Compare XPUs

Select up to 5 XPUs to compare side-by-side

Select XPUs to Compare

Clear all (5)Jump to results

Filter by Vendor

Showing 128 XPUs • 5 selected

Alibaba

Hanguang 800

AMD

MI100

23.1 TFLOPs

AMD

MI210

181 TFLOPs

AMD

MI250X

383 TFLOPs

AMD

MI300X

1,307 TFLOPs

AMD

MI325X

1,400 TFLOPs

AMD

MI350X

2,100 TFLOPs

AMD

MI355X

5,300 TFLOPs

AMD

Radeon Pro V520

23.04 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

FuriosaAI

RNGD (Renegade)

256 TFLOPs

FuriosaAI

Warboy

Google

TPU v4

275 TFLOPs

Google

TPU v5e

197 TFLOPs

Google

TPU v5p

459 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	Cerebras WSE-3	NVIDIA GeForce RTX 4070 Super	NVIDIA B100	NVIDIA GeForce GT 710	NVIDIA RTX 5000
Architecture	Wafer Scale Engine	Ada Lovelace	Blackwell	Kepler	Turing
Form Factor	—	PCIe	SXM	PCIe	PCIe
VRAM	44 GB	12 GB	192 GB	2 GB	16 GB
Memory Bandwidth	—	504 GB/s	8,000 GB/s	14.4 GB/s	448 GB/s
TFLOPs (FP32)	—	—	45	0.366	11.2
TFLOPs (FP16)	—	—	—	—	—
TFLOPs	—	70	1,800	0.366	89.2
TFLOPs (FP8)	—	—	—	—	—
TDP	23000 W	220 W	700 W	19 W	265 W
Launch Date	Mar 2024	Jan 2024	Mar 2024	Mar 2014	Aug 2018

Efficiency Metrics

Metric	WSE-3	GeForce RTX 4070 Super	B100	GeForce GT 710	RTX 5000
TFLOPs per Watt (FP32-eq)	—	0.16	1.29	0.02	0.17
Memory Bandwidth per GB	—	42.0 GB/s	41.7 GB/s	7.2 GB/s	28.0 GB/s

Performance Equivalence

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

To match 1x Cerebras WSE-3

NVIDIA GeForce RTX 4070 Super

VRAM

3.67x

Need 3.67x GeForce RTX 4070 Super

NVIDIA B100

VRAM

0.23x

B100 has 4.36x more

NVIDIA GeForce GT 710

VRAM

22.00x

Need 22.00x GeForce GT 710

NVIDIA RTX 5000

VRAM

2.75x

Need 2.75x RTX 5000

To match 1x NVIDIA GeForce RTX 4070 Super

Cerebras WSE-3

VRAM

0.27x

WSE-3 has 3.67x more

NVIDIA B100

Compute (FP32-eq)

0.04x

B100 is 25.71x faster

VRAM

0.06x

B100 has 16.00x more

Memory Bandwidth

0.06x

B100 has 15.87x more

NVIDIA GeForce GT 710

Compute (FP32-eq)

95.63x

Need 95.63x GeForce GT 710

VRAM

6.00x

Need 6.00x GeForce GT 710

Memory Bandwidth

35.00x

Need 35.00x GeForce GT 710

NVIDIA RTX 5000

Compute (FP32-eq)

0.78x

RTX 5000 is 1.27x faster

VRAM

0.75x

RTX 5000 has 1.33x more

Memory Bandwidth

1.13x

Need 1.13x RTX 5000

To match 1x NVIDIA B100

Cerebras WSE-3

VRAM

4.36x

Need 4.36x WSE-3

NVIDIA GeForce RTX 4070 Super

Compute (FP32-eq)

25.71x

Need 25.71x GeForce RTX 4070 Super

VRAM

16.00x

Need 16.00x GeForce RTX 4070 Super

Memory Bandwidth

15.87x

Need 15.87x GeForce RTX 4070 Super

NVIDIA GeForce GT 710

Compute (FP32-eq)

2459.02x

Need 2459.02x GeForce GT 710

FP32 Compute

122.95x

Need 122.95x GeForce GT 710

VRAM

96.00x

Need 96.00x GeForce GT 710

Memory Bandwidth

555.56x

Need 555.56x GeForce GT 710

NVIDIA RTX 5000

Compute (FP32-eq)

20.18x

Need 20.18x RTX 5000

FP32 Compute

4.02x

Need 4.02x RTX 5000

VRAM

12.00x

Need 12.00x RTX 5000

Memory Bandwidth

17.86x

Need 17.86x RTX 5000

To match 1x NVIDIA GeForce GT 710

Cerebras WSE-3

VRAM

0.05x

WSE-3 has 22.00x more

NVIDIA GeForce RTX 4070 Super

Compute (FP32-eq)

0.01x

GeForce RTX 4070 Super is 95.63x faster

VRAM

0.17x

GeForce RTX 4070 Super has 6.00x more

Memory Bandwidth

0.03x

GeForce RTX 4070 Super has 35.00x more

NVIDIA B100

Compute (FP32-eq)

0.00x

B100 is 2459.02x faster

FP32 Compute

0.01x

B100 is 122.95x faster

VRAM

0.01x

B100 has 96.00x more

Memory Bandwidth

0.00x

B100 has 555.56x more

NVIDIA RTX 5000

Compute (FP32-eq)

0.01x

RTX 5000 is 121.86x faster

FP32 Compute

0.03x

RTX 5000 is 30.60x faster

VRAM

0.13x

RTX 5000 has 8.00x more

Memory Bandwidth

0.03x

RTX 5000 has 31.11x more

To match 1x NVIDIA RTX 5000

Cerebras WSE-3

VRAM

0.36x

WSE-3 has 2.75x more

NVIDIA GeForce RTX 4070 Super

Compute (FP32-eq)

1.27x

Need 1.27x GeForce RTX 4070 Super

VRAM

1.33x

Need 1.33x GeForce RTX 4070 Super

Memory Bandwidth

0.89x

GeForce RTX 4070 Super has 1.13x more

NVIDIA B100

Compute (FP32-eq)

0.05x

B100 is 20.18x faster

FP32 Compute

0.25x

B100 is 4.02x faster

VRAM

0.08x

B100 has 12.00x more

Memory Bandwidth

0.06x

B100 has 17.86x more

NVIDIA GeForce GT 710

Compute (FP32-eq)

121.86x

Need 121.86x GeForce GT 710

FP32 Compute

30.60x

Need 30.60x GeForce GT 710

VRAM

8.00x

Need 8.00x GeForce GT 710

Memory Bandwidth

31.11x

Need 31.11x GeForce GT 710

Pricing

Price Type	WSE-3	GeForce RTX 4070 Super	B100	GeForce GT 710	RTX 5000
CAPEX (Street Price)	—	—	—	—	—
OPEX (per hour)	—	—	—	$0.07/hr	$0.82/hr
Price per TFLOPs (FP32-eq)	—	—	—	—	—