Re:BeagleBoard is nice, but the PandaBoard is better (dual C...
by pmjordan on Tuesday October 05, 2010 @04:46AM
(#1759041)
That looks like a very impressive piece of kit. I was planning to get a Hawkboard Lite for a project, but might wait for this to become available instead, if the price is decent. 8 times the RAM!
[ Parent
| on HN ]
Re:BeagleBoard is nice, but the PandaBoard is better (dual C...
by pietrofmaggi on Tuesday October 05, 2010 @05:07AM
(#1759071)
The different amount of RAM is the last think to notice here. The OMAP L138 on the hawkboard is an ARM9 (+ C674xDSP) at 375/456MHz, the OMAP4 of the pandaboard is a dual core Cortex-A9 at 1GHz.
They are completely in two different class. I too hope for the price.
Re:BeagleBoard is nice, but the PandaBoard is better (dual C...
by pmjordan on Tuesday October 05, 2010 @05:42AM
(#1759121)
The different amount of RAM is the last think to notice here.
That depends mainly on the use you have in mind. For my purposes, the Hawkboard's CPU would probably be fine. But yeah, the Pandaboard is quite a beast - twice the computing power of an iPad, if I'm not mistaken. (not counting the DSP)
Re:BeagleBoard is nice, but the PandaBoard is better (dual C...
by codedivine on Tuesday October 05, 2010 @06:35AM
(#1759206)
More than twice I would say. iPad's A4 is a tweaked Cortex A8. Cortex A9 is out-of-order compared to in-order A8 and will likely be faster per clock.
Re:BeagleBoard is nice, but the PandaBoard is better (dual C...
by mansr on Tuesday October 05, 2010 @09:26AM
(#1759563)
The A9 is readily 25% faster per clock than the A8 for scalar integer operations. Scalar floating-point is 5-10x faster depending on the exact operations. Vector instructions are mostly the same or slower than A8 due to lack of the limited dual-issue present in A8. I have yet to run a proper benchmark to compare the difference in vector performance in realistic situations.