Optimize Kubernetes clusters with Pod Efficiency Metrics ->

trn1.32xlarge

The trn1.32xlarge instance is in the machine learning asic instances family with 128 vCPUs, 512.0 GiB of memory and 8 Gibps of bandwidth starting at $21.5 per hour.

paid

Pricing

On Demand

Spot

1 Yr Reserved

3 Yr Reserved

$5983.56 per month (-62%) with Autopilot

Learn more

dns

Family Sizes

Size	vCPUs	Memory (GiB)
trn1.2xlarge	8	32
trn1.32xlarge	128	512

Compare trn1.32xlarge to other Instances

dns

Instance Variants

Variant
trn1
trn1n

info

Instance Details

Compute	Value
vCPUs	128
Memory (GiB)	512.0
Memory per vCPU (GiB)	4.0
Physical Processor	Intel Xeon Scalable (Icelake)
Clock Speed (GHz)	3.5
CPU Architecture	x86_64
GPU	16
GPU Architecture	AWS Inferentia
Video Memory (GiB)	512
GPU Compute Capability (?)	0
FPGA	0

Networking	Value
Network Performance (Gibps)	8
Enhanced Networking	True
IPV6	True
Placement Group (?)	True

Storage	Value
EBS Optimized	True
Max Bandwidth (Mbps) on (EBS)	80000
Max Throughput (MB/s) on EBS	10000.0
Max I/O Operations/second (IOPS)	260000
Baseline Bandwidth (Mbps) on (EBS)	80000
Baseline Throughput (MB/s) on EBS	10000.0
Baseline I/O Operations/second (IOPS)	260000
Devices	4
Swap Partition	False
NVME Drive	True
Disk Space (GiB)	1900
SSD	True
Initialize Storage	False

Amazon	Value
Generation	current
Instance Type	trn1.32xlarge
Family	Machine Learning ASIC Instances
Name	TRN1 32xlarge
Elastic Map Reduce (EMR)	False