p3.16xlarge
⚡ 8x GPU • 128 GiBPart of the P3 Accelerated Computing family — Hardware accelerators (GPUs, FPGAs) for specialized compute.
Name:
pGPU Deep Learning & HPCp = GPU Deep Learning & HPCHigh-end NVIDIA GPUs (Blackwell, Hopper, Ampere, Volta) built for frontier AI/LLM training, HPC simulations, and generative AI.33rd Generation3 = 3rd GenerationAWS 3rd generation hardware platform with updated CPU/system architecture.16xlarge16 Extra Large16xlarge = 16 Extra Large64 vCPUs • 488 GB RAM • EBS only • 25 Gbps Network • 8x NVIDIA Tesla V100 (128.0 GB VRAM)Price (Linux)
$24.4800/ hr
~$17,870.40/mo
Compute
64vCPUs
64 threads • x86_64
Memory
488GB
7.6 GB / vCPU
Value
$0.3825/ vCPU-hr
$0.0502/GB-hr
Network
25 Gbps
Standard
Storage & EBS
EBSOnly
EBS only
Technical Specifications & Limits
Hardware specs, container thresholds, and system limitsCompute & Processor
ProcessorIntel Xeon E5-2686 v4 (Broadwell)
Architecturex86_64 (x86_64)
vCPUs / Threads64 vCPUs
Physical Cores—
Clock Speed2.3 GHz
Intel AVX / AVX2AVX & AVX2 Supported
Intel Turbo BoostYes
Bare MetalNo (Virtualized)
Performance & Benchmarks
CoreMark ScoreNot benchmarked
CoreMark / vCPU—
FFmpeg Transcoding—
CoreMark per Dollar—
Cost per vCPU-hr$0.3825
Cost per GB-hr$0.0502
Containers & Kubernetes
EKS Max Pods234 pods
Max ECS Tasks7 tasks
Max ENIs—
IPv4/v6 per ENI—
Total Addressable IPs—
CNI Prefix DelegationStandard CNI
Networking & Bandwidth
Network Performance25 Gbps
Baseline Bandwidth—
Burst / Peak Bandwidth—
Enhanced NetworkingYes
EFA SupportNo
IPv6 SupportNative Dual-Stack Supported
Storage & EBS Optimization
Storage ConfigurationEBS only
Local NVMe DiskNone (EBS only)
EBS Baseline Throughput—
EBS Max Throughput—
EBS Baseline IOPS—
EBS Max / Burst IOPS—
Dedicated EBS Throughput14000 Mbps
GPU & Hardware Acceleration
GPU ModelNVIDIA Tesla V100
Accelerator Count8x GPU(s)
Total GPU Memory128.0 GB
Memory ArchitectureGDDR / HBM
CUDA Compute Capabilityv7.0
Acceleration RuntimesCUDA, TensorRT, cuDNN, PyTorch, PyJAX
Platform & Architecture Details
Hypervisor / SystemXen Hypervisor
Nitro EnclavesNot Supported
Normalization Factor128
Generation StatusCurrent Generation (Recommended)
Date Introduced2017-10-25
Regional Pricing Matrix
Live on-demand rates across all global AWS cloud regionsLowest Price Region
$24.4800/ hr
US East (N. Virginia) (us-east-1)Highest Price Region
$33.8720/ hr
Asia Pacific (Seoul) (ap-northeast-2)Global Fleet Average
$29.2207/ hr
Across all active regionsRegional Coverage
12regions
Available worldwide| Region | On-Demand Price |
|---|---|
ap-northeast-2 HighestAsia Pacific (Seoul) | $33.8720 |
ap-southeast-1 HighestAsia Pacific (Singapore) | $33.8720 |
ap-southeast-2 HighestAsia Pacific (Sydney) | $33.8720 |
ap-northeast-1 Asia Pacific (Tokyo) | $33.5520 |
us-gov-west-1 AWS GovCloud (US-West) | $29.3760 |
ca-central-1 Canada (Central) | $26.9280 |
eu-central-1 EU (Frankfurt) | $30.5840 |
eu-west-1 EU (Ireland) | $26.4400 |
eu-west-2 EU (London) | $28.7120 |
us-east-1 LowestUS East (N. Virginia) | $24.4800 |
us-east-2 LowestUS East (Ohio) | $24.4800 |
us-west-2 LowestUS West (Oregon) | $24.4800 |
| Region | On-Demand Price |
|---|---|
ap-northeast-2 HighestAsia Pacific (Seoul) | $36.8160 |
ap-southeast-1 HighestAsia Pacific (Singapore) | $36.8160 |
ap-southeast-2 HighestAsia Pacific (Sydney) | $36.8160 |
ap-northeast-1 Asia Pacific (Tokyo) | $36.4960 |
us-gov-west-1 AWS GovCloud (US-West) | $32.3200 |
ca-central-1 Canada (Central) | $29.8720 |
eu-central-1 EU (Frankfurt) | $33.5280 |
eu-west-1 EU (Ireland) | $29.3840 |
eu-west-2 EU (London) | $31.6560 |
us-east-1 LowestUS East (N. Virginia) | $27.4240 |
us-east-2 LowestUS East (Ohio) | $27.4240 |
us-west-2 LowestUS West (Oregon) | $27.4240 |
No regions matched your search filter.