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	AMD MI250X	Google TPU v5e	NVIDIA B200	AMD MI300X
Architecture	CDNA 2	TPU v5	Blackwell	CDNA 3
Form Factor	OAM	Mezzanine	SXM	OAM
VRAM	128 GB	—	192 GB	192 GB
Memory Bandwidth	3,277 GB/s	—	8,000 GB/s	5,300 GB/s
TFLOPs (FP32)	47.9	—	90	163
TFLOPs (FP16)	383	—	2,250	1,307
TFLOPs	383	197	2,250	1,307
TFLOPs (FP8)	—	—	4,500	2,614
TDP	560 W	250 W	1000 W	750 W
Launch Date	Nov 2021	Aug 2023	Mar 2024	Dec 2023

Efficiency Metrics

Metric	MI250X	TPU v5e	B200	MI300X
TFLOPs per Watt (FP32-eq)	0.34	0.39	1.13	0.87
Memory Bandwidth per GB	25.6 GB/s	—	41.7 GB/s	27.6 GB/s

Performance Equivalence

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

To match 1x AMD MI250X

Google TPU v5e

Compute (FP32-eq)

1.94x

Need 1.94x TPU v5e

NVIDIA B200

Compute (FP32-eq)

0.17x

B200 is 5.87x faster

FP32 Compute

0.53x

B200 is 1.88x faster

VRAM

0.67x

B200 has 1.50x more

Memory Bandwidth

0.41x

B200 has 2.44x more

AMD MI300X

Compute (FP32-eq)

0.29x

MI300X is 3.41x faster

FP32 Compute

0.29x

MI300X is 3.40x faster

VRAM

0.67x

MI300X has 1.50x more

Memory Bandwidth

0.62x

MI300X has 1.62x more

To match 1x Google TPU v5e

AMD MI250X

Compute (FP32-eq)

0.51x

MI250X is 1.94x faster

NVIDIA B200

Compute (FP32-eq)

0.09x

B200 is 11.42x faster

AMD MI300X

Compute (FP32-eq)

0.15x

MI300X is 6.63x faster

To match 1x NVIDIA B200

AMD MI250X

Compute (FP32-eq)

5.87x

Need 5.87x MI250X

FP32 Compute

1.88x

Need 1.88x MI250X

VRAM

1.50x

Need 1.50x MI250X

Memory Bandwidth

2.44x

Need 2.44x MI250X

Google TPU v5e

Compute (FP32-eq)

11.42x

Need 11.42x TPU v5e

AMD MI300X

Compute (FP32-eq)

1.72x

Need 1.72x MI300X

FP32 Compute

0.55x

MI300X is 1.81x faster

VRAM

1.00x

MI300X has 1.00x more

Memory Bandwidth

1.51x

Need 1.51x MI300X

To match 1x AMD MI300X

AMD MI250X

Compute (FP32-eq)

3.41x

Need 3.41x MI250X

FP32 Compute

3.40x

Need 3.40x MI250X

VRAM

1.50x

Need 1.50x MI250X

Memory Bandwidth

1.62x

Need 1.62x MI250X

Google TPU v5e

Compute (FP32-eq)

6.63x

Need 6.63x TPU v5e

NVIDIA B200

Compute (FP32-eq)

0.58x

B200 is 1.72x faster

FP32 Compute

1.81x

Need 1.81x B200

VRAM

1.00x

B200 has 1.00x more

Memory Bandwidth

0.66x

B200 has 1.51x more

Pricing

Price Type	MI250X	TPU v5e	B200	MI300X
CAPEX (Street Price)	$12,000	—	$70,000	$35,000
OPEX (per hour)	$2.00/hr	$1.80/hr	—	$10.40/hr
Price per TFLOPs (FP32-eq)	$63	—	$62	$54