[Arm-netbook] Rpi vs cortexA8

[Gordan Bobic]

On 06/03/2012 12:11 PM, Enrico wrote:
> On Sun, Jun 3, 2012 at 12:45 PM, Gordan Bobic<gordan at bobich.net>  wrote:
>> On 06/03/2012 11:42 AM, Enrico wrote:
>>> On Sun, Jun 3, 2012 at 12:17 PM, Tsvetan Usunov, OLIMEX LTD
>>> <usunov at olimex.com>    wrote:
>>>>> http://www.engadget.com/2012/06/01/raspberry-pi-impressions-the-35-linux-computer-and-tinker-toy/
>>>>>
>>>>> "And, in case there was any doubt in your mind about how painful just
>>>>> web browsing could be on this thing, we ran SunSpider (which also
>>>>> pinned the CPU) and got a score of 44,230. By comparison, our OG Droid
>>>>> (which is clocked at just 550MHz, but has the advantage of being a
>>>>> Cortex A8 chip) pulled a 11,188."
>>>>
>>>> so ARM11 @700Mhz is 4 times slower than Cortex-A8 @ 550Mhz?
>>>>
>>>> http://en.wikipedia.org/wiki/List_of_ARM_microprocessor_cores
>>>>
>>>> says 965 DMIPS @ 772 MHz for ARM1176 and 2.0 DMIPS/MHz for Cortex A8, so
>>>> Cortex A8 @ 550 should give 1100 DMIPS or about 15% faster
>>>
>>> Just speculating, but one of the things that could give a speed boost
>>> in the cortex-a8 is the neon unit.
>>
>> Not if the same software is used in both instances since if it's running
>> on ARMv6 it won't be issuing NEON instructions.
>
> So they used the same debian squeeze (that i think is built for armv5)
> on rpi and OG Droid and run SunSpider?

I would assume so. Otherwise the comparison is largely meaningless anyway.

Gordan