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	AMD MI210	AMD MI325X	AWS Trainium2	AMD MI300A
Architecture	CDNA 2	CDNA 3.5	Trainium Gen2	CDNA 3
Form Factor	PCIe	OAM	—	OAM
VRAM	64 GB	256 GB	—	128 GB
Memory Bandwidth	1,638 GB/s	6,000 GB/s	—	5,300 GB/s
TFLOPs (FP32)	45.3	175	—	122.6
TFLOPs (FP16)	181	1,400	—	980.6
TFLOPs	181	1,400	680	980.6
TFLOPs (FP8)	—	2,800	—	—
TDP	300 W	800 W	300 W	760 W
Launch Date	Jan 2022	Oct 2024	Nov 2023	Dec 2023

Efficiency Metrics

Metric	MI210	MI325X	Trainium2	MI300A
TFLOPs per Watt (FP32-eq)	0.30	0.88	1.13	0.65
Memory Bandwidth per GB	25.6 GB/s	23.4 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 AMD MI210

AMD MI325X

Compute (FP32-eq)

0.13x

MI325X is 7.73x faster

FP32 Compute

0.26x

MI325X is 3.86x faster

VRAM

0.25x

MI325X has 4.00x more

Memory Bandwidth

0.27x

MI325X has 3.66x more

AWS Trainium2

Compute (FP32-eq)

0.27x

Trainium2 is 3.76x faster

AMD MI300A

Compute (FP32-eq)

0.18x

MI300A is 5.42x faster

FP32 Compute

0.37x

MI300A is 2.71x faster

VRAM

0.50x

MI300A has 2.00x more

Memory Bandwidth

0.31x

MI300A has 3.24x more

To match 1x AMD MI325X

AMD MI210

Compute (FP32-eq)

7.73x

Need 7.73x MI210

FP32 Compute

3.86x

Need 3.86x MI210

VRAM

4.00x

Need 4.00x MI210

Memory Bandwidth

3.66x

Need 3.66x MI210

AWS Trainium2

Compute (FP32-eq)

2.06x

Need 2.06x Trainium2

AMD MI300A

Compute (FP32-eq)

1.43x

Need 1.43x MI300A

FP32 Compute

1.43x

Need 1.43x MI300A

VRAM

2.00x

Need 2.00x MI300A

Memory Bandwidth

1.13x

Need 1.13x MI300A

To match 1x AWS Trainium2

AMD MI210

Compute (FP32-eq)

3.76x

Need 3.76x MI210

AMD MI325X

Compute (FP32-eq)

0.49x

MI325X is 2.06x faster

AMD MI300A

Compute (FP32-eq)

0.69x

MI300A is 1.44x faster

To match 1x AMD MI300A

AMD MI210

Compute (FP32-eq)

5.42x

Need 5.42x MI210

FP32 Compute

2.71x

Need 2.71x MI210

VRAM

2.00x

Need 2.00x MI210

Memory Bandwidth

3.24x

Need 3.24x MI210

AMD MI325X

Compute (FP32-eq)

0.70x

MI325X is 1.43x faster

FP32 Compute

0.70x

MI325X is 1.43x faster

VRAM

0.50x

MI325X has 2.00x more

Memory Bandwidth

0.88x

MI325X has 1.13x more

AWS Trainium2

Compute (FP32-eq)

1.44x

Need 1.44x Trainium2

Pricing

Price Type	MI210	MI325X	Trainium2	MI300A
CAPEX (Street Price)	$6,000	$45,000	—	—
OPEX (per hour)	—	—	$1.34/hr	—
Price per TFLOPs (FP32-eq)	$66	$64	—	—