Windows 10 WSL vs. Docker on Windows 10 vs. Bare Metal Linux Performance
With the recent Windows 10 Fall Creator's Update there were some improvements to the Windows Subsystem for Linux (WSL) particularly around boosting the I/O performance (though further WSL performance work is coming), so this week I've been carrying out some fresh benchmarks of Windows 10 WSL with its openSUSE and Fedora options. For additional perspective I also compared the performance to running benchmarks with Linux containers on Docker under Windows 10 and lastly the "bare metal" Linux performance.
Available currently from the Windows 10 Store with the Fall Creator's Update are Ubuntu 16.04 LTS, openSUSE Leap 42.3, and SUSE Linux Enterprise but unfortunately Fedora still has not yet appeared. For testing of Windows 10 WSL I was using both Ubuntu 16.04 LTS and openSUSE Leap 42.3.
After testing Windows 10 WSL with Ubuntu and openSUSE, I then turned to testing them using the Ubuntu 16.04 and openSUSE Leap 42.3 via Docker stable running on this Windows 10 installation. At this point I also tossed in Intel's Clear Linux distribution given its generally leading out-of-the-box Linux performance and curious in how this translated to running within Docker on Windows.
Following those Windows 10 WSL and Docker benchmarks, next I then tested Ubuntu 16.04 / openSUSE / Clear Linux when installing each of them "bare metal" on the system to see how their raw Linux performance is without any virtualization or the like.
All of this benchmarking took place from the same Intel Core i9 7980XE system with ASUS PRIME X299-A motherboard, 4 x 4GB DDR4-3200 system memory, and 120GB Corsair Force MP500 NVMe SSD.
All of these benchmarks of the different Linux environments were carried out in a fully-automated and reproducible manner using the open-source Phoronix Test Suite benchmarking software.