d********w
发帖数: 363 | 1 职位：performance engineer
1) 3-tier web application, analyze possible bottleneck
2) calculate power(int x, int y)
3) sorted array,find intersection,
4) jvm heap management, gc mechanism, block, c++/java difference
5) merge sorted array, how to optimize in large scale.
6) given a client and server, server transfer a large data file e.g. (10M)
, traffic is full in network, how to improve the performance. i.e. response
time reduce to half of original version.
7) Wordcount how to improve performance, multithread, parallel
8) many queue, one process VS. each queue has one process, throughput is the
same as 1, which one is better? poisson distribution?
9) Given large amount of data points, it records response time. How to
measure the system is available? (statistics: average, median, percentile)
10) how to measure the total response time when visiting a web page in
browser? (Track user behavior, analyze which step consume more time)
11) SSO, when someone login, it go to OAuth server and validade the user and
password. The goal is to achieve 10000 visit times/min, how to design?
12) Given some performance graph, CPU/load graph, Throughput/Load graph, ask
why the line are stable within the increasing of work load. | j********x
发帖数: 2330 | | z*********8
发帖数: 2070 | | p*****2
发帖数: 21240 | | d********w
发帖数: 363 | 5 要对系统方面有研究，熟悉IO cost, memory model, performance metrics:latency,
throughput, CPU load. concurrency, multithread, parallel system,这些还是要学
的挺多的
【在 p*****2 的大作中提到】
: 现在S挺缺PE的。
| r*****k
发帖数: 1281 | 6 考察的知识面挺广啊
感觉比Google facebook还要难。。 | p*****2
发帖数: 21240 | 7
太专了。没做过的很难回答好。中小公司很多都这样吧。
【在 r*****k 的大作中提到】
: 考察的知识面挺广啊
: 感觉比Google facebook还要难。。
| l*****a
发帖数: 14598 | 8 最近这么多面试？
快毕业了？
)
response
【在 d********w 的大作中提到】
: 职位：performance engineer
: 1) 3-tier web application, analyze possible bottleneck
: 2) calculate power(int x, int y)
: 3) sorted array,find intersection,
: 4) jvm heap management, gc mechanism, block, c++/java difference
: 5) merge sorted array, how to optimize in large scale.
: 6) given a client and server, server transfer a large data file e.g. (10M)
: , traffic is full in network, how to improve the performance. i.e. response
: time reduce to half of original version.
: 7) Wordcount how to improve performance, multithread, parallel
| d********w
发帖数: 363 | 9 快了
【在 l*****a 的大作中提到】
: 最近这么多面试？
: 快毕业了？
:
: )
: response
|
|
