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 A10	NVIDIA B200	Cerebras WSE-3	Graphcore IPU-M2000
Architecture	Ampere	Blackwell	Wafer Scale Engine	Colossus Mk2
Form Factor	PCIe	SXM	—	—
VRAM	24 GB	192 GB	44 GB	3.6 GB
Memory Bandwidth	600 GB/s	8,000 GB/s	—	—
TFLOPs (FP32)	31.2	90	—	—
TFLOPs (FP16)	125	2,250	—	500
TFLOPs	125	2,250	—	—
TFLOPs (FP8)	—	4,500	—	—
TDP	150 W	1000 W	23000 W	180 W
Launch Date	Apr 2021	Mar 2024	Mar 2024	Jul 2020

Efficiency Metrics

Metric	A10	B200	WSE-3	IPU-M2000
TFLOPs per Watt (FP32-eq)	0.42	1.13	—	—
Memory Bandwidth per GB	25.0 GB/s	41.7 GB/s	—	—

Performance Equivalence

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

To match 1x NVIDIA A10

NVIDIA B200

Compute (FP32-eq)

0.06x

B200 is 18.00x faster

FP32 Compute

0.35x

B200 is 2.88x faster

VRAM

0.13x

B200 has 8.00x more

Memory Bandwidth

0.07x

B200 has 13.33x more

Cerebras WSE-3

VRAM

0.55x

WSE-3 has 1.83x more

Graphcore IPU-M2000

VRAM

6.67x

Need 6.67x IPU-M2000

To match 1x NVIDIA B200

NVIDIA A10

Compute (FP32-eq)

18.00x

Need 18.00x A10

FP32 Compute

2.88x

Need 2.88x A10

VRAM

8.00x

Need 8.00x A10

Memory Bandwidth

13.33x

Need 13.33x A10

Cerebras WSE-3

VRAM

4.36x

Need 4.36x WSE-3

Graphcore IPU-M2000

VRAM

53.33x

Need 53.33x IPU-M2000

To match 1x Cerebras WSE-3

NVIDIA A10

VRAM

1.83x

Need 1.83x A10

NVIDIA B200

VRAM

0.23x

B200 has 4.36x more

Graphcore IPU-M2000

VRAM

12.22x

Need 12.22x IPU-M2000

To match 1x Graphcore IPU-M2000

NVIDIA A10

VRAM

0.15x

A10 has 6.67x more

NVIDIA B200

VRAM

0.02x

B200 has 53.33x more

Cerebras WSE-3

VRAM

0.08x

WSE-3 has 12.22x more

Pricing

Price Type	A10	B200	WSE-3	IPU-M2000
CAPEX (Street Price)	—	$70,000	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	—	$62	—	—