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	AMD MI325X	NVIDIA GB200 NVL72	Huawei Ascend 910B
Architecture	Ampere	CDNA 3.5	Blackwell	Da Vinci
Form Factor	PCIe	OAM	Custom	—
VRAM	24 GB	256 GB	13824 GB	64 GB
Memory Bandwidth	600 GB/s	6,000 GB/s	130,000 GB/s	—
TFLOPs (FP32)	31.2	175	6,480	—
TFLOPs (FP16)	125	1,400	360,000	640
TFLOPs	125	1,400	360,000	—
TFLOPs (FP8)	—	2,800	—	—
TDP	150 W	800 W	120000 W	400 W
Launch Date	Apr 2021	Oct 2024	Mar 2024	Aug 2023

Efficiency Metrics

Metric	A10	MI325X	GB200 NVL72	Ascend 910B
TFLOPs per Watt (FP32-eq)	0.42	0.88	1.50	—
Memory Bandwidth per GB	25.0 GB/s	23.4 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 A10

AMD MI325X

Compute (FP32-eq)

0.09x

MI325X is 11.20x faster

FP32 Compute

0.18x

MI325X is 5.61x faster

VRAM

0.09x

MI325X has 10.67x more

Memory Bandwidth

0.10x

MI325X has 10.00x more

NVIDIA GB200 NVL72

Compute (FP32-eq)

0.00x

GB200 NVL72 is 2880.00x faster

FP32 Compute

0.00x

GB200 NVL72 is 207.69x faster

VRAM

0.00x

GB200 NVL72 has 576.00x more

Memory Bandwidth

0.00x

GB200 NVL72 has 216.67x more

Huawei Ascend 910B

VRAM

0.38x

Ascend 910B has 2.67x more

To match 1x AMD MI325X

NVIDIA A10

Compute (FP32-eq)

11.20x

Need 11.20x A10

FP32 Compute

5.61x

Need 5.61x A10

VRAM

10.67x

Need 10.67x A10

Memory Bandwidth

10.00x

Need 10.00x A10

NVIDIA GB200 NVL72

Compute (FP32-eq)

0.00x

GB200 NVL72 is 257.14x faster

FP32 Compute

0.03x

GB200 NVL72 is 37.03x faster

VRAM

0.02x

GB200 NVL72 has 54.00x more

Memory Bandwidth

0.05x

GB200 NVL72 has 21.67x more

Huawei Ascend 910B

VRAM

4.00x

Need 4.00x Ascend 910B

To match 1x NVIDIA GB200 NVL72

NVIDIA A10

Compute (FP32-eq)

2880.00x

Need 2880.00x A10

FP32 Compute

207.69x

Need 207.69x A10

VRAM

576.00x

Need 576.00x A10

Memory Bandwidth

216.67x

Need 216.67x A10

AMD MI325X

Compute (FP32-eq)

257.14x

Need 257.14x MI325X

FP32 Compute

37.03x

Need 37.03x MI325X

VRAM

54.00x

Need 54.00x MI325X

Memory Bandwidth

21.67x

Need 21.67x MI325X

Huawei Ascend 910B

VRAM

216.00x

Need 216.00x Ascend 910B

To match 1x Huawei Ascend 910B

NVIDIA A10

VRAM

2.67x

Need 2.67x A10

AMD MI325X

VRAM

0.25x

MI325X has 4.00x more

NVIDIA GB200 NVL72

VRAM

0.00x

GB200 NVL72 has 216.00x more

Pricing

Price Type	A10	MI325X	GB200 NVL72	Ascend 910B
CAPEX (Street Price)	—	$45,000	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	—	$64	—	—