AWS Associate Architect Certification Lecture 9— EC2 part 6 — Accelerated computing instances

In this lecture , we are discussing about GPU enabled accelerated computing instances.

Features :

Accelerated computing instances family uses accelerated hardwares or co- processors to perform some functions such as floating point, data pattern matching , graphics processing more efficiently

Uses :
USED FOR live streaming, 3D gaming, AI , machine learning, deep learning.

There are 3 series in this category :

P — series
F — series
G — series

Lets explore each one by one :

F — series
Only one option in this series : F1 instances

Features :

customizable hardware acceleration with field programmable gate arrays (FPGA) for signal processing.

Uses :

1. used in image processing
2. used for live streaming.
3. Each FPGA contains 2.5 million logic elements and 6800 DSP engines
   designed to accelerate computationally high intensive algorithms like data flow or hugh parallel operations like facebook live, online gaming
4. Gnomic research, AI and machine learning, deep learning, video processing

Configurations :

VCPU — 8 to 64
FPGA — 1 to 8
RAM — 122 to 976 GB
storage — NVM SSD

— — — — — — — — — — — — — — — — — — — — — — — — — — —

P — series

features :
1. NVidia tesla t60 graphics card is present
2. high bandwidth,
3. deep learning, flow dynamics

Different types in this category :

P2 instances :
Configurations :

vCPU 4 to 64
GPU 1 to 16
RAM 61 to 732
GPU ram 12 to 192
network bandwidth 25 gbps

P3 instances :

Configurations :
vcpu 8 to 96
GPU 1 to 8
RAM 61 to 768
no GPU RAM defined
network bandwidth 10 GBPS
P3 is latest than P2

Use cases — DL, ML, AI, scesmic learning, molecular modelling,

Imp note. : API used in p3 — cuda 9 AND OPEN CL , IN p2 — CUDA and OPEN CL -1.2v

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

G — series
Types in this category : G2 and G3 instances

Uses :

1. for 3D applications like 3d gaming
2. graphic intensive
3. 3d visualization, facebook live

Configurations :

1. NVidia m60 processors
2. VCPU — 4 to 64
3. GPU 1 to 4
4. RAM 30 to 488
5. GpU memory — 8 to 32 gb
6. network 25gbps — max network badwidth provided by aws

Usage —