OS & Concurrency

After C and pointers, operating systems and concurrency are among the most reliable topics at Intel, and nearly every track gets some of it. Firmware candidates meet the ideas as interrupts and shared buffers. AI software candidates get threads and Python's GIL, and systems software and driver roles go deepest of all.

Most of it is talked through on a whiteboard, but a few topics turn into good code. A deadlock becomes a cycle in a graph of which thread waits on which. A Premium question becomes a counter you can check by hand against a reference string. The cache shows up in a matrix multiply, where loop order alone decides how often it misses.

Writing the mechanism helps with the follow-up questions. When the interviewer asks which deadlock condition you would break, the edges are already on the screen. When they ask whether more memory can ever mean more faults, you can run FIFO on 1,2,3,4,1,2,5,1,2,3,4,5 with three frames and then with four, and read off nine faults and then ten.

The Intel angle

Intel builds the hardware these simulations model, so the follow-ups go down to the silicon. A Premium question count leads to the Accessed bit the MMU sets in each x86 page-table entry, because no CPU can afford exact LRU. The matrix multiply leads to cache lines, the size of L1, and the tiled kernels in oneAPI's math library. Plan to finish the code with time left, then spend that time on what the real machine does differently from your model.

IntelPremium lesson · Coding round

The 4 Intel problems are below.

1. 1A Premium questionCoding · Medium · Frequently asked
2. 2A Premium questionCoding · Medium
3. 3A Premium questionCoding · Medium

+ 1 more in this lesson

Each with hints, the follow-up and a tested solution.

Get the Intel questions

$49 a year · all 100 companies · cancel anytime

From 20,000+ candidate reports, forum threads behind logins and write-ups in Chinese.

Read a free Intel lesson firstAlready a member? Sign in