Sub-microsecond timing on EC2 is way messier than I expected

Syntactic sugar of member function binding?

Fast GPU Linear Algebra via Compile Time Expression Fusion

atomic_queue benchmarks SMT vs no-SMT performance

Discussion: Using Python as a control plane for high-performance systems - Is the overhead of the C-API still the main bottleneck?

I tried compile-time heapsort in TMP. It basically became selection sort.

boost::container::hub ACCEPTED into Boost