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 A100 SXM	NVIDIA L4	AMD MI325X	Cerebras WSE-3
Architecture	Ampere	Ada Lovelace	CDNA 3.5	Wafer Scale Engine
Form Factor	SXM	PCIe	OAM	—
VRAM	80 GB	24 GB	256 GB	44 GB
Memory Bandwidth	2,039 GB/s	300 GB/s	6,000 GB/s	—
TFLOPs (FP32)	19.5	30.3	175	—
TFLOPs (FP16)	312	242	1,400	—
TFLOPs	312	121	1,400	—
TFLOPs (FP8)	—	485	2,800	—
TDP	400 W	72 W	800 W	23000 W
Launch Date	May 2020	Mar 2023	Oct 2024	Mar 2024

Efficiency Metrics

Metric	A100 SXM	L4	MI325X	WSE-3
TFLOPs per Watt (FP32-eq)	0.39	0.84	0.88	—
Memory Bandwidth per GB	25.5 GB/s	12.5 GB/s	23.4 GB/s	—

Performance Equivalence

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

To match 1x NVIDIA A100 SXM

NVIDIA L4

Compute (FP32-eq)

2.58x

Need 2.58x L4

FP32 Compute

0.64x

L4 is 1.55x faster

VRAM

3.33x

Need 3.33x L4

Memory Bandwidth

6.80x

Need 6.80x L4

AMD MI325X

Compute (FP32-eq)

0.22x

MI325X is 4.49x faster

FP32 Compute

0.11x

MI325X is 8.97x faster

VRAM

0.31x

MI325X has 3.20x more

Memory Bandwidth

0.34x

MI325X has 2.94x more

Cerebras WSE-3

VRAM

1.82x

Need 1.82x WSE-3

To match 1x NVIDIA L4

NVIDIA A100 SXM

Compute (FP32-eq)

0.39x

A100 SXM is 2.58x faster

FP32 Compute

1.55x

Need 1.55x A100 SXM

VRAM

0.30x

A100 SXM has 3.33x more

Memory Bandwidth

0.15x

A100 SXM has 6.80x more

AMD MI325X

Compute (FP32-eq)

0.09x

MI325X is 11.57x faster

FP32 Compute

0.17x

MI325X is 5.78x faster

VRAM

0.09x

MI325X has 10.67x more

Memory Bandwidth

0.05x

MI325X has 20.00x more

Cerebras WSE-3

VRAM

0.55x

WSE-3 has 1.83x more

To match 1x AMD MI325X

NVIDIA A100 SXM

Compute (FP32-eq)

4.49x

Need 4.49x A100 SXM

FP32 Compute

8.97x

Need 8.97x A100 SXM

VRAM

3.20x

Need 3.20x A100 SXM

Memory Bandwidth

2.94x

Need 2.94x A100 SXM

NVIDIA L4

Compute (FP32-eq)

11.57x

Need 11.57x L4

FP32 Compute

5.78x

Need 5.78x L4

VRAM

10.67x

Need 10.67x L4

Memory Bandwidth

20.00x

Need 20.00x L4

Cerebras WSE-3

VRAM

5.82x

Need 5.82x WSE-3

To match 1x Cerebras WSE-3

NVIDIA A100 SXM

VRAM

0.55x

A100 SXM has 1.82x more

NVIDIA L4

VRAM

1.83x

Need 1.83x L4

AMD MI325X

VRAM

0.17x

MI325X has 5.82x more

Pricing

Price Type	A100 SXM	L4	MI325X	WSE-3
CAPEX (Street Price)	$15,000	$4,000	$45,000	—
OPEX (per hour)	$4.05/hr	$0.80/hr	—	—
Price per TFLOPs (FP32-eq)	$96	$66	$64	—