tuxalin
I dont have anymore the exact numbers, usually when you have around 1000 commands/calls you won't really feel a difference, in that case even using std::sort is fast enough and there's no point in using radix sort, however when you go over 10k commands/calls you'll feel it (can be up to a msec on the dispatch side).
You can check this great post that provides numbers (see part 4): https://blog.molecular-matters.com/2014/11/06/stateless-laye...
The performance should be very similar and it even goes further by adding thread local storage to reduce false sharing even more, which I didn't add but should be considered as it's easy to do.
Regarding the implemention, I presumed you asked about the lock-free logic, apart from m_currentIndex which is used with the m_commands vector, the LinearAllocator also has to be lock-free, see it's alloc function. The allocator is used for storing the command packets (along with their command data) and also their auxiliary data (if they have that).
Another implementation detail of the allocator is that it has two modes, an auto-alignment mode which requires two atomic operations (load acquire and compare_exchange_weak) and will be slower but lower memory usage, therother mode uses the alignment of the CommandPacket class which will be faster (one fetch_add) but consume more memory.
And another detail is the CommandPacket uses a intrusive list approach, that's how you actually chain multiple commands together and do an indirect function call (instead of virtual or other blasphemies) with dispatch of the command.
Hope that clears some things up about it.