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 L4	Intel Habana Gaudi 3	NVIDIA GB200 NVL72
Architecture	Ada Lovelace	Gaudi Gen3	Blackwell
Form Factor	PCIe	OAM	Custom
VRAM	24 GB	128 GB	13824 GB
Memory Bandwidth	300 GB/s	3,700 GB/s	130,000 GB/s
TFLOPs (FP32)	30.3	—	6,480
TFLOPs (FP16)	242	—	360,000
TFLOPs	121	1,835	360,000
TFLOPs (FP8)	485	3,670	—
TDP	72 W	900 W	120000 W
Launch Date	Mar 2023	Apr 2024	Mar 2024

Efficiency Metrics

Metric	L4	Gaudi 3	GB200 NVL72
TFLOPs per Watt (FP32-eq)	0.84	1.02	1.50
Memory Bandwidth per GB	12.5 GB/s	28.9 GB/s	9.4 GB/s

Performance Equivalence

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

To match 1x NVIDIA L4

Intel Habana Gaudi 3

Compute (FP32-eq)

0.07x

Gaudi 3 is 15.17x faster

VRAM

0.19x

Gaudi 3 has 5.33x more

Memory Bandwidth

0.08x

Gaudi 3 has 12.33x more

NVIDIA GB200 NVL72

Compute (FP32-eq)

0.00x

GB200 NVL72 is 2975.21x faster

FP32 Compute

0.00x

GB200 NVL72 is 213.86x faster

VRAM

0.00x

GB200 NVL72 has 576.00x more

Memory Bandwidth

0.00x

GB200 NVL72 has 433.33x more

To match 1x Intel Habana Gaudi 3

NVIDIA L4

Compute (FP32-eq)

15.17x

Need 15.17x L4

VRAM

5.33x

Need 5.33x L4

Memory Bandwidth

12.33x

Need 12.33x L4

NVIDIA GB200 NVL72

Compute (FP32-eq)

0.01x

GB200 NVL72 is 196.19x faster

VRAM

0.01x

GB200 NVL72 has 108.00x more

Memory Bandwidth

0.03x

GB200 NVL72 has 35.14x more

To match 1x NVIDIA GB200 NVL72

NVIDIA L4

Compute (FP32-eq)

2975.21x

Need 2975.21x L4

FP32 Compute

213.86x

Need 213.86x L4

VRAM

576.00x

Need 576.00x L4

Memory Bandwidth

433.33x

Need 433.33x L4

Intel Habana Gaudi 3

Compute (FP32-eq)

196.19x

Need 196.19x Gaudi 3

VRAM

108.00x

Need 108.00x Gaudi 3

Memory Bandwidth

35.14x

Need 35.14x Gaudi 3

Pricing

Price Type	L4	Gaudi 3	GB200 NVL72
CAPEX (Street Price)	$4,000	$15,000	—
OPEX (per hour)	$0.80/hr	$1.20/hr	—
Price per TFLOPs (FP32-eq)	$66	$16	—