trn1.2xlarge

Analysis

The trn1.2xlarge is a machine learning asic instances instance in the trn1 series with 8 vCPUs and 32 GiB of memory on Intel Xeon Scalable (Icelake) (up to Up to 3.5 GHz), which makes it balanced at 4 GiB of memory per vCPU.

It carries 1 GPU with 32 GB of GPU memory, so most of the price is the accelerator rather than the CPUs. It has local instance storage (1 x 475 GB NVMe SSD), which is fast but lost when the instance stops. Network: 12500 Megabit.

Its price is almost flat across the 3 regions that offer it ($1.34 to $1.34/hour), so latency and data residency should decide where it runs. It is not offered in Canada Central, São Paulo, Ireland, London, Frankfurt, Mumbai, Singapore, Sydney and Tokyo, so plan around that if you need a presence there.

Per vCPU it is one of the dearer machine learning asic instances types: 77% of the 13 current types in the family cost less than its $0.1680 per vCPU-hour. The premium mostly pays for the accelerators. On demand it is $1.34/hour in US East (N. Virginia), or $981 for a 730-hour month.

Looking beyond its own line, 5 of 7 current types with 8 vCPUs and similar memory are cheaper. The cheapest of those is the g6f.2xlarge on AMD EPYC 7R13 at $0.4750/hour.