Very basic HTTP load generator that avoids coordinated omission.
./hyperflood <host name> <port num> [-r <requests per second>] [-t <thread count>]
Example quickstart - to start a dummy HTTP server:
make server
./server
In another terminal, start the client:
./hyperflood "127.0.0.1" 8081 -r 10
This loads your server with 10 requests per second; no more, no less. If your server cannot handle the load in a given time slice (e.g. it could only process 5 instead of 10), the unprocessed requests will accumulate and be resent in later time slices, and the latency will reflect that.