A transfer estimate becomes misleading when a gigabit rate is divided into a gigabyte total without conversion. The safest method is to convert the file total to bytes and the rate to bytes per second, then divide. Keep the origin of the rate with the result.
This is arithmetic, not a benchmark of a named drive. A speed entered from your own measurement is your observation; the calculator does not relabel it as a test conducted by this site.
Work the units all the way through
| Inputs | Conversion | Time |
|---|---|---|
| 10 GB at 100 MB/s | 10,000,000,000 B ÷ 100,000,000 B/s | 100 s |
| 10 GiB at 100 MiB/s | 10,737,418,240 B ÷ 104,857,600 B/s | 102.4 s |
| 10 GB at nominal 1 Gbps | 10,000,000,000 B ÷ 125,000,000 B/s | 80 s theoretical |
The decimal and binary examples deliberately use different divisors. Preserve the unit capitalization: b means bits and B means bytes in these rate labels.
Choose the correct calculation mode
A nominal interface rate gives a theoretical lower bound if it is the only limit considered. It does not include the actual workload's device and protocol constraints. If you assume 80% usable efficiency for the 1 Gbps example, the estimate is 10,000,000,000 ÷ (125,000,000 × 0.8) = 100 seconds. Label 80% as your assumption.
A known effective rate has already described usable throughput under its observation conditions. If it is 100 MB/s, the first example is 100 seconds. Applying an automatic 80% factor again would double-discount the rate and produce a different, unjustified estimate.
Keep the workload with the number
Record whether the rate concerns reading or writing, the approximate file count, total bytes and the circumstances under which the effective rate was obtained. A large sequential transfer and many small files need not behave alike. Caching, software and device state can affect the actual elapsed time.
Do not infer a rate from a device name. A manufacturer's “up to” claim is a conditional specification, not your observed sustained copy speed.
Make the estimate reusable
The tool's copy and print outputs include size, rate, converted units, mode and efficiency assumption. Save that full record rather than only “about two minutes.” Zero total size can return zero seconds, while the rate must stay positive. An estimate whose units and assumptions travel with it is easier to revise when better data becomes available.