Rewrite the cpp example to do the same thing as the nim example

[det]

The current C++ example is unnecessarily complex and slower than a naive implementation. This patch makes the C++ example work the same way as the Nim example by using a vector with 1 byte per bool and using globals.

[lgeek]

I'm not Jonathan so I can only provide a datapoint, but for me this is 15% slower than the current CPP implementation on ARM (Cortex-A15). Access to globals is slower and the template-based current implementation passes the visited bitset by value for small graphs (it's a single half-word), including the one used for benchmarking, which is faster than accessing as a global or passing by reference.

[logicchains]

I've found the same; on ARM it's around 3.5 seconds for this
implementation, compared to around 3.0 for the previous one.

On 3 January 2015 at 09:39, Cosmin Gorgovan notifications@github.com
wrote:

I'm not Jonathan so I can only provide a datapoint, but for me this is 15%
slower than the current CPP implementation on ARM (Cortex-A15).

—
Reply to this email directly or view it on GitHub
#60 (comment).