CInstRunOnce::vCompile>error:using more than 40000 bytes

infuzion · 5 Jan 2008

I am sure the usual "Send the file to [email removed]" applies here.

malc · 7 Jan 2008

Indeed, if you can send me a the schematic I can take a look.

Many thanks.

TomC · 13 Jan 2008

malc wrote: Indeed, if you can send me a the schematic I can take a look.

Many thanks.

Done!

Matthieu · 12 Aug 2008

Hi

I am getting this same 40000 bytes problem.. While using 32 "state variable" filter. (16 to the left channel, 16 to the right) I dont get that problem if i unplug one of left or right channels.

Is there any filter replacement that isnt based on a String table?

One that has at least 5 filter types, cut, rez, in and outs?

Some verry stable version thought:)

(using IL's FL Synthmaker)

oddson · 13 Aug 2008

Matthieu wrote: Is there any filter replacement that isn't based on a String table?

Pardon? 😕

malc · 13 Aug 2008

Matthieu,

Are you using an old version? This was corrected in 1.0.10.

Malc

Matthieu · 4 Mar 2010

Wow :P Way later after the post beginning XD .. I still get the same error code today!

I guess it hasnt been solved like stated in the lastest message.

I am using FL version of synthmaker... Mayby it isnt as up to date as the official release?

My problem is still caused by the use of many effects, 16, then 8, 2 times, one line for L one for R, and when all those code components i get the message "CInstDirectSoundOut::Compile> error: code>40000.

I tried to merge together all 16 firsts by simplification so the new flow line is one code component with 16 ins 16 outs and 16 tuning ins, routed to 8 effects. So it hasnt anything to do with amount of code component.. Anyway why would it at all XD

The effects are of many lines... Afterall it's just delays and filters. I was dealing with the one-piece delays that arent modifiable, but as they wirk with int only, it wasnt good sounding enough, so i tried coming backward and see what went wrong... I guess there's nothing to do without help. So Help. (i mean it... Don't just read 3 words and answer something totaly outside the subject like most comon humans usualy do, please! :P )

>>Hey TomC lol :P ça fait cool de voir des gens qui parlent Français de temps en temps sur cet internet remplit d'internationnaux anglophones XD HAHA Je te serre la patte de bien loin! à+

oddson · 4 Mar 2010

I believe there was a very recent fix in the beta update for the main version for this error... I believe it occurs when the code module uses more RAM than an arbitrary maximum. He's upping the limit by a factor of five and considering how to make it unlimited.

Perhaps Malc will be doing a new build for FLSM soon.

Matthieu · 4 Mar 2010

Very interresting words here 🙂 I noticed that the thread's error code and mine are diferant - - anyway XD

I was wondering how comes the synth ends up taking a whole gigabyte of ram... Usual synths even the heaviest usualy are way more gentle on the ram... I know that it has nothing to do with loaded images, neither most modules... But something somewere ends up taking heavy amounts of ram... My synth takes 7% of 2x2ghz right after loading, without playback, and climbs up to 20 or 50% deppending on what components are on,, Then, if i load the filters, it reaches 89% and start clipping, and short peaks reaches 100%... The synth itself is under 5Mb... mayby in order to work, it has to decompress or something?

And i use oscillators that i made with sin1(X) that is the only way i found to oscillate the frequency yet... And it is heavy on CPU too... But my synth is very beautiful :P (in truth, it's currently in a messful halfway to its new interface)... It has a year old now and passed throught so many malfunctions XD

Doing it all on my own with instinct, hear and look... It's like the second time i come here with questions :P

Yet i realy whish the SM team and Malc (I have never read about him.. Is he the one behind the SM masterpiece?) I realy whish them all the soul and luck to bring nice surprises soon :P 🙂

Mayby as i got your attention, whitch makes my joy, i would do good to ask a few questions..

What factors are to be kept in sight when using the code component in order to save on ram and cpu? Like, does "0.33" use less power than "1/3"? And does the amount of characters and lines have a great influence too?

I must figure out a way to use all objects planned in the synth's desing with saving on weight...

MichaelBenjamin · 4 Mar 2010

if the cpu climbs up like that i would guess you have some denormals in the filters or somewhere else. you can look for this with a number readout, usually if the numbers get into regions of e-45 after release you have denormals and your cpu usage will be up 50% and more. to solve this you can either insert low level noise at the start of your chain right at the oscillators, or individually add some small number at each problematic modules (usually delays with feedback and filters). but there are also threads here that can explain it further.

and a division uses about 4x as much cpu as a multiplication, so when you have a static division like 1/3 you can easily save that by writing 0.333333, like you said.

Matthieu · 4 Mar 2010

I just noticed that the bug only occurs with monophonic plugs. After i tried moving the effect behind the mono chain, it stopped doing the error... ( Very recent, still need testing and check ups but i quite belive it is fine now)