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 H200 SXM	NVIDIA A40	FuriosaAI Warboy	NVIDIA RTX 6000 Ada Generation
Architecture	Hopper	Ampere	FuriosaAI NPU	Ada Lovelace
Form Factor	SXM	PCIe	—	PCIe
VRAM	141 GB	48 GB	32 GB	48 GB
Memory Bandwidth	4,800 GB/s	696 GB/s	—	960 GB/s
TFLOPs (FP32)	67	37.4	—	91.1
TFLOPs (FP16)	1,979	150	—	182.2
TFLOPs	1,979	150	—	182.2
TFLOPs (FP8)	3,958	—	—	—
TDP	700 W	300 W	80 W	300 W
Launch Date	Nov 2023	Oct 2020	Oct 2021	Dec 2022

Efficiency Metrics

Metric	H200 SXM	A40	Warboy	RTX 6000 Ada Generation
TFLOPs per Watt (FP32-eq)	1.41	0.25	—	0.30
Memory Bandwidth per GB	34.0 GB/s	14.5 GB/s	—	20.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA H200 SXM

NVIDIA A40

Compute (FP32-eq)

13.19x

Need 13.19x A40

FP32 Compute

1.79x

Need 1.79x A40

VRAM

2.94x

Need 2.94x A40

Memory Bandwidth

6.90x

Need 6.90x A40

FuriosaAI Warboy

VRAM

4.41x

Need 4.41x Warboy

NVIDIA RTX 6000 Ada Generation

Compute (FP32-eq)

10.86x

Need 10.86x RTX 6000 Ada Generation

FP32 Compute

0.74x

RTX 6000 Ada Generation is 1.36x faster

VRAM

2.94x

Need 2.94x RTX 6000 Ada Generation

Memory Bandwidth

5.00x

Need 5.00x RTX 6000 Ada Generation

To match 1x NVIDIA A40

NVIDIA H200 SXM

Compute (FP32-eq)

0.08x

H200 SXM is 13.19x faster

FP32 Compute

0.56x

H200 SXM is 1.79x faster

VRAM

0.34x

H200 SXM has 2.94x more

Memory Bandwidth

0.14x

H200 SXM has 6.90x more

FuriosaAI Warboy

VRAM

1.50x

Need 1.50x Warboy

NVIDIA RTX 6000 Ada Generation

Compute (FP32-eq)

0.82x

RTX 6000 Ada Generation is 1.21x faster

FP32 Compute

0.41x

RTX 6000 Ada Generation is 2.44x faster

VRAM

1.00x

RTX 6000 Ada Generation has 1.00x more

Memory Bandwidth

0.72x

RTX 6000 Ada Generation has 1.38x more

To match 1x FuriosaAI Warboy

NVIDIA H200 SXM

VRAM

0.23x

H200 SXM has 4.41x more

NVIDIA A40

VRAM

0.67x

A40 has 1.50x more

NVIDIA RTX 6000 Ada Generation

VRAM

0.67x

RTX 6000 Ada Generation has 1.50x more

To match 1x NVIDIA RTX 6000 Ada Generation

NVIDIA H200 SXM

Compute (FP32-eq)

0.09x

H200 SXM is 10.86x faster

FP32 Compute

1.36x

Need 1.36x H200 SXM

VRAM

0.34x

H200 SXM has 2.94x more

Memory Bandwidth

0.20x

H200 SXM has 5.00x more

NVIDIA A40

Compute (FP32-eq)

1.21x

Need 1.21x A40

FP32 Compute

2.44x

Need 2.44x A40

VRAM

1.00x

A40 has 1.00x more

Memory Bandwidth

1.38x

Need 1.38x A40

FuriosaAI Warboy

VRAM

1.50x

Need 1.50x Warboy

Pricing

Price Type	H200 SXM	A40	Warboy	RTX 6000 Ada Generation
CAPEX (Street Price)	$40,000	—	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	$40	—	—	—