Robustness of V5 ?

[nebelfuerst]

Hi,

I've one machine with a 14 DSP+ 4 DSP Pulsar in it. I installed the latest V5.
Although it works quite well, I often experience lockups or get "recursion limit 1000" when wireing the equipement.
I have to kill Scope with taskmanager to continue...

Is this normal or a different problem of my setup ?
bye

[astroman]

the error you describe isn't even version 5 specific, though it may be introduced by device modifications (or new devices) in that version.
Most likely it's a limit of nesting processing levels, as that's what a recursion is about.
They want to prevent an infinite loop by setting the limit to 1000 (in that case you experience)
Basically a recursion is a feedback loop, a function that calls itself at the end of processing starting allover again.
Usually this is a very 'elegant' approach to program, as long as the 'exit condition' is met...
The message kind of tells you 'last exit missed'. Could happen in any device of any version.

cheers, Tom

[nebelfuerst]

Yep. I know what recursion is, but I didn't expect to have Scope to bring such an error if I wire something wrong by accident. I played a lot with V4 but I never had so much serious locks like I've seen with V5.

With V4 I got some errors, too. But the could be fixed with some clicks. V5 sometime raises 100% CPUload and I have to shoot it by taskmanager. Nevertheless I want to work with V5 on the long run, but I'll wait for the next update until I retry.

[astroman]

just forget about it...

[Gordon Gekko]

its always amazing to read of how much tech specs you can pull while only guessing at the problem. this guy witnessed a big difference between two versions of the same software, running on the same hardware, unless i'm mistaken. please Tom, try to look at it from his angle as it is valuable info for SC, even if they know about it already. sorry for the rant

[maky325]

its always amazing to read of how much tech specs you can pull while only guessing at the problem.

Tell me about it...

[astroman]

sorry Gordon, I have no idea what kind of intent you interpret into my reply...
I have answered strictly from his point of view picking up his questions

...Yep. I know what recursion is, but I didn't expect to have Scope to bring such an error if I wire something wrong by accident. I played a lot with V4 but I never had so much serious locks like I've seen with V5.

Scope 5 is developed on XITE with compatibility in mind - a reasonable assumption, isn't it ?
I pointed out a significant difference between the chips, which I consider capable to cause such 'strange' errors.
That's nothing but an explanation... what's so bad about it ?

With V4 I got some errors, too. But the could be fixed with some clicks. V5 sometime raises 100% CPUload and I have to shoot it by taskmanager. ...

100% CPU load suggests that the system is doing a hell of whatever.
That's just plain wrong - it does almost nothing while looping (almost) idle, but it can't exit the loop.
This way the CPU is locked out completely for other processes, not available at all - from the perspective of another process this looks identically to a 100% busy CPU performing a 'real' task.

Again this is just an explanation for what's happening.
It's highly probable that it's a variation of the problem described further up.
Imho he seemed interested in an explanation for the strange error, if you consider it rubbish... provide a better one

cheers, Tom

[Gordon Gekko]

it's just that some of your posts sound to me like you try to dismiss issues for SC, drown them in some (perhaps valid) tech analysis.. if not never mind my rant

[julian]

the recursion bug should be fixed in the final release...
(if we talk about the same bug that is )
there are some bugs in the GUI elements that caused this (like the drag and drop text field from the sampleOsc or vdat...)

julian
--
SC

[Mr Arkadin]

Might help us and SC to actually state your system. Was it indeed the exact same system for v4 and v5? i have none of these errors so it's hard to blame v5 alone.

[fidox]

well, i can just talk from my experiences,
i think first release of version 5 was in May and later second one,
that's now 6 months and in all that time i didn't experiences any errors at all,

for me , it's already stable for what i need and work with,
so i guess it's something else,


best,

matej

[Gordon Gekko]

+1
Scope 5/win7 seems like a very robust combo . I unplugged/re-plugged the pcmcia card from its slot while V5 not loaded (this is a laptop/magma conf) and it did not blue screen like it did on a V4.5/winXP combo. It would be interesting to know on what OS nebelfuerst installed it. unfortunately for me and whether the fault is on win7, magma driver or V5 the system can't load the same number of plug-ins (as in masterverb test) as before.. still experimenting though, on a triple boot fedora/winXP-scope4.5/win7-scope5/

[time_chase]

I think the new drivers may have solved my blue screen crashes - the crashes appear to only occur when i use the onboard sound card (realtek) at the same time with scope, eg. watching digital tv or divx movie. Was playing around with SFP5 for several hours last night with a divx movie playing in the background, no crashes.

So far it seems faster - loading and shutting down. The only crash i had was when i started up nuendo and it was scanning the scope vst stuff (xtc??), crashed at scScope.sys or something. I'm guessing it was because i had scope running in the background? Anyway i ended up deleting the vst stuff scope installed in my vst folder (dont use xtc).

The 'problem' with the save project prompt is still there. If i saved it before i exit, why do i need to be prompted to save it again? surely it's not that hard to have a check flag in the program to detect changes and to signal for a save?? sigh...

[garyb]

I think the new drivers may have solved my blue screen crashes - the crashes appear to only occur when i use the onboard sound card (realtek) at the same time with scope, eg. watching digital tv or divx movie. Was playing around with SFP5 for several hours last night with a divx movie playing in the background, no crashes.

So far it seems faster - loading and shutting down. The only crash i had was when i started up nuendo and it was scanning the scope vst stuff (xtc??), crashed at scScope.sys or something. I'm guessing it was because i had scope running in the background? Anyway i ended up deleting the vst stuff scope installed in my vst folder (dont use xtc).

The 'problem' with the save project prompt is still there. If i saved it before i exit, why do i need to be prompted to save it again? surely it's not that hard to have a check flag in the program to detect changes and to signal for a save?? sigh...

shared irq?

[katano]

is the recursion bug fixed in the final version?

[julian]

it should be...
hard to say if i only hear of "recursion limit" without the filename causing it
there had been a bug in one of the RC files that is fixed thou.

julian