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 L4	Meta MTIA v1	NVIDIA GB200 Superchip	AMD Radeon PRO W7900
Architecture	Ada Lovelace	Meta Training & Inference Accelerator	Blackwell	RDNA 3
Form Factor	PCIe	—	SXM	PCIe
VRAM	24 GB	128 GB	384 GB	48 GB
Memory Bandwidth	300 GB/s	—	8,000 GB/s	864 GB/s
TFLOPs (FP32)	30.3	—	90	61
TFLOPs (FP16)	242	—	5,000	122
TFLOPs	121	—	5,000	122
TFLOPs (FP8)	485	—	—	—
TDP	72 W	400 W	2700 W	295 W
Launch Date	Mar 2023	May 2023	Mar 2024	Apr 2023

Efficiency Metrics

Metric	L4	MTIA v1	GB200 Superchip	Radeon PRO W7900
TFLOPs per Watt (FP32-eq)	0.84	—	0.93	0.21
Memory Bandwidth per GB	12.5 GB/s	—	20.8 GB/s	18.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA L4

Meta MTIA v1

VRAM

0.19x

MTIA v1 has 5.33x more

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.02x

GB200 Superchip is 41.32x faster

FP32 Compute

0.34x

GB200 Superchip is 2.97x faster

VRAM

0.06x

GB200 Superchip has 16.00x more

Memory Bandwidth

0.04x

GB200 Superchip has 26.67x more

AMD Radeon PRO W7900

Compute (FP32-eq)

0.99x

Radeon PRO W7900 is 1.01x faster

FP32 Compute

0.50x

Radeon PRO W7900 is 2.01x faster

VRAM

0.50x

Radeon PRO W7900 has 2.00x more

Memory Bandwidth

0.35x

Radeon PRO W7900 has 2.88x more

To match 1x Meta MTIA v1

NVIDIA L4

VRAM

5.33x

Need 5.33x L4

NVIDIA GB200 Superchip

VRAM

0.33x

GB200 Superchip has 3.00x more

AMD Radeon PRO W7900

VRAM

2.67x

Need 2.67x Radeon PRO W7900

To match 1x NVIDIA GB200 Superchip

NVIDIA L4

Compute (FP32-eq)

41.32x

Need 41.32x L4

FP32 Compute

2.97x

Need 2.97x L4

VRAM

16.00x

Need 16.00x L4

Memory Bandwidth

26.67x

Need 26.67x L4

Meta MTIA v1

VRAM

3.00x

Need 3.00x MTIA v1

AMD Radeon PRO W7900

Compute (FP32-eq)

40.98x

Need 40.98x Radeon PRO W7900

FP32 Compute

1.48x

Need 1.48x Radeon PRO W7900

VRAM

8.00x

Need 8.00x Radeon PRO W7900

Memory Bandwidth

9.26x

Need 9.26x Radeon PRO W7900

To match 1x AMD Radeon PRO W7900

NVIDIA L4

Compute (FP32-eq)

1.01x

Need 1.01x L4

FP32 Compute

2.01x

Need 2.01x L4

VRAM

2.00x

Need 2.00x L4

Memory Bandwidth

2.88x

Need 2.88x L4

Meta MTIA v1

VRAM

0.38x

MTIA v1 has 2.67x more

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.02x

GB200 Superchip is 40.98x faster

FP32 Compute

0.68x

GB200 Superchip is 1.48x faster

VRAM

0.13x

GB200 Superchip has 8.00x more

Memory Bandwidth

0.11x

GB200 Superchip has 9.26x more

Pricing

Price Type	L4	MTIA v1	GB200 Superchip	Radeon PRO W7900
CAPEX (Street Price)	$4,000	—	—	—
OPEX (per hour)	$0.80/hr	—	—	—
Price per TFLOPs (FP32-eq)	$66	—	—	—