Various benchmarks mostly for streams of interleaved stores, in support of the blog post What has your microcode done for you lately?.

Currently it only works on Linux, but I am interested in porting it to Windows. It should be enough to run make:

make 

Run with no arguments for usage info, as follows:

Must provide 1 or 3 arguments Usage:	bench TEST_NAME TEST_NAME is one of: interleaved	basic interleaved stores (1 fixed 1 variable) interleaved-pf-fixed	interleaved with fixed region prefetch interleaved-pf-var	interleaved with variable region prefetch interleaved-pf-both	interleaved with both region prefetch interleaved-u2	interleaved unrolled by 2x interleaved-u4	interleaved unrolled by 4x interleaved-sfenceA	interleaved with 1 sfence interleaved-sfenceB	interleaved with 1 sfence interleaved-sfenceC	interleaved with 2 sfences wrandom1	single region random stores wrandom1-unroll	wrandom1 but unrolled and fast/cheaty RNG wlinear1	linear 64B stide writes over one stream wlinearHL	linear with lfence wlinearHS	linear with sfence wlinear1-sfence	linear with sfence rlinear1	linear 64B stride reads over one region lcg	raw LCG test pcg	raw PCG test 

At a minimum you need to provide the test name from the list above.

If you provide only the test name, default starting and stopping sizes for the region are used (4 KiB to 512 KiB):

./bench interleaved 

Otherwise, you can provide your starting and stopping points as the 2nd and 3rd arguments, in KiB. The plots in the blog post all use 1 to 100,000 KiB as follows:

./bench interleaved 1 100000 

Values are rounded up to the next power of two, so 100,000 becomes 131,072.

The /scripts directory contains a bunch of .sh scripts that I use to generate the various plots. In particular, to generate all plots it should be as simple as running the all.sh script:

SUFFIX=new scripts/all.sh 

The SUFFIX here is appended to each plot name and indicates whether an old or new microcode version was used (the exact microcode revion is also automatically added to the plot title). The output appears in the /assets directory.

You can also run any of the individual plots that all.sh creates directly, e.g., scripts/rwrite-1-vs-2.sh will generate the first plot from the post. My default these individual scripts will pop up an interactive window with the plot, but you can write to a file by setting the OUTFILE environment variable. There are a variety of other variables you can set too, for example STOP=1000 scripts/rwrite-1-vs-2.sh will change the stopping point to 1000 KiB which results in much faster plot generation. You can take a peek at all.sh for some examples or other variables.

We support recording various Intel performance counters using pmu-tools' jevents library. To record events, set them in the CPU_COUNTERS environment variable, using the short name for as shown in the supported events table: