llvm-project
Size distributions for memory functions under specific workloads
This folder contains a set of files that are included from
.
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
. They contain a single line of comma separated real values representing the probability that a particular size occurs. e.g.
indicates that only the size"0,1"
occurs,1
indicates sizes"0.5,0.5"
and0
occur with the same frequency.1
These files usually contains sizes from
to
inclusive. To save on space trailing zeros are discarded.
Workloads
As identified in the automemcpy paper:
<->GoogleAservice 4
<->GoogleBdatabase 1
<->GoogleDstorage
<->GoogleLlogging
<->GoogleMservice 2
<->GoogleQdatabase 2
<->GoogleSdatabase 3
<->GoogleUservice 3
<->GoogleWservice 1
is an additional synthetic workload that simply returns a uniform repartition of the sizes from
to
inclusive.
Note
Except for
, all distributions are gathered over one week worth of data.