DEPLOYDatabase

Chip comparison

AWS Trainium (Trn1) vs Google TPU v6 (Trillium)

Side-by-side specs

Straight from each vendor's datasheet. Winner in bold: higher throughput, memory, or transistor count; lower power draw or price. Sparse (2:4) throughput roughly doubles the FP8/INT8 numbers shown.

FieldAWS Trainium (Trn1)Google TPU v6 (Trillium)
Process nodeTSMC N7
TDP500 W200 W
Memory32 GB HBM2e32 GB HBM3
Memory bandwidth820 GB/s1,600 GB/s
BF16 (dense)210 TFLOPS918 TFLOPS
FP8 (dense)1,836 TFLOPS
INT8 (dense)1,836 TOPS
Form factorTrn1 EC2 instanceOAM (Trillium pod, up to 256 chips)
Announced2020-12-012024-05-14
Released2022-10-012024-12-01

Robots running AWS Trainium (Trn1)

No robots publicly running it yet.

Robots running Google TPU v6 (Trillium)

No robots publicly running it yet.

Data centers with AWS Trainium (Trn1)

No data centers publicly running it yet.

Data centers with Google TPU v6 (Trillium)

Common questions

AWS Trainium (Trn1) vs Google TPU v6 (Trillium): which is faster for training?

Google TPU v6 (Trillium) has 8.74x the dense FP16/BF16 throughput of the other (AWS Trainium (Trn1): 210 TFLOPS; Google TPU v6 (Trillium): 1,836 TFLOPS). Real-world training speed also depends on memory bandwidth, interconnect, and software stack: see the spec diff for those inputs.

AWS Trainium (Trn1) vs Google TPU v6 (Trillium): which has more memory?

Google TPU v6 (Trillium) carries more HBM (AWS Trainium (Trn1): 32 GB HBM2e; Google TPU v6 (Trillium): 32 GB HBM3). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

AWS Trainium (Trn1) vs Google TPU v6 (Trillium): which is more power-efficient?

Dense FP16/BF16 TFLOPS per watt: AWS Trainium (Trn1) 0.42 (210 ÷ 500 W); Google TPU v6 (Trillium) 9.18 (1,836 ÷ 200 W). Google TPU v6 (Trillium) wins at the die level; system-level efficiency also depends on cooling and interconnect.

AWS Trainium (Trn1) vs Google TPU v6 (Trillium): which is more widely used?

AWS Trainium (Trn1): 0 robots and 0 data centers publicly running it. Google TPU v6 (Trillium): 0 robots and 1 data center. Empty here means no operator has said publicly, not zero adoption in the market.

See also: every chip comparison · AWS Trainium (Trn1) full page · Google TPU v6 (Trillium) full page.