←back to thread

Google's “Director of Engineering” Hiring Test

(www.gwan.com)
1764 points fatihky | 3 comments | 13 Oct 16 15:15 UTC | HN request time: 0.527s | source
Show context
Tempest1981 ◴[13 Oct 16 16:04 UTC] No.12701678[source]▶
On #9, is it 16-bit or 64-bit values? Sounds like you may have been answering a different question. Isn't the standard answer to use a 65536 entry lookup table?

Edit: oh, are you reinterpreting the array of 10000 16-bit integers as 2500 64-bit integers? But then what operation do you use on each?

Either way, if you find yourself arguing with the recruiter, it's probably a bad sign.

replies(2): >>12702694 #>>12703171 #
landhar ◴[13 Oct 16 19:02 UTC] No.12703171[source]▶
I thought the reason it was generally accepted that the 8-bit lookup table is optimal is because it can fit in the L1 cache.
replies(1): >>12703666 #
gcp ◴[13 Oct 16 20:13 UTC] No.12703666[source]▶
(Very) embedded developer speaking here. What's this L1 cache you speak of?
replies(1): >>12706639 #
1. joss82 ◴[14 Oct 16 07:43 UTC] No.12706639[source]▶
It's level 1 CPU cache, the fastest (and smallest) cache.

It caches both instructions and data.

https://en.wikipedia.org/wiki/CPU_cache

replies(2): >>12707477 #>>12722595 #
2. phpnode ◴[14 Oct 16 11:48 UTC] No.12707477[source]▶
GP is making the point that not all CPUs have a cache.
3. ummonk ◴[17 Oct 16 03:54 UTC] No.12722595[source]▶
It was a joke. Some embedded processors lack a cache.