Sung M. KimAnthony has shared wonderful tips on interviewing at Google.
Within the fourth tip, Brainstorm solutions and their time/space complexity, Anthony stressed on "space-time tradeoff principle", which I was vaguely familiar with.
While attempting to understand it, I drew a sketchnote.
As it's not sharable, I came up with emoji version.
Space-time Tradeoff Table
| ⏰🔽💽🔼 | (faster + more space) | vs | ⏰🔼💽🔽 | (slower + less space) |
|---|---|---|---|---|
| 🙄📑 | (Lookup table) | 🔃💻 | (Recalculation) | |
| 💾 x 1 | (Uncompressed Data) | 💾 x 0.5 | (Compressed Data) | |
| 🏪🖼 | (Stored Image) | 🔃📜 | (Re-render) | |
| ➿➡➰ | (Loop Unrolling) | 👨💻 x 0.5 | (Smaller code) |
Gist of each tradeoff
🙄📑(Lookup table) vs. 🔃💻 (Recalculation)
Saving computed data in memory saves time but increases the space usage.💾 x 1 (Uncompressed Data) vs. 💾 x 0.5 (Compressed Data)
Faster to access uncompressed data as no decompression step is needed but requires more space.🏪🖼 (Stored Image) vs. 🔃📜 (Re-render)
Storing image in cache is faster than re-rendering but requires more space (in disk/memory).➿➡➰ (Loop Unrolling) vs. 👨💻 x 0.5 (Smaller code)
Loop unrolling can optimize execution speed at the cost of increased binary size.
It might be a bit too much of emojis but helped me to understand a bit better what tradeoffs are available.
Feel free to share your version of table or sketchnote 🙂
