llvm-project/libc/benchmarks/distributions

Size distributions for memory functions under specific workloads

This folder contains a set of files that are included from libc/benchmarks/MemorySizeDistributions.cpp.

Offloading this data to individual files helps

• C++ editors (large arrays are usually not well handled by editors),
• and allows processing data by other tools to perform analysis or graph rendering.

Format

Most filenames are of the form {MemoryFunctionName}{WorkloadID}.csv. They contain a single line of comma separated real values representing the probability that a particular size occurs. e.g.

• "0,1" indicates that only the size 1 occurs,
• "0.5,0.5" indicates sizes 0 and 1 occur with the same frequency.

These files usually contains sizes from 0 to 4096 inclusive. To save on space trailing zeros are discarded.

Workloads

As identified in the automemcpy paper:

• GoogleA <-> service 4
• GoogleB <-> database 1
• GoogleD <-> storage
• GoogleL <-> logging
• GoogleM <-> service 2
• GoogleQ <-> database 2
• GoogleS <-> database 3
• GoogleU <-> service 3
• GoogleW <-> service 1

Uniform384To4096 is an additional synthetic workload that simply returns a uniform repartition of the sizes from 384 to 4096 inclusive.

Note

Except for GoogleD, all distributions are gathered over one week worth of data.