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 A40	AMD MI325X	Cerebras WSE-3	AWS Inferentia2
Architecture	Ampere	CDNA 3.5	Wafer Scale Engine	Inferentia Gen2
Form Factor	PCIe	OAM	—	—
VRAM	48 GB	256 GB	44 GB	32 GB
Memory Bandwidth	696 GB/s	6,000 GB/s	—	—
TFLOPs (FP32)	37.4	175	—	—
TFLOPs (FP16)	150	1,400	—	—
TFLOPs	150	1,400	—	190
TFLOPs (FP8)	—	2,800	—	—
TDP	300 W	800 W	23000 W	150 W
Launch Date	Oct 2020	Oct 2024	Mar 2024	Nov 2022

Efficiency Metrics

Metric	A40	MI325X	WSE-3	Inferentia2
TFLOPs per Watt (FP32-eq)	0.25	0.88	—	0.63
Memory Bandwidth per GB	14.5 GB/s	23.4 GB/s	—	—

Performance Equivalence

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

To match 1x NVIDIA A40

AMD MI325X

Compute (FP32-eq)

0.11x

MI325X is 9.33x faster

FP32 Compute

0.21x

MI325X is 4.68x faster

VRAM

0.19x

MI325X has 5.33x more

Memory Bandwidth

0.12x

MI325X has 8.62x more

Cerebras WSE-3

VRAM

1.09x

Need 1.09x WSE-3

AWS Inferentia2

Compute (FP32-eq)

0.79x

Inferentia2 is 1.27x faster

VRAM

1.50x

Need 1.50x Inferentia2

To match 1x AMD MI325X

NVIDIA A40

Compute (FP32-eq)

9.33x

Need 9.33x A40

FP32 Compute

4.68x

Need 4.68x A40

VRAM

5.33x

Need 5.33x A40

Memory Bandwidth

8.62x

Need 8.62x A40

Cerebras WSE-3

VRAM

5.82x

Need 5.82x WSE-3

AWS Inferentia2

Compute (FP32-eq)

7.37x

Need 7.37x Inferentia2

VRAM

8.00x

Need 8.00x Inferentia2

To match 1x Cerebras WSE-3

NVIDIA A40

VRAM

0.92x

A40 has 1.09x more

AMD MI325X

VRAM

0.17x

MI325X has 5.82x more

AWS Inferentia2

VRAM

1.38x

Need 1.38x Inferentia2

To match 1x AWS Inferentia2

NVIDIA A40

Compute (FP32-eq)

1.27x

Need 1.27x A40

VRAM

0.67x

A40 has 1.50x more

AMD MI325X

Compute (FP32-eq)

0.14x

MI325X is 7.37x faster

VRAM

0.13x

MI325X has 8.00x more

Cerebras WSE-3

VRAM

0.73x

WSE-3 has 1.38x more

Pricing

Price Type	A40	MI325X	WSE-3	Inferentia2
CAPEX (Street Price)	—	$45,000	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	—	$64	—	—