Competitive programming

Small inputs: no difference

At 1,000 numbers every version finishes in well under a tenth of a second. The plain input() loop is fine, and simpler to write.

Large inputs: a big one

At a million numbers, reading everything at once is about 3.5 times faster in Python. That can be the gap between passing and a time-limit failure.

The surprise

On this machine's older compiler, Python's bulk read beat C++ cin even with synchronisation turned off. The language matters less than the reading method.

Why C++ cin is slower by default

By default C++ streams stay in step with C's input functions. cppreference explains the switch: "If the synchronization is turned off, the C++ standard streams are allowed to buffer their I/O independently, which may be considerably faster in some cases."

In our run, turning it off cut 1.434 s to 0.840 s.

The lesson students keep

Measure on input as large as the constraints allow, not on the sample. Simple code is right for small inputs; switch to bulk reading only when the numbers say it matters.

Our timings depend on the machine and compiler, so students repeat them on their own.

Reviewing an attempt

After a timed set, each learner explains one wrong submission: what the judge said, what they guessed, and what the fault really was.

Writing the fault down turns one mistake into a rule they keep.

Not every failure is a slip

Sometimes the algorithm itself is too slow or wrong, and no amount of tidying will fix it. Telling the two apart quickly is a skill that grows with practice.

The review asks which kind each failure was.