Linux kernel developers have been working tirelessly to squeeze more performance out of IO_uring and the block / I/O code in general . IO_uring lead developer Jens Axboe who also serves as the Linux block subsystem's maintainer (among other roles and major contributions over the years) has used his system as a baseline for evaluating such kernel improvements. He's now moved to using AMD Zen 3 while sticking to Intel Optane storage and is seeing a mighty speed boost out of AMD's latest processors.

Rather than talking in the 3.5~3.8M IOPS per-core range for evaluating kernel improvements, he is now at over 5 million IOPS per core by upgrading to AMD Zen 3 hardware.

His system up to this point for doing the major Linux I/O testing was an AMD Ryzen Threadripper 3970X with Intel Optane Gen2 storage. He has now moved to using an AMD Ryzen 9 5950X with Intel Optane Gen2.

The Linux IOPS per-core performance for his new baseline is now well above five million.... 5.1M IOPS is what he is seeing out of his new configuration at this point in Linux kernel development. This upgrade to AMD Zen 3 is now pushing the limits of the underlying Optane storage itself.

Answer is 5.1M IOPS, which is also the limit of the device. Stock 5950X build, not overclocked.

IOPS=5121792, BW=2500MiB/s, IOS/call=32/31, inflight=(128 128)
IOPS=5121792, BW=2500MiB/s, IOS/call=32/32, inflight=(128 101)

New baseline. #io_uring #amd #optane https://t.co/ujDhwxX3Vx

— Jens Axboe (@axboe) September 30, 2021