Phoronix posted an article comparing OpenSuSE leap 15, leap 42.3 and tumbleweed.
I don’t have these operating systems on my benchmark system, but make the following notes about what the benchmarks are measuring.
These comparisons were on an Intel i9 system with 18 cores and 16 GB memory. These tests are of one OpenSuSE version vs. another rather than comparisons with Windows or other distributions.
The following table provides an summary
|Benchmark||Phoronix observations||My observations||Analysis|
|fio||Random and sequential read/write are all faster||Latency is key issue for I/O not CPU time; can be affected by spectre/meltdown so good it is faster.||Analysis|
|sqlite||leap 15 slightly faster||Database in one big file with file locking, testing kernel performance.||Analysis|
|blake2||leap 15 has slower performance||quick running test, less than 1/2 second. Single-threaded||Analysis|
|go-benchmark||http much faster, json slightly faster, build slower and build slower||Mostly reflects the different versions of go used.||Analysis|
|x264||leap 15 is slower||On_Cpu 15% with frontend stalls and IPC of 1.3||Analysis|
|stockfish||leap 15 is a hair faster||frontend stalls and speculation misses are key issues otherwise On_Cpu 100%||Analysis|
|ebizzy||leap 15 is a hair faster||mostly a test of kernel memory management||Analysis|
|build-php||leap 15 is a fair amount slower||newer gcc versions compile slower.||Analysis|
|phpbench||leap 15 is slightly slower||single-threaded; high IPC with ~15% frontend and ~15% backend stalls||Analysis|
|git||leap 15 is slower||mostly a test of the filesystem||Analysis|
|compress-zstd||leap 15 is faster||single threaded with many backend stalls and an IPC of 0.71||Analysis|
|tinymembench||leap 15 memcpy is faster||single threaded memory test of varying sizes||Analysis|