Thanks Mike,

I was hoping that all that hardware would cure a problem I've been having... I've made an indicator which is a
Bollinger% B3 of t4tcum. (Boll%B3 was published here in one of Gabryele's threads).
If set for cobc false, ninja crashes... I thought more computer power and more memory would cure that.

What do you think?

superpete 

Thanks Mike, I was hoping that all that hardware would cure a problem I've been having... I've made an indicator which is a Bollinger% B3 of t4tcum. (Boll%B3 was published here in one of Gabryele's threads). If set for cobc false, ninja crashes... I thought more computer power and more memory would cure that. What do you think?

my "old" t4tCUM... have little problemsto give permanentvalid values to the external data series... when COBC = "false", then its give during the internal sync state moments with invalid and not accessable values...
-> many indicators dont check/validate the current array values...

I have this now corrected by continue using the "old" values from last bar as start value of any new bar, when COBC = false. This give everytime accessable in combination valid values and if "insync" all tickvalues well be used for correct new tickbased calculations.

=> for using other indicators its very important the use of a only one time on StartUp initialized variable!
-> in my t4tCUM thread i have published a demo source(with a update of t4tCumUD->v15) for a indicator with external using the t4tCumUD15... this show this coding style and work better as t4tCumUD13...
-> t4tCumUD2 well come this month

Thank you Timetrade, I didn't know this indie was yours... I'll wait for the newer version