LTTng Execution traces for ten Phoronix benchmarks (part 3)

The  network-loopback.zip2 file contains the second part of the LTTng execution traces for the network-loopback Phoronix benchmark.

The traces have been generated on a desktop machine with a x86-64 Xeon E3-1225@3.20GHz processor,  32GB of memory,  Gigabit Ethernet connection and  SSD storage

Each zip file contains 32x3 traces : 3 tracing configurations, 32 runs per configuration.

- the all-events configuration traces the kernel.

- the libc configuration traces  memory-related function calls (malloc, free, ...)

- the perf-trace configuration traces  the hardware counters for memory accesses (L1-dcache-loads, L1-dcache-stores) and for instructions (Instructions).

To obtain the full set of execution traces for the benchmark, you should also download the file at LTTng Execution traces for ten Phoronix benchmarks (part 2) (https://doi.org/10.5281/zenodo.437179) and concatenate the two files (cat network-loopback.zip1 network-loopback.zip2 > network-loopback.zip)

The traces for the nine other Phoronix benchmarks, captured on the x86 machine, can be downloaded from https://doi.org/10.5281/zenodo.437170.

The file juno.zip contains the traces generated on a Juno board ({http://www.arm.com/products/tools/development-boards/). It has one dual core Cortex-A57 processor, one quad core Cortex-A53 processor, 8GB of memory,  Gigabit Ethernet connection and  SSD storage. The traces correspond to one run per configuration (all-events, libc, perf-trace) for the ten Phoronix benchmarks (compress-gzip, ffmpeg, iozone, network-loopback, phpbench, pybench, ramspeed, scimark2, stream, unpack-linux). In total there are therefore 3x10 = 30 traces.