Re: [eigen] Help on solving a race condition

[Ilja Honkonen <ilja.honkonen@xxxxxxxxxxx>]

> > static std::ptrdiff_t m_l1CacheSize = 0;
> > static std::ptrdiff_t m_l2CacheSize = 0;
> > static bool initialized = false;
> > if(!initialized)
....
> There is no guarantee that the "initialized = true" gets seen by another
> thread after the update to "m_l1CacheSize" or "m_l2CacheSize" gets seen
> by that thread. The order of the stores can be differently perceived by
....

There is also quite a good explanation of these things at 
http://en.cppreference.com/w/cpp/atomic/memory_order
Pre C++11 behavior seems to be memory_order_relaxed but without atomic 
variables so pretty much anything goes. A concrete example:

t1 sees
initialize == false
and does
m_l1CacheSize = manage_caching...
....
initialized = true;

"after" that t2 does
m_l1CacheSize = 0;
....
but already sees
initialized == true;

resulting (possibly) in obviously wrong sizes of 0.

Ilja