Skip to content

Latest commit

 

History

History
executable file
·
157 lines (95 loc) · 5.22 KB

day57.md

File metadata and controls

executable file
·
157 lines (95 loc) · 5.22 KB

On the fifty seventh day, I learned the following things about Cloud Computing.

5. Accelerated Computing Instances

  • Accelerated computing instance families use hardware accelerators or co-processors to perform some functions such as floating point number calculation, graphics processing or data pattern matching more efficiently than it is possible in software running on CPUs.

  • This instance is used in AI, ML, and DL. It is also used for live streaming that requires fast processing.

  • There are three types of series in it.

    • F series
      • F1
    • P series
      • P2
      • P3
    • G series
      • G2
      • G3

F series:

  • F1 instance offers customizable hardware acceleration with Field Programmable Gate Arrays(FPGA). FPGA fastly processes an image and make changes in it.

  • Each FPGA contains 2.5 million logic elements(logic gates like AND, NOT gates, etc) and 6800 Digital signal processing(DSP) engines.

  • Designed to accelerate computationally intensive algorithms such as data flow or highly parallel operations. Games usually requires parallel processing like button is pressed and action is done live.

  • F1 provides local NVM SSD storage.

    • vCPU -> 8 to 64
    • FPGA -> 1 to 8
    • RAM -> 122 to 976 GB
    • Instance storage -> NVMe SSD

  • It is used in Genomics research, financial analytics, real time video recording, and big data search.

P series:

  • It uses NVIDIA Tesla GPUs

  • It provides high bandwidth networking.

  • Upto 32GB of memory per GPU which makes them ideal for deep learning and computational fluid dynamics.

  • P2 instance

    • vCPU -> 4 to 64
    • GPU -> 1 to 16
    • RAM -> 61 to 732 GB
    • GPU RAM -> 12 to 192 GB
    • Network Bandwidth -> 25 Gbps

  • P3 instance

    • vCPU -> 8 to 96
    • GPU -> 1 to 8
    • RAM -> 61 to 768 GB
    • Storage -> SSD and EBS

  • It is used in machine learning, databases, seismic analysis(earthquake study), genomics, molecular modeling(chemistry related), AI, Deep learning.

  • Note: P3 support CUDA 9 and OpenCL API. P2 supports CUDA 8 and OpenCL 1.2

G series:

  • It is optimized for graphics intensive applications.

  • It is well suited for apps like 3D visualization applications.

  • G3 instances use NVIDIA Tesla M60 GPU and provide a cost effective high performance platform for graphic applications.

    • vCPU -> 6 to 64
    • GPU -> 1 to 4
    • RAM -> 30.5 to 488 GB
    • GPU Memory -> 8 to 32 GB
    • Network Performance -> 25 Gbps

  • It is used in video creation services, 3D visualization, streaming graphics-intensive application.

6. High Memory Instances

  • High memory instances are purpose built to run large-in-memory databases, including production development of SAP HANA in the cloud.

  • It has only one series that is U-series. U-series has 3 subtypes that U6, U9 and U12.

Note

  • High memory instances are bare metal instances and do not run on a hypervisor.

  • The dedicated host will be provided for running only your server in it.

  • The dedicated host is available in purchasing category for 3 years term.

  • OS is directly installed on hardware.

Features

  • Latest generation intel xeon pentium 8176M processor.

  • 6, 9, 12 TB of instance memory, the largest of any EC2 instance.

  • Powered by the AWS nitro system, a combination of dedicated hardware and lightweight hypervisor.

  • Bare metal performance with direct access to host hardware.

  • EBS is optimized by default at no additional cost.

  • Network performance is 25 Gbps.

  • Dedicated EBS bandwidth is 14 Gbps. It will read and write the data from storage fastly.

  • Each instance offer 448 logical processor.

7. Previous Version Instances

  • These are the previous instances that you can still purchase and use them.

  • If you bought any one of them previously as a dedicated instance but now they're added to the previous version, the price of these instances won't change. They will be the same

    • T1
    • M1
    • C1
    • CC2
    • M2
    • CR1
    • CG1
    • i2
    • HS1
    • M3
    • C3
    • R3

Important Questions

  • Q. When the EC2 bill starts and stops or by which mean we have to pay?

    A. EC2 instances will start counting the bill if it is booted and it will shut down when you terminate.

  • Q. If you stop or shutdown the server or instance but didn't terminate, will I have to pay the whole bill?

    A. No, but you just have to pay the storage fee of the instance, not the processing fee because the server is not running.

  • Q. Is the pay per second option available for windows?

    A. No, it is currently available for linux servers/instances only. For windows, it is pay per hour.

  • Q. Will the bill be provided based on seconds usage or minutes or hourly usage?

    A. The bill will be provided to you every month and they will be based on the hourly usage. The minutes and seconds will be calculated in those hours.

  • Q. Does the billing contains the taxes?

    A. No, the taxes are excluded from the bills. Only the instance payment is fixed. The taxes depends from country to country.

Explaining it in a video

Here you can get an explanation in a video. 57/60 Day of DevOps Challenge