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 MI250X	Meta MTIA v1	Etched Sohu	AMD Radeon RX 7900 XTX
Architecture	CDNA 2	Meta Training & Inference Accelerator	Transformer ASIC	RDNA 3
Form Factor	OAM	—	PCIe	PCIe
VRAM	128 GB	128 GB	144 GB	24 GB
Memory Bandwidth	3,277 GB/s	—	4,800 GB/s	960 GB/s
TFLOPs (FP32)	47.9	—	10,000	61
TFLOPs (FP16)	383	—	—	122
TFLOPs	383	—	10,000	122
TFLOPs (FP8)	—	—	—	—
TDP	560 W	400 W	700 W	355 W
Launch Date	Nov 2021	May 2023	Jun 2024	Dec 2022

Efficiency Metrics

Metric	MI250X	MTIA v1	Sohu	Radeon RX 7900 XTX
TFLOPs per Watt (FP32-eq)	0.34	—	14.29	0.17
Memory Bandwidth per GB	25.6 GB/s	—	33.3 GB/s	40.0 GB/s

Performance Equivalence

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

To match 1x AMD MI250X

Meta MTIA v1

VRAM

1.00x

MTIA v1 has 1.00x more

Etched Sohu

Compute (FP32-eq)

0.02x

Sohu is 52.22x faster

FP32 Compute

0.00x

Sohu is 208.77x faster

VRAM

0.89x

Sohu has 1.13x more

Memory Bandwidth

0.68x

Sohu has 1.46x more

AMD Radeon RX 7900 XTX

Compute (FP32-eq)

3.14x

Need 3.14x Radeon RX 7900 XTX

FP32 Compute

0.79x

Radeon RX 7900 XTX is 1.27x faster

VRAM

5.33x

Need 5.33x Radeon RX 7900 XTX

Memory Bandwidth

3.41x

Need 3.41x Radeon RX 7900 XTX

To match 1x Meta MTIA v1

AMD MI250X

VRAM

1.00x

MI250X has 1.00x more

Etched Sohu

VRAM

0.89x

Sohu has 1.13x more

AMD Radeon RX 7900 XTX

VRAM

5.33x

Need 5.33x Radeon RX 7900 XTX

To match 1x Etched Sohu

AMD MI250X

Compute (FP32-eq)

52.22x

Need 52.22x MI250X

FP32 Compute

208.77x

Need 208.77x MI250X

VRAM

1.13x

Need 1.13x MI250X

Memory Bandwidth

1.46x

Need 1.46x MI250X

Meta MTIA v1

VRAM

1.13x

Need 1.13x MTIA v1

AMD Radeon RX 7900 XTX

Compute (FP32-eq)

163.93x

Need 163.93x Radeon RX 7900 XTX

FP32 Compute

163.93x

Need 163.93x Radeon RX 7900 XTX

VRAM

6.00x

Need 6.00x Radeon RX 7900 XTX

Memory Bandwidth

5.00x

Need 5.00x Radeon RX 7900 XTX

To match 1x AMD Radeon RX 7900 XTX

AMD MI250X

Compute (FP32-eq)

0.32x

MI250X is 3.14x faster

FP32 Compute

1.27x

Need 1.27x MI250X

VRAM

0.19x

MI250X has 5.33x more

Memory Bandwidth

0.29x

MI250X has 3.41x more

Meta MTIA v1

VRAM

0.19x

MTIA v1 has 5.33x more

Etched Sohu

Compute (FP32-eq)

0.01x

Sohu is 163.93x faster

FP32 Compute

0.01x

Sohu is 163.93x faster

VRAM

0.17x

Sohu has 6.00x more

Memory Bandwidth

0.20x

Sohu has 5.00x more

Pricing

Price Type	MI250X	MTIA v1	Sohu	Radeon RX 7900 XTX
CAPEX (Street Price)	$12,000	—	—	—
OPEX (per hour)	$2.00/hr	—	—	—
Price per TFLOPs (FP32-eq)	$63	—	—	—