On 27/04/2018 17:15, Martin Gregorie wrote:
> On Fri, 27 Apr 2018 12:55:15 +0100, Gareth's Downstairs Computer wrote:
>
>> On 27/04/2018 11:30, Jan Panteltje wrote:
>>> On a sunny day (Fri, 27 Apr 2018 11:46:30 +0200) it happened Andreas
>>> Neumann  wrote in :
>>>
>>>> Andrew Gabriel wrote:
>>>>
>>>>> Why do you want a 64 bit build?
>>>>
>>>> Because 64 bit are twice as good as 32 bit?
>>>>
>>>> SCNR
>>>
>>> Not even in theory, I did some web searching for that question myself,
>>> wanted benchmarks.
>>>
>>> started reading here:
>>>    https://raspberrypi.stackexchange.com/questions/49466/raspberry-pi-3-
> and-64-bit-kernel-differences-between-armv7-and-armv8
>>>    https://www.raspberrypi.org/forums/viewtopic.php?
> f=66&t=138385&sid=9aaac4bc075db741ec8137e188b2956e
>>>
>>> Things may even slow down.
>>>
>>>
>> Yes, but will improve all the time. In some ways like the specification
>> for the processing of GSM mobile communication. ISTR that the needed
>> technology was not around at the time of specifying, but every
>> confidence was that it would appear.
>
> In this case the only likely effect is that compiled executables will be
> a bit bigger because a 64 bit pointer is 8 bytes compared with 4 for a 32
> bit pointer and will run a bit slower because (I think) the 64 bit
> pointer will need two fetches compared with one for a 32 bit pointer.
>
> And, seeing that a 31 bits is a bigger address space than there is RAM on
> an RPi3 (4GB  address space, 1GB RAM less that used by the GPU and screen
> buffer), using 64 bit pointers gains you nothing.
>
>


1Gb now, 32GB in 5 years' time when the 64 bit is the only RPi still
in production?

Remeber Billy Gates? 640k is more than enough ...

--- SoupGate-Win32 v1.05