COMPSCI 92L Lecture Notes - Lecture 9: Graphics Processing Unit, General-Purpose Computing On Graphics Processing Units, Software As A Service

21 views1 pages
11 May 2016
of 1
Spring 2016
Programming in Scratch/Cloud Computing, Autonomous Vehicles
Chips have multi-core interfaces, where slow chips put together make one chip that is
Can layer chips both vertically and horizontally, meaning there’s “multi-core”
Ex. a 2014 MacBook pro uses an 2.6 GHz Intel Core i5
The number of things connected to the Internet is doubling every 2 years
Moore’s Law - speed, power of computer chips will double every two years because the
number of transistors in a dense integrated circuit doubles every two years and the size
of hardware will half every 15 years
Cache - list of entries accumulated over search history, ex. Clearing the cache on your
computer removes all addresses previously listed
Concurrency - do many things at the same time
Pipeline computing: people pass a task around
Concurrency: found in computer chips where actions happen all at the same time
Graphics processing unit (GPU) - concentrated amount of chips used to decipher
graphics in sophisticated video games
Ex. CPU - has multiple cores (ex. 2^3 = 8), GPU - thousands of cores (ex. 2^10)
GP/GPU: general purpose graphics processing unit, not only for video games
Costs $700-1000
CPU found in phone and laptop
Cloud computing is a model for enabling convenient, on-demand network access to a
shared pool of configurable computing resources (ex. Networks, servers, storage,
applications, services) that can be rapidly provisioned and released with minimal
management effort or service provider interaction
Ex. Microsoft’s Azure, Amazon web services’ EC2, Google Cloud platform
SaaS - software as a service, consumer facing
PasS - platform as a service developer facing
IaaS - infrastructure as a service, computing power
find more resources at
find more resources at