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 A100 SXM	NVIDIA GB200 Superchip	NVIDIA H200 PCIe
Architecture	Ampere	Blackwell	Hopper
Form Factor	SXM	SXM	PCIe
VRAM	80 GB	384 GB	141 GB
Memory Bandwidth	2,039 GB/s	8,000 GB/s	4,800 GB/s
TFLOPs (FP32)	19.5	90	51
TFLOPs (FP16)	312	5,000	1,513
TFLOPs	312	5,000	1,513
TFLOPs (FP8)	—	—	—
TDP	400 W	2700 W	350 W
Launch Date	May 2020	Mar 2024	Mar 2024

Efficiency Metrics

Metric	A100 SXM	GB200 Superchip	H200 PCIe
TFLOPs per Watt (FP32-eq)	0.39	0.93	2.16
Memory Bandwidth per GB	25.5 GB/s	20.8 GB/s	34.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA A100 SXM

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.06x

GB200 Superchip is 16.03x faster

FP32 Compute

0.22x

GB200 Superchip is 4.62x faster

VRAM

0.21x

GB200 Superchip has 4.80x more

Memory Bandwidth

0.25x

GB200 Superchip has 3.92x more

NVIDIA H200 PCIe

Compute (FP32-eq)

0.21x

H200 PCIe is 4.85x faster

FP32 Compute

0.38x

H200 PCIe is 2.62x faster

VRAM

0.57x

H200 PCIe has 1.76x more

Memory Bandwidth

0.42x

H200 PCIe has 2.35x more

To match 1x NVIDIA GB200 Superchip

NVIDIA A100 SXM

Compute (FP32-eq)

16.03x

Need 16.03x A100 SXM

FP32 Compute

4.62x

Need 4.62x A100 SXM

VRAM

4.80x

Need 4.80x A100 SXM

Memory Bandwidth

3.92x

Need 3.92x A100 SXM

NVIDIA H200 PCIe

Compute (FP32-eq)

3.30x

Need 3.30x H200 PCIe

FP32 Compute

1.76x

Need 1.76x H200 PCIe

VRAM

2.72x

Need 2.72x H200 PCIe

Memory Bandwidth

1.67x

Need 1.67x H200 PCIe

To match 1x NVIDIA H200 PCIe

NVIDIA A100 SXM

Compute (FP32-eq)

4.85x

Need 4.85x A100 SXM

FP32 Compute

2.62x

Need 2.62x A100 SXM

VRAM

1.76x

Need 1.76x A100 SXM

Memory Bandwidth

2.35x

Need 2.35x A100 SXM

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.30x

GB200 Superchip is 3.30x faster

FP32 Compute

0.57x

GB200 Superchip is 1.76x faster

VRAM

0.37x

GB200 Superchip has 2.72x more

Memory Bandwidth

0.60x

GB200 Superchip has 1.67x more

Pricing

Price Type	A100 SXM	GB200 Superchip	H200 PCIe
CAPEX (Street Price)	$15,000	—	—
OPEX (per hour)	$4.05/hr	—	—
Price per TFLOPs (FP32-eq)	$96	—	—