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	Biren Technology BR100	NVIDIA H100 SXM	Google TPU v6e (Trillium)
Architecture	Ampere	Biren GPU	Hopper	Trillium
Form Factor	PCIe	—	SXM	Custom
VRAM	48 GB	64 GB	80 GB	32 GB
Memory Bandwidth	696 GB/s	—	3,350 GB/s	1,600 GB/s
TFLOPs (FP32)	37.4	—	67	459
TFLOPs (FP16)	150	1,000	1,979	918
TFLOPs	150	—	1,979	918
TFLOPs (FP8)	—	—	3,958	—
TDP	300 W	550 W	700 W	200 W
Launch Date	Oct 2020	Aug 2022	Sep 2022	May 2024

Efficiency Metrics

Metric	A40	BR100	H100 SXM	TPU v6e (Trillium)
TFLOPs per Watt (FP32-eq)	0.25	—	1.41	2.29
Memory Bandwidth per GB	14.5 GB/s	—	41.9 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

Biren Technology BR100

VRAM

0.75x

BR100 has 1.33x more

NVIDIA H100 SXM

Compute (FP32-eq)

0.08x

H100 SXM is 13.19x faster

FP32 Compute

0.56x

H100 SXM is 1.79x faster

VRAM

0.60x

H100 SXM has 1.67x more

Memory Bandwidth

0.21x

H100 SXM has 4.81x 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 Biren Technology BR100

NVIDIA A40

VRAM

1.33x

Need 1.33x A40

NVIDIA H100 SXM

VRAM

0.80x

H100 SXM has 1.25x more

Google TPU v6e (Trillium)

VRAM

2.00x

Need 2.00x TPU v6e (Trillium)

To match 1x NVIDIA H100 SXM

NVIDIA A40

Compute (FP32-eq)

13.19x

Need 13.19x A40

FP32 Compute

1.79x

Need 1.79x A40

VRAM

1.67x

Need 1.67x A40

Memory Bandwidth

4.81x

Need 4.81x A40

Biren Technology BR100

VRAM

1.25x

Need 1.25x BR100

Google TPU v6e (Trillium)

Compute (FP32-eq)

2.16x

Need 2.16x TPU v6e (Trillium)

FP32 Compute

0.15x

TPU v6e (Trillium) is 6.85x faster

VRAM

2.50x

Need 2.50x TPU v6e (Trillium)

Memory Bandwidth

2.09x

Need 2.09x 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

Biren Technology BR100

VRAM

0.50x

BR100 has 2.00x more

NVIDIA H100 SXM

Compute (FP32-eq)

0.46x

H100 SXM is 2.16x faster

FP32 Compute

6.85x

Need 6.85x H100 SXM

VRAM

0.40x

H100 SXM has 2.50x more

Memory Bandwidth

0.48x

H100 SXM has 2.09x more

Pricing

Price Type	A40	BR100	H100 SXM	TPU v6e (Trillium)
CAPEX (Street Price)	—	—	$30,000	—
OPEX (per hour)	—	—	$3.50/hr	—
Price per TFLOPs (FP32-eq)	—	—	$30	—