3 ms·
I checked again with O_DIRECT and now I stand corrected. I didn't know that O_DIRECT could make such a huge difference. Thanks!
by jstimpfle 2y ago
I checked again with O_DIRECT and now I stand corrected. I didn't know that O_DIRECT could make such a huge difference. Thanks!
- jstimpfle 2y agoOops, O_DIRECT does not actually make that big of a difference. I had updated my ad-hoc test to use O_DIRECT, but didn't check that write() now returned errors because of wrong alignment ;-) As mentioned in the sibling comment, syncs are still slow. My initial 1-2ms number came from a desktop I bought in 2018, to which I added an NVME drive connected to an M.1 slot in 2022. On my current test system I'm seeing avg latencies of around 250us, sometimes a lot more (there a fluctuations). # put the following in a file "fio.job" and run "fio fio.job" # enable either direct=1 (O_DIRECT) or fsync=1 (fsync() after each write()) [Job1] #direct=1 fsync=1 readwrite=randwrite bs=64k # size of each write() size=256m # total size written