DEPLOYDatabase

Chip comparison

AWS Trainium (Trn1) vs Groq LPU (Language Processing Unit)

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)Groq LPU (Language Processing Unit)
Process nodeTSMC N7GlobalFoundries 14nm
TDP500 W375 W
Memory32 GB HBM2e0.23 GB SRAM (230 MB on-die)
Memory bandwidth820 GB/s80,000 GB/s
FP16 (dense)188 TFLOPS
BF16 (dense)210 TFLOPS
INT8 (dense)750 TOPS
Form factorTrn1 EC2 instancePCIe (GroqCard) / rack-scale GroqRack
Announced2020-12-012020-01-01
Released2022-10-012021-06-01

Robots running AWS Trainium (Trn1)

No robots publicly running it yet.

Robots running Groq LPU (Language Processing Unit)

No robots publicly running it yet.

Data centers with AWS Trainium (Trn1)

No data centers publicly running it yet.

Data centers with Groq LPU (Language Processing Unit)

No data centers publicly running it yet.

Common questions

AWS Trainium (Trn1) vs Groq LPU (Language Processing Unit): which is faster for training?

AWS Trainium (Trn1) has 1.12x the dense FP16/BF16 throughput of the other (AWS Trainium (Trn1): 210 TFLOPS; Groq LPU (Language Processing Unit): 188 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 Groq LPU (Language Processing Unit): which has more memory?

AWS Trainium (Trn1) carries more HBM (AWS Trainium (Trn1): 32 GB HBM2e; Groq LPU (Language Processing Unit): 0.23 GB SRAM (230 MB on-die)). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

AWS Trainium (Trn1) vs Groq LPU (Language Processing Unit): which is more power-efficient?

Dense FP16/BF16 TFLOPS per watt: AWS Trainium (Trn1) 0.42 (210 ÷ 500 W); Groq LPU (Language Processing Unit) 0.50 (188 ÷ 375 W). Groq LPU (Language Processing Unit) wins at the die level; system-level efficiency also depends on cooling and interconnect.

See also: every chip comparison · AWS Trainium (Trn1) full page · Groq LPU (Language Processing Unit) full page.