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 A10	Google TPU v5e	AWS Inferentia2	AMD MI300A
Architecture	Ampere	TPU v5	Inferentia Gen2	CDNA 3
Form Factor	PCIe	Mezzanine	—	OAM
VRAM	24 GB	—	32 GB	128 GB
Memory Bandwidth	600 GB/s	—	—	5,300 GB/s
TFLOPs (FP32)	31.2	—	—	122.6
TFLOPs (FP16)	125	—	—	980.6
TFLOPs	125	197	190	980.6
TFLOPs (FP8)	—	—	—	—
TDP	150 W	250 W	150 W	760 W
Launch Date	Apr 2021	Aug 2023	Nov 2022	Dec 2023

Efficiency Metrics

Metric	A10	TPU v5e	Inferentia2	MI300A
TFLOPs per Watt (FP32-eq)	0.42	0.39	0.63	0.65
Memory Bandwidth per GB	25.0 GB/s	—	—	41.4 GB/s

Performance Equivalence

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

To match 1x NVIDIA A10

Google TPU v5e

Compute (FP32-eq)

0.63x

TPU v5e is 1.58x faster

AWS Inferentia2

Compute (FP32-eq)

0.66x

Inferentia2 is 1.52x faster

VRAM

0.75x

Inferentia2 has 1.33x more

AMD MI300A

Compute (FP32-eq)

0.13x

MI300A is 7.84x faster

FP32 Compute

0.25x

MI300A is 3.93x faster

VRAM

0.19x

MI300A has 5.33x more

Memory Bandwidth

0.11x

MI300A has 8.83x more

To match 1x Google TPU v5e

NVIDIA A10

Compute (FP32-eq)

1.58x

Need 1.58x A10

AWS Inferentia2

Compute (FP32-eq)

1.04x

Need 1.04x Inferentia2

AMD MI300A

Compute (FP32-eq)

0.20x

MI300A is 4.98x faster

To match 1x AWS Inferentia2

NVIDIA A10

Compute (FP32-eq)

1.52x

Need 1.52x A10

VRAM

1.33x

Need 1.33x A10

Google TPU v5e

Compute (FP32-eq)

0.96x

TPU v5e is 1.04x faster

AMD MI300A

Compute (FP32-eq)

0.19x

MI300A is 5.16x faster

VRAM

0.25x

MI300A has 4.00x more

To match 1x AMD MI300A

NVIDIA A10

Compute (FP32-eq)

7.84x

Need 7.84x A10

FP32 Compute

3.93x

Need 3.93x A10

VRAM

5.33x

Need 5.33x A10

Memory Bandwidth

8.83x

Need 8.83x A10

Google TPU v5e

Compute (FP32-eq)

4.98x

Need 4.98x TPU v5e

AWS Inferentia2

Compute (FP32-eq)

5.16x

Need 5.16x Inferentia2

VRAM

4.00x

Need 4.00x Inferentia2

Pricing

Price Type	A10	TPU v5e	Inferentia2	MI300A
CAPEX (Street Price)	—	—	—	—
OPEX (per hour)	—	$1.80/hr	—	—
Price per TFLOPs (FP32-eq)	—	—	—	—