Description:

from playing around with bugs 124577 and 129753 i have the idea that the use of 'threaded' and / or 'openCL' computing is somewhere hardcoded to step in up from 100 rows with identical formulas, (former versions of calc had an option switch to choose the value), 

i assume that value is choosen to balance the perfomance benefit from parallelizing against the effort to divide and dispatch the jobs, 

but why 100? that's a logical threshold value for humans, but computers rely on powers of two? 

thus i think either 64 or 128 will fit better as decisive point whether to use or avoid parallelizing? or maybe other powers of two? 

as this area needs a workthrough anyway, to get out the autocalculate errors, it might be appropriate to process that in one wash? 

just an idea ... 

reg. 

b. 

Steps to Reproduce:
1. no real 'bug' to be seen, 
2. just an idea about meaningful design, 
3. analysis materil may be found in mentioned bug reports, 

Actual Results:
an arbitrary 'humane' value is established

Expected Results:
might be an 'computer fair' value would fit better? 


Reproducible: Always


User Profile Reset: No



Additional Info:
Version: 6.4.0.2 (x64)
Build ID: 08d19fecdc7a2298d051e19cfdb7c35544855fc3
CPU threads: 8; OS: Windows 6.1 Service Pack 1 Build 7601; UI render: default; VCL: win; 
Locale: de-DE (de_DE); UI-Language: en-US
Calc: CL