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	Microsoft Maia 100	Cambricon MLU370	NVIDIA GB200 Superchip	Etched Sohu
Architecture	Maia	MLU Core	Blackwell	Transformer ASIC
Form Factor	—	—	SXM	PCIe
VRAM	64 GB	48 GB	384 GB	144 GB
Memory Bandwidth	—	—	8,000 GB/s	4,800 GB/s
TFLOPs (FP32)	—	—	90	10,000
TFLOPs (FP16)	—	—	5,000	—
TFLOPs	700	256	5,000	10,000
TFLOPs (FP8)	—	—	—	—
TDP	500 W	300 W	2700 W	700 W
Launch Date	Nov 2023	Nov 2022	Mar 2024	Jun 2024

Efficiency Metrics

Metric	Maia 100	MLU370	GB200 Superchip	Sohu
TFLOPs per Watt (FP32-eq)	0.70	0.43	0.93	14.29
Memory Bandwidth per GB	—	—	20.8 GB/s	33.3 GB/s

Performance Equivalence

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

To match 1x Microsoft Maia 100

Cambricon MLU370

Compute (FP32-eq)

2.73x

Need 2.73x MLU370

VRAM

1.33x

Need 1.33x MLU370

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.14x

GB200 Superchip is 7.14x faster

VRAM

0.17x

GB200 Superchip has 6.00x more

Etched Sohu

Compute (FP32-eq)

0.04x

Sohu is 28.57x faster

VRAM

0.44x

Sohu has 2.25x more

To match 1x Cambricon MLU370

Microsoft Maia 100

Compute (FP32-eq)

0.37x

Maia 100 is 2.73x faster

VRAM

0.75x

Maia 100 has 1.33x more

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.05x

GB200 Superchip is 19.53x faster

VRAM

0.13x

GB200 Superchip has 8.00x more

Etched Sohu

Compute (FP32-eq)

0.01x

Sohu is 78.13x faster

VRAM

0.33x

Sohu has 3.00x more

To match 1x NVIDIA GB200 Superchip

Microsoft Maia 100

Compute (FP32-eq)

7.14x

Need 7.14x Maia 100

VRAM

6.00x

Need 6.00x Maia 100

Cambricon MLU370

Compute (FP32-eq)

19.53x

Need 19.53x MLU370

VRAM

8.00x

Need 8.00x MLU370

Etched Sohu

Compute (FP32-eq)

0.25x

Sohu is 4.00x faster

FP32 Compute

0.01x

Sohu is 111.11x faster

VRAM

2.67x

Need 2.67x Sohu

Memory Bandwidth

1.67x

Need 1.67x Sohu

To match 1x Etched Sohu

Microsoft Maia 100

Compute (FP32-eq)

28.57x

Need 28.57x Maia 100

VRAM

2.25x

Need 2.25x Maia 100

Cambricon MLU370

Compute (FP32-eq)

78.13x

Need 78.13x MLU370

VRAM

3.00x

Need 3.00x MLU370

NVIDIA GB200 Superchip

Compute (FP32-eq)

4.00x

Need 4.00x GB200 Superchip

FP32 Compute

111.11x

Need 111.11x GB200 Superchip

VRAM

0.38x

GB200 Superchip has 2.67x more

Memory Bandwidth

0.60x

GB200 Superchip has 1.67x more

Pricing

Price Type	Maia 100	MLU370	GB200 Superchip	Sohu
CAPEX (Street Price)	—	—	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	—	—	—	—