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 L4	Google TPU v4	NVIDIA B200	Biren Technology BR100
Architecture	Ada Lovelace	TPU v4	Blackwell	Biren GPU
Form Factor	PCIe	Mezzanine	SXM	—
VRAM	24 GB	—	192 GB	64 GB
Memory Bandwidth	300 GB/s	—	8,000 GB/s	—
TFLOPs (FP32)	30.3	—	90	—
TFLOPs (FP16)	242	—	2,250	1,000
TFLOPs	121	275	2,250	—
TFLOPs (FP8)	485	—	4,500	—
TDP	72 W	300 W	1000 W	550 W
Launch Date	Mar 2023	May 2021	Mar 2024	Aug 2022

Efficiency Metrics

Metric	L4	TPU v4	B200	BR100
TFLOPs per Watt (FP32-eq)	0.84	0.46	1.13	—
Memory Bandwidth per GB	12.5 GB/s	—	41.7 GB/s	—

Performance Equivalence

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

To match 1x NVIDIA L4

Google TPU v4

Compute (FP32-eq)

0.44x

TPU v4 is 2.27x faster

NVIDIA B200

Compute (FP32-eq)

0.05x

B200 is 18.60x faster

FP32 Compute

0.34x

B200 is 2.97x faster

VRAM

0.13x

B200 has 8.00x more

Memory Bandwidth

0.04x

B200 has 26.67x more

Biren Technology BR100

VRAM

0.38x

BR100 has 2.67x more

To match 1x Google TPU v4

NVIDIA L4

Compute (FP32-eq)

2.27x

Need 2.27x L4

NVIDIA B200

Compute (FP32-eq)

0.12x

B200 is 8.18x faster

To match 1x NVIDIA B200

NVIDIA L4

Compute (FP32-eq)

18.60x

Need 18.60x L4

FP32 Compute

2.97x

Need 2.97x L4

VRAM

8.00x

Need 8.00x L4

Memory Bandwidth

26.67x

Need 26.67x L4

Google TPU v4

Compute (FP32-eq)

8.18x

Need 8.18x TPU v4

Biren Technology BR100

VRAM

3.00x

Need 3.00x BR100

To match 1x Biren Technology BR100

NVIDIA L4

VRAM

2.67x

Need 2.67x L4

NVIDIA B200

VRAM

0.33x

B200 has 3.00x more

Pricing

Price Type	L4	TPU v4	B200	BR100
CAPEX (Street Price)	$4,000	—	$70,000	—
OPEX (per hour)	$0.80/hr	$3.00/hr	—	—
Price per TFLOPs (FP32-eq)	$66	—	$62	—