On Thu, Jan 19, 2012 at 9:56 AM, Dimiter 'malkia' Stanev
<malkia@gmail.com> wrote:
Smaller numbers are bad.
I'm think Mike mentioned that ARM's hard-floating point is not used fully (vfp, or what was the other stuff - neon?).
Neon is their SIMD instruction set, not sure if it has anything to do with FP implementation.
On 1/18/2012 7:42 PM, David Burgess wrote:
Thank you. One question are smaller numbers good or bad?
On Thu, Jan 19, 2012 at 10:10 AM, Michal Kottman<k0mpjut0r@gmail.com> wrote:
On 18 January 2012 23:54, David Burgess<dburgess@gmail.com> wrote:
I have been doing some testing on an Marvell Armada 510.
The thing that interests me is that the 510, has a vfp with good
double performance.
(So, I immediately thought that Lua is going to like this.)
The question I have is whether there is any Lua benchmarks which I
could use to compare Lua on a Armada 510 vs other processors?
One such benchmark is SciMark 2, ported to Lua by Mike Pall. It is
available here: http://luajit.org/download/scimark.lua
If ran under LuaJIT, it takes advantage of the FFI for greater efficiency.