AWS Amazon EC2

hpc8a.96xlarge

Part of the HPC8A High-Performance Computing family — Optimized for tightly-coupled, large-scale HPC workloads.

Name:hpcHigh Performance Computing88th GenerationaAMD EPYC96xlarge96 Extra Large

On-demand, Linux

$7.9200/ hr

~$5,781.60/mo · lowest in US East (Ohio), us-east-2

Savings Plan, Linux

$4.89993/ hr

From, with a 3-year Compute Savings Plan, No Upfront: −38% vs on-demand rate (us-east-2)

Per unit

$0.0413/ vCPU-hr

$0.0103/GB-hr

Compute

192vCPUs

192 cores • x86_64

Memory

768GB

4.0 GB / vCPU

Network

300 Gbps

Baseline: 300 Gbps

Storage & EBS

260.625MB/s

Max: 11,000 IOPS

On Google Cloud

The nearest Google Cloud machines to hpc8a.96xlarge by vCPU and memory, at On-demand Linux rates in Ohio (us-east-2 ↔ us-east5). hpc8a.96xlarge isn't sold in us-east-1, so this compares the paired region where its rate is lowest.

Machine	Size	On-demand Linux	vs hpc8a.96xlarge
hpc8a.96xlargeThis page · AWS, us-east-2	192 vCPU · 768 GiB	$7.9200/hr	—
n2d-standard-224224 vCPU and 896 GiB for 192 vCPU and 768 GiB, the nearest size up; AMD on both; 192 physical cores vs 112	224 vCPU · 896 GiB	$9.463104/hr	19% more
c4d-highmem-192192 vCPU and 1512 GiB for 192 vCPU and 768 GiB, the nearest size up; AMD on both; 192 physical cores vs 96	192 vCPU · 1,512 GiB	$11.954135/hr	51% more
z4d-highmem-192-standardlssd192 vCPU and 1512 GiB for 192 vCPU and 768 GiB, the nearest size up; AMD on both; 192 physical cores vs 96; local disk on z4d-highmem-192-standardlssd only	192 vCPU · 1,512 GiB	$13.084024/hr	65% more

Matched on vCPU and memory, then CPU vendor, then rate; a larger size is offered only when nothing is the same size. On-demand Linux only: Spot and commitments work differently on the two clouds, so they aren't compared here.

Technical Specifications & Limits

Hardware specs, container thresholds, and system limits

Compute & Processor

ProcessorAMD EPYC 9R45 Processor

Architecturex86_64 (x86_64)

vCPUs / Threads192 vCPUs

Physical Cores192

Clock Speed4.5 GHz

Intel AVX / AVX2No

Intel Turbo BoostNo

Bare MetalNo (Virtualized)

Performance & Benchmarks

CoreMark ScoreNot benchmarked

CoreMark / vCPU—

FFmpeg Transcoding—

CoreMark per Dollar—

Cost per vCPU-hr$0.0413

Cost per GB-hr$0.0103

Containers & Kubernetes

EKS Max Pods100 pods

Max ENIs4

IPv4/v6 per ENI50

Total Addressable IPs200 total IPs

CNI Prefix DelegationSupported (High Density)

Networking & Bandwidth

Network Performance300 Gbps

Baseline Bandwidth300 Gbps

Burst / Peak Bandwidth300 Gbps

Enhanced NetworkingYes

EFA SupportNo

IPv6 SupportNative Dual-Stack Supported

Storage & EBS Optimization

Storage ConfigurationEBS only

Local NVMe DiskNone (EBS only)

EBS Baseline Throughput10.875 MB/s

EBS Max Throughput260.625 MB/s

EBS Baseline IOPS500

EBS Max / Burst IOPS11,000

Dedicated EBS Throughput87 Mbps

Platform & Architecture Details

Hypervisor / SystemAWS Nitro System (PCIe Offload)

Nitro EnclavesNot Supported

Normalization Factor384

Generation StatusCurrent Generation (Recommended)

Rates by region

Live on-demand rates across all global AWS cloud regions

Lowest Price Region

$7.9200/ hr

US East (Ohio) (us-east-2)

Highest Price Region

$10.2299/ hr

Asia Pacific (Tokyo) (ap-northeast-1)

Global Fleet Average

$9.2163/ hr

Across all active regions

Regional Coverage

5regions

Available worldwide

Linux — Available in all 5 regions
Region	On-Demand Price	% vs On-Demand Rate
ap-southeast-1 Asia Pacific (Singapore)	$9.9000
ap-northeast-1 Asia Pacific (Tokyo) Highest	$10.2299
us-gov-west-1 AWS GovCloud (US-West)	$9.5341
eu-north-1 EU (Stockholm)	$8.4977
us-east-2 US East (Ohio) Lowest	$7.9200

Windows — Available in all 5 regions
Region	On-Demand Price	% vs On-Demand Rate
ap-southeast-1 Asia Pacific (Singapore)	$18.7320
ap-northeast-1 Asia Pacific (Tokyo) Highest	$19.0619
us-gov-west-1 AWS GovCloud (US-West)	$18.3661
eu-north-1 EU (Stockholm)	$17.3297
us-east-2 US East (Ohio) Lowest	$16.7520

Savings Plans: what to know

A Savings Plan commits you to a dollar amount per hour of spend for one or three years, not to an instance: usage up to that amount is billed at the plan's rates and the rest at on-demand, and unused commitment is still billed.
A Compute Savings Plan follows your usage across instance families, sizes, regions and operating systems; an EC2 Instance Savings Plan is locked to one family in one region in exchange for a lower rate (what that flexibility costs).
In US East (Ohio) (us-east-2), the 3-year Compute Savings Plan rate (No Upfront) for hpc8a.96xlarge on Linux is 62% of the on-demand rate, so the instance has to run at least 62% of the hours for the commitment to cost less than on-demand.
On Windows a plan discounts only the compute part of the rate: the Windows license part costs the same as on demand, so the percentage off is smaller than on Linux.
Rates are AWS's effective hourly rate for each payment option; All Upfront pays the whole term's commitment at purchase, which is why its hourly rate is lower; Partial Upfront (at least half up front) isn't shown on this page. The Savings Plans calculator has all three and sizes a commitment for this instance or a mix of them. How AWS applies a plan each hour: how Compute Savings Plans decide what they cover. Sizing one for elastic workloads: managing Compute Savings Plans.