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 B200	SambaNova SN40L	NVIDIA GeForce RTX 4080	Google TPU v6e (Trillium)
Architecture	Blackwell	Reconfigurable Dataflow Unit (RDU)	Ada Lovelace	Trillium
Form Factor	SXM	—	PCIe	Custom
VRAM	192 GB	640 GB	16 GB	32 GB
Memory Bandwidth	8,000 GB/s	—	717 GB/s	1,600 GB/s
TFLOPs (FP32)	90	—	48.7	459
TFLOPs (FP16)	2,250	—	97.5	918
TFLOPs	2,250	—	97.5	918
TFLOPs (FP8)	4,500	—	—	—
TDP	1000 W	700 W	320 W	200 W
Launch Date	Mar 2024	Jan 2024	Nov 2022	May 2024

Efficiency Metrics

Metric	B200	SN40L	GeForce RTX 4080	TPU v6e (Trillium)
TFLOPs per Watt (FP32-eq)	1.13	—	0.15	2.29
Memory Bandwidth per GB	41.7 GB/s	—	44.8 GB/s	50.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA B200

SambaNova SN40L

VRAM

0.30x

SN40L has 3.33x more

NVIDIA GeForce RTX 4080

Compute (FP32-eq)

23.08x

Need 23.08x GeForce RTX 4080

FP32 Compute

1.85x

Need 1.85x GeForce RTX 4080

VRAM

12.00x

Need 12.00x GeForce RTX 4080

Memory Bandwidth

11.16x

Need 11.16x GeForce RTX 4080

Google TPU v6e (Trillium)

Compute (FP32-eq)

2.45x

Need 2.45x TPU v6e (Trillium)

FP32 Compute

0.20x

TPU v6e (Trillium) is 5.10x faster

VRAM

6.00x

Need 6.00x TPU v6e (Trillium)

Memory Bandwidth

5.00x

Need 5.00x TPU v6e (Trillium)

To match 1x SambaNova SN40L

NVIDIA B200

VRAM

3.33x

Need 3.33x B200

NVIDIA GeForce RTX 4080

VRAM

40.00x

Need 40.00x GeForce RTX 4080

Google TPU v6e (Trillium)

VRAM

20.00x

Need 20.00x TPU v6e (Trillium)

To match 1x NVIDIA GeForce RTX 4080

NVIDIA B200

Compute (FP32-eq)

0.04x

B200 is 23.08x faster

FP32 Compute

0.54x

B200 is 1.85x faster

VRAM

0.08x

B200 has 12.00x more

Memory Bandwidth

0.09x

B200 has 11.16x more

SambaNova SN40L

VRAM

0.03x

SN40L has 40.00x more

Google TPU v6e (Trillium)

Compute (FP32-eq)

0.11x

TPU v6e (Trillium) is 9.42x faster

FP32 Compute

0.11x

TPU v6e (Trillium) is 9.43x faster

VRAM

0.50x

TPU v6e (Trillium) has 2.00x more

Memory Bandwidth

0.45x

TPU v6e (Trillium) has 2.23x more

To match 1x Google TPU v6e (Trillium)

NVIDIA B200

Compute (FP32-eq)

0.41x

B200 is 2.45x faster

FP32 Compute

5.10x

Need 5.10x B200

VRAM

0.17x

B200 has 6.00x more

Memory Bandwidth

0.20x

B200 has 5.00x more

SambaNova SN40L

VRAM

0.05x

SN40L has 20.00x more

NVIDIA GeForce RTX 4080

Compute (FP32-eq)

9.42x

Need 9.42x GeForce RTX 4080

FP32 Compute

9.43x

Need 9.43x GeForce RTX 4080

VRAM

2.00x

Need 2.00x GeForce RTX 4080

Memory Bandwidth

2.23x

Need 2.23x GeForce RTX 4080

Pricing

Price Type	B200	SN40L	GeForce RTX 4080	TPU v6e (Trillium)
CAPEX (Street Price)	$70,000	—	—	—
OPEX (per hour)	—	—	—	—
Price per TFLOPs (FP32-eq)	$62	—	—	—