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 A40	Alibaba Hanguang 800	NVIDIA GB200 Superchip	Google TPU v6e (Trillium)
Architecture	Ampere	XuanTie	Blackwell	Trillium
Form Factor	PCIe	—	SXM	Custom
VRAM	48 GB	—	384 GB	32 GB
Memory Bandwidth	696 GB/s	—	8,000 GB/s	1,600 GB/s
TFLOPs (FP32)	37.4	—	90	459
TFLOPs (FP16)	150	—	5,000	918
TFLOPs	150	—	5,000	918
TFLOPs (FP8)	—	—	—	—
TDP	300 W	150 W	2700 W	200 W
Launch Date	Oct 2020	Sep 2019	Mar 2024	May 2024

Efficiency Metrics

Metric	A40	Hanguang 800	GB200 Superchip	TPU v6e (Trillium)
TFLOPs per Watt (FP32-eq)	0.25	—	0.93	2.29
Memory Bandwidth per GB	14.5 GB/s	—	20.8 GB/s	50.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA A40

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.03x

GB200 Superchip is 33.33x faster

FP32 Compute

0.42x

GB200 Superchip is 2.41x faster

VRAM

0.13x

GB200 Superchip has 8.00x more

Memory Bandwidth

0.09x

GB200 Superchip has 11.49x more

Google TPU v6e (Trillium)

Compute (FP32-eq)

0.16x

TPU v6e (Trillium) is 6.12x faster

FP32 Compute

0.08x

TPU v6e (Trillium) is 12.27x faster

VRAM

1.50x

Need 1.50x TPU v6e (Trillium)

Memory Bandwidth

0.43x

TPU v6e (Trillium) has 2.30x more

To match 1x NVIDIA GB200 Superchip

NVIDIA A40

Compute (FP32-eq)

33.33x

Need 33.33x A40

FP32 Compute

2.41x

Need 2.41x A40

VRAM

8.00x

Need 8.00x A40

Memory Bandwidth

11.49x

Need 11.49x A40

Google TPU v6e (Trillium)

Compute (FP32-eq)

5.45x

Need 5.45x TPU v6e (Trillium)

FP32 Compute

0.20x

TPU v6e (Trillium) is 5.10x faster

VRAM

12.00x

Need 12.00x TPU v6e (Trillium)

Memory Bandwidth

5.00x

Need 5.00x TPU v6e (Trillium)

To match 1x Google TPU v6e (Trillium)

NVIDIA A40

Compute (FP32-eq)

6.12x

Need 6.12x A40

FP32 Compute

12.27x

Need 12.27x A40

VRAM

0.67x

A40 has 1.50x more

Memory Bandwidth

2.30x

Need 2.30x A40

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.18x

GB200 Superchip is 5.45x faster

FP32 Compute

5.10x

Need 5.10x GB200 Superchip

VRAM

0.08x

GB200 Superchip has 12.00x more

Memory Bandwidth

0.20x

GB200 Superchip has 5.00x more

Pricing

Price Type	A40	Hanguang 800	GB200 Superchip	TPU v6e (Trillium)
CAPEX (Street Price)	—	—	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	—	—	—	—