DEPLOYDatabase

Chip comparison

AWS Trainium (Trn1) vs Huawei Ascend 910C

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)Huawei Ascend 910C
Process nodeTSMC N7SMIC N+2 (~7nm)
TDP500 W500 W
Memory32 GB HBM2e128 GB HBM2e
Memory bandwidth820 GB/s3,200 GB/s
FP16 (dense)—800 TFLOPS
BF16 (dense)210 TFLOPS800 TFLOPS
INT8 (dense)—1,600 TOPS
Form factorTrn1 EC2 instanceOAM
Announced2020-12-012024-09-19
Released2022-10-012025-01-01

Robots running AWS Trainium (Trn1)

No robots publicly running it yet.

Robots running Huawei Ascend 910C

No robots publicly running it yet.

Data centers with AWS Trainium (Trn1)

No data centers publicly running it yet.

Data centers with Huawei Ascend 910C

No data centers publicly running it yet.

Common questions

AWS Trainium (Trn1) vs Huawei Ascend 910C: which is faster for training?

Huawei Ascend 910C has 3.81x the dense FP16/BF16 throughput of the other (AWS Trainium (Trn1): 210 TFLOPS; Huawei Ascend 910C: 800 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 Huawei Ascend 910C: which has more memory?

Huawei Ascend 910C carries more HBM (AWS Trainium (Trn1): 32 GB HBM2e; Huawei Ascend 910C: 128 GB HBM2e). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

AWS Trainium (Trn1) vs Huawei Ascend 910C: which is more power-efficient?

Dense FP16/BF16 TFLOPS per watt: AWS Trainium (Trn1) 0.42 (210 ÷ 500 W); Huawei Ascend 910C 1.60 (800 ÷ 500 W). Huawei Ascend 910C wins at the die level; system-level efficiency also depends on cooling and interconnect.

See also: every chip comparison · AWS Trainium (Trn1) full page · Huawei Ascend 910C full page.