Disk I/O Throughput Calculator
Convert sustained IOPS, block size and latency into MB/s, queue depth and time per I/O.
Throughput by IOPS and block size
Throughput is IOPS multiplied by block size, so the same device moves far more bytes with larger I/Os — but a bigger block usually costs more latency per operation.
| Block | 1,000 IOPS | 10,000 IOPS | 100,000 IOPS |
|---|
Throughput in megabytes per second is the operation rate times the bytes per operation, with MB meaning 1,000,000 bytes and MiB meaning 1,048,576 bytes. Queue depth follows Little's law: the average number of outstanding I/Os is the completion rate times the average latency, so 10,000 IOPS at 0.2 ms need about two requests in flight, and a queue that is too shallow caps the IOPS the device can sustain at that latency. The figures describe a steady state — they do not model bursts, read/write mixes, sequential versus random access, RAID write penalties, controller caches, filesystem overhead or network storage. Everything is calculated in your browser; the page does not benchmark, store or upload anything.