2) Statistically better stability(for a given number of units)
There's a reason why certain CPUs will overclock better than others of the same bin, and why the manufacturer does not sell them at a higher bin, for a CPU manufacturer it all depends on yields. For a device manufacturer, it all depends on what speed will give the best stability for a large number of units, while not compromising performance.
shorter lifecycle and higher risk of data corruption. Data corruption is IMO worst because i think that we would have lots of really really weird bug reports in bugzilla. Data corruption may seem like bugs in applications and sometimes itīs really hard to spot and it can lead to lotīs of wasted time on bug hunting.
Reasons:
a) They can buy cheaper components to build same model.
b) They can produce new models cheaper.
(battery? stock battery life is 5h, even less, under moderate stress. If you work more than 4h59min/day, you need two-three bats or a charger int your pocket... so no matter if OC under heavy stress longs 2h30min, the 5800 Xpress Charger does the job... as with stock kernels)
the N800 and 810 kernels for diablo were "overclocked" compared to the chinook kernels.
maybe nokia will release an official higer clocked kernel.
maybe they wont.
stability is the most important thing for most....
Even today, there is a higher speed version of OMAP3530 (a chip which is similar to OMAP3430 used in N900), which can run stable at 720MHz instead of standard 600MHz. But this higher speed version of OMAP3 chips became available only recently, and I guess it was too late for N900: http://newscenter.ti.com/Blogs/newsr...ns-228688.aspx
