 load another instance, set to a high voice count
did it break on #9 ? probably not
set one instance of minimax to a # of voices that fill (say) 8 dsps.
according to your hypothesis the next masterverb pro will break on the memory of chip #9
did it break ? I bet not...
btw Masterverb Pro's don't count for benchmarking - use Masterverbs instead to receive comparable results - MV Pro is an optional device.
I guess it's significantly more demanding, so your 7 instances are probably very good.
regarding the logic:
reverbs are built from delay lines
delay lines need memory - some on chip, some over the PCI bus
both resources are limited
you'll never need that much instances of a top reverb - it's simply a waste of resources.
No human ear will be able to distinguish them in a production - in other words: you could achieve the same result with less demanding devices.
there's the solo singer and the background vocals - the solo Stradivarious and the violin section. There is a reason to isolate the 'better' performer
cheers, tom
Load a Minimax, set it to max # of voices, this should fill a significant amount of your dsps.
if it didn't fill dsp #9 load another instance, set to a high voice count
did it break on #9 ? probably not

set one instance of minimax to a # of voices that fill (say) 8 dsps.
according to your hypothesis the next masterverb pro will break on the memory of chip #9
did it break ? I bet not...
btw Masterverb Pro's don't count for benchmarking - use Masterverbs instead to receive comparable results - MV Pro is an optional device.
I guess it's significantly more demanding, so your 7 instances are probably very good.
regarding the logic:
reverbs are built from delay lines
delay lines need memory - some on chip, some over the PCI bus
both resources are limited
you'll never need that much instances of a top reverb - it's simply a waste of resources.
No human ear will be able to distinguish them in a production - in other words: you could achieve the same result with less demanding devices.
there's the solo singer and the background vocals - the solo Stradivarious and the violin section. There is a reason to isolate the 'better' performer

cheers, tom