Examples
Examples
Try the starting example
- Requests: name, dependencies joined by | or none, latency ms, size kB, Mbps
- HTML,none,100,10,8 CSS,HTML,20,20,8 JS,HTML,30,40,8 API,JS,100,0,8
- Concurrent request slots
- 2
- Simulated completion time (ms)
- 280
- Unlimited-slot dependency critical path (ms)
- 280
- Request start and finish times (ms)
- HTML: 0–110; CSS: 110–150; JS: 110–180; API: 180–280
- Total request service time (ms)
- 320
Try another scenario
- Requests: name, dependencies joined by | or none, latency ms, size kB, Mbps
- HTML,none,100,10,8 CSS,HTML,20,20,8 JS,HTML,30,40,8 API,JS,100,0,8
- Concurrent request slots
- 1
- Simulated completion time (ms)
- 320
- Unlimited-slot dependency critical path (ms)
- 280
- Request start and finish times (ms)
- HTML: 0–110; CSS: 110–150; JS: 150–220; API: 220–320
- Total request service time (ms)
- 320
Calculation method
Calculation method
Simulated completion time (ms)
Run ready requests in input order when slots are free; each request consumes latency + size kB × 8 / Mbps milliseconds.
Unlimited-slot dependency critical path (ms)
Longest dependency-chain duration with unlimited parallel requests.
Request start and finish times (ms)
Release dependants at parent completion, then use the first available slots.
Total request service time (ms)
Sum all request durations, excluding waiting.
Symbols and units
- a
- Requests: name, dependencies joined by | or none, latency ms, size kB, Mbps
- b
- Concurrent request slots
Use the units shown. A list keeps the order in which you enter its values. The symbol ^ means “raised to the power of”.
Assumptions & limits
Assumptions & limits
- A dependency-aware fixed-duration list schedule with shared concurrency slots. Transfers retain their entered per-request speed even in parallel; DNS, handshakes, multiplexing, parsing and shared bandwidth are not inferred. The limited schedule is not claimed optimal.
Are these inputs private?
These calculations run in your browser. A shared link contains the values you choose to share. Only questions submitted to the separate AI finder are sent to OpenAI.
Further reading
W3C: resource request timing concepts ↗
Background reading; this does not imply endorsement or independent review of SumFizz.