Compare XPUs

Select up to 5 XPUs to compare side-by-side

Select XPUs to Compare

Clear all (3)Jump to results

Filter by Vendor

Showing 72 XPUs • 3 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	AMD Radeon PRO W7900	NVIDIA GeForce RTX 5090
Architecture	Blackwell	RDNA 3	Blackwell
Form Factor	SXM	PCIe	PCIe
VRAM	192 GB	48 GB	32 GB
Memory Bandwidth	8,000 GB/s	864 GB/s	1,792 GB/s
TFLOPs (FP32)	90	61	104.8
TFLOPs (FP16)	2,250	122	209.5
TFLOPs	2,250	122	209.5
TFLOPs (FP8)	4,500	—	—
TDP	1000 W	295 W	575 W
Launch Date	Mar 2024	Apr 2023	Jan 2025

Efficiency Metrics

Metric	B200	Radeon PRO W7900	GeForce RTX 5090
TFLOPs per Watt (FP32-eq)	1.13	0.21	0.18
Memory Bandwidth per GB	41.7 GB/s	18.0 GB/s	56.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA B200

AMD Radeon PRO W7900

Compute (FP32-eq)

18.44x

Need 18.44x Radeon PRO W7900

FP32 Compute

1.48x

Need 1.48x Radeon PRO W7900

VRAM

4.00x

Need 4.00x Radeon PRO W7900

Memory Bandwidth

9.26x

Need 9.26x Radeon PRO W7900

NVIDIA GeForce RTX 5090

Compute (FP32-eq)

10.74x

Need 10.74x GeForce RTX 5090

FP32 Compute

0.86x

GeForce RTX 5090 is 1.16x faster

VRAM

6.00x

Need 6.00x GeForce RTX 5090

Memory Bandwidth

4.46x

Need 4.46x GeForce RTX 5090

To match 1x AMD Radeon PRO W7900

NVIDIA B200

Compute (FP32-eq)

0.05x

B200 is 18.44x faster

FP32 Compute

0.68x

B200 is 1.48x faster

VRAM

0.25x

B200 has 4.00x more

Memory Bandwidth

0.11x

B200 has 9.26x more

NVIDIA GeForce RTX 5090

Compute (FP32-eq)

0.58x

GeForce RTX 5090 is 1.72x faster

FP32 Compute

0.58x

GeForce RTX 5090 is 1.72x faster

VRAM

1.50x

Need 1.50x GeForce RTX 5090

Memory Bandwidth

0.48x

GeForce RTX 5090 has 2.07x more

To match 1x NVIDIA GeForce RTX 5090

NVIDIA B200

Compute (FP32-eq)

0.09x

B200 is 10.74x faster

FP32 Compute

1.16x

Need 1.16x B200

VRAM

0.17x

B200 has 6.00x more

Memory Bandwidth

0.22x

B200 has 4.46x more

AMD Radeon PRO W7900

Compute (FP32-eq)

1.72x

Need 1.72x Radeon PRO W7900

FP32 Compute

1.72x

Need 1.72x Radeon PRO W7900

VRAM

0.67x

Radeon PRO W7900 has 1.50x more

Memory Bandwidth

2.07x

Need 2.07x Radeon PRO W7900

Pricing

Price Type	B200	Radeon PRO W7900	GeForce RTX 5090
CAPEX (Street Price)	$70,000	—	—
OPEX (per hour)	—	—	—
Price per TFLOPs (FP32-eq)	$62	—	—