Compare XPUs

Select up to 5 XPUs to compare side-by-side

Select XPUs to Compare

Clear all (4)Jump to results

Filter by Vendor

Showing 72 XPUs • 4 selected

Alibaba

Hanguang 800

AMD

MI100

23.1 TFLOPs

AMD

MI210

181 TFLOPs

AMD

MI250X

383 TFLOPs

AMD

MI300A

980.6 TFLOPs

AMD

MI300X

1,307 TFLOPs

AMD

MI325X

1,400 TFLOPs

AMD

MI350X

2,100 TFLOPs

AMD

Radeon PRO W7900

122 TFLOPs

AMD

Radeon RX 7900 XT

104 TFLOPs

AMD

Radeon RX 7900 XTX

122 TFLOPs

AWS

Inferentia2

190 TFLOPs

AWS

Trainium

190 TFLOPs

AWS

Trainium2

680 TFLOPs

Baidu

Kunlun II

Biren Technology

BR100

Cambricon

MLU370

256 TFLOPs

Cerebras

WSE-3

Enflame Technology

CloudBlazer T20

Etched

Sohu

10,000 TFLOPs

FuriosaAI

Warboy

Google

TPU v4

275 TFLOPs

Google

TPU v5e

197 TFLOPs

Google

TPU v5p

459 TFLOPs

Google

TPU v6e (Trillium)

918 TFLOPs

Graphcore

Bow IPU

Graphcore

IPU-M2000

Groq

LPU Inference Engine

Huawei

Ascend 910B

Iluvatar CoreX

BI-V150

300 TFLOPs

Intel

Data Center GPU Max 1100

177 TFLOPs

Intel

Data Center GPU Max 1550

419 TFLOPs

Intel Habana

Gaudi 2

432 TFLOPs

Intel Habana

Gaudi 3

1,835 TFLOPs

Multi-Metric Comparison

Relative performance across 5 key metrics (normalized to 100 = best in comparison)

Compute Performance (BF16)

Memory Capacity

Power Consumption

Power Efficiency

Specifications

Specification	NVIDIA A100 SXM	AWS Trainium2	NVIDIA GeForce RTX 4060 Ti	Etched Sohu
Architecture	Ampere	Trainium Gen2	Ada Lovelace	Transformer ASIC
Form Factor	SXM	—	PCIe	PCIe
VRAM	80 GB	—	8 GB	144 GB
Memory Bandwidth	2,039 GB/s	—	288 GB/s	4,800 GB/s
TFLOPs (FP32)	19.5	—	22.1	10,000
TFLOPs (FP16)	312	—	44.2	—
TFLOPs	312	680	44.2	10,000
TFLOPs (FP8)	—	—	—	—
TDP	400 W	300 W	160 W	700 W
Launch Date	May 2020	Nov 2023	May 2023	Jun 2024

Efficiency Metrics

Metric	A100 SXM	Trainium2	GeForce RTX 4060 Ti	Sohu
TFLOPs per Watt (FP32-eq)	0.39	1.13	0.14	14.29
Memory Bandwidth per GB	25.5 GB/s	—	36.0 GB/s	33.3 GB/s

Performance Equivalence

How many units of each GPU are needed to match the performance of the others?

To match 1x NVIDIA A100 SXM

AWS Trainium2

Compute (FP32-eq)

0.46x

Trainium2 is 2.18x faster

NVIDIA GeForce RTX 4060 Ti

Compute (FP32-eq)

7.06x

Need 7.06x GeForce RTX 4060 Ti

FP32 Compute

0.88x

GeForce RTX 4060 Ti is 1.13x faster

VRAM

10.00x

Need 10.00x GeForce RTX 4060 Ti

Memory Bandwidth

7.08x

Need 7.08x GeForce RTX 4060 Ti

Etched Sohu

Compute (FP32-eq)

0.02x

Sohu is 64.10x faster

FP32 Compute

0.00x

Sohu is 512.82x faster

VRAM

0.56x

Sohu has 1.80x more

Memory Bandwidth

0.42x

Sohu has 2.35x more

To match 1x AWS Trainium2

NVIDIA A100 SXM

Compute (FP32-eq)

2.18x

Need 2.18x A100 SXM

NVIDIA GeForce RTX 4060 Ti

Compute (FP32-eq)

15.38x

Need 15.38x GeForce RTX 4060 Ti

Etched Sohu

Compute (FP32-eq)

0.03x

Sohu is 29.41x faster

To match 1x NVIDIA GeForce RTX 4060 Ti

NVIDIA A100 SXM

Compute (FP32-eq)

0.14x

A100 SXM is 7.06x faster

FP32 Compute

1.13x

Need 1.13x A100 SXM

VRAM

0.10x

A100 SXM has 10.00x more

Memory Bandwidth

0.14x

A100 SXM has 7.08x more

AWS Trainium2

Compute (FP32-eq)

0.07x

Trainium2 is 15.38x faster

Etched Sohu

Compute (FP32-eq)

0.00x

Sohu is 452.49x faster

FP32 Compute

0.00x

Sohu is 452.49x faster

VRAM

0.06x

Sohu has 18.00x more

Memory Bandwidth

0.06x

Sohu has 16.67x more

To match 1x Etched Sohu

NVIDIA A100 SXM

Compute (FP32-eq)

64.10x

Need 64.10x A100 SXM

FP32 Compute

512.82x

Need 512.82x A100 SXM

VRAM

1.80x

Need 1.80x A100 SXM

Memory Bandwidth

2.35x

Need 2.35x A100 SXM

AWS Trainium2

Compute (FP32-eq)

29.41x

Need 29.41x Trainium2

NVIDIA GeForce RTX 4060 Ti

Compute (FP32-eq)

452.49x

Need 452.49x GeForce RTX 4060 Ti

FP32 Compute

452.49x

Need 452.49x GeForce RTX 4060 Ti

VRAM

18.00x

Need 18.00x GeForce RTX 4060 Ti

Memory Bandwidth

16.67x

Need 16.67x GeForce RTX 4060 Ti

Pricing

Price Type	A100 SXM	Trainium2	GeForce RTX 4060 Ti	Sohu
CAPEX (Street Price)	$15,000	—	—	—
OPEX (per hour)	$4.05/hr	$1.34/hr	—	—
Price per TFLOPs (FP32-eq)	$96	—	—	—