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 B200	Tenstorrent Wormhole	Meta MTIA v1	NVIDIA RTX 5000 Ada Generation
Architecture	Blackwell	Tensix Core	Meta Training & Inference Accelerator	Ada Lovelace
Form Factor	SXM	—	—	PCIe
VRAM	192 GB	24 GB	128 GB	32 GB
Memory Bandwidth	8,000 GB/s	—	—	576 GB/s
TFLOPs (FP32)	90	—	—	65.3
TFLOPs (FP16)	2,250	—	—	130.6
TFLOPs	2,250	364	—	130.6
TFLOPs (FP8)	4,500	—	—	—
TDP	1000 W	160 W	400 W	250 W
Launch Date	Mar 2024	Oct 2023	May 2023	Aug 2023

Efficiency Metrics

Metric	B200	Wormhole	MTIA v1	RTX 5000 Ada Generation
TFLOPs per Watt (FP32-eq)	1.13	1.14	—	0.26
Memory Bandwidth per GB	41.7 GB/s	—	—	18.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA B200

Tenstorrent Wormhole

Compute (FP32-eq)

6.18x

Need 6.18x Wormhole

VRAM

8.00x

Need 8.00x Wormhole

Meta MTIA v1

VRAM

1.50x

Need 1.50x MTIA v1

NVIDIA RTX 5000 Ada Generation

Compute (FP32-eq)

17.23x

Need 17.23x RTX 5000 Ada Generation

FP32 Compute

1.38x

Need 1.38x RTX 5000 Ada Generation

VRAM

6.00x

Need 6.00x RTX 5000 Ada Generation

Memory Bandwidth

13.89x

Need 13.89x RTX 5000 Ada Generation

To match 1x Tenstorrent Wormhole

NVIDIA B200

Compute (FP32-eq)

0.16x

B200 is 6.18x faster

VRAM

0.13x

B200 has 8.00x more

Meta MTIA v1

VRAM

0.19x

MTIA v1 has 5.33x more

NVIDIA RTX 5000 Ada Generation

Compute (FP32-eq)

2.79x

Need 2.79x RTX 5000 Ada Generation

VRAM

0.75x

RTX 5000 Ada Generation has 1.33x more

To match 1x Meta MTIA v1

NVIDIA B200

VRAM

0.67x

B200 has 1.50x more

Tenstorrent Wormhole

VRAM

5.33x

Need 5.33x Wormhole

NVIDIA RTX 5000 Ada Generation

VRAM

4.00x

Need 4.00x RTX 5000 Ada Generation

To match 1x NVIDIA RTX 5000 Ada Generation

NVIDIA B200

Compute (FP32-eq)

0.06x

B200 is 17.23x faster

FP32 Compute

0.73x

B200 is 1.38x faster

VRAM

0.17x

B200 has 6.00x more

Memory Bandwidth

0.07x

B200 has 13.89x more

Tenstorrent Wormhole

Compute (FP32-eq)

0.36x

Wormhole is 2.79x faster

VRAM

1.33x

Need 1.33x Wormhole

Meta MTIA v1

VRAM

0.25x

MTIA v1 has 4.00x more

Pricing

Price Type	B200	Wormhole	MTIA v1	RTX 5000 Ada Generation
CAPEX (Street Price)	$70,000	—	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	$62	—	—	—