Compare XPUs

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

Select XPUs to Compare

Clear all (3)Jump to results

Filter by Vendor

Showing 128 XPUs • 3 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	NVIDIA RTX 6000 Blackwell	NVIDIA A100 SXM	NVIDIA GeForce GTX 1660 Super
Architecture	Blackwell	Ampere	Turing
Form Factor	PCIe	SXM	PCIe
VRAM	96 GB	80 GB	6 GB
Memory Bandwidth	1,536 GB/s	2,039 GB/s	336 GB/s
TFLOPs (FP32)	181	19.5	5.027
TFLOPs (FP16)	—	312	—
TFLOPs	362	312	5.027
TFLOPs (FP8)	—	—	—
TDP	300 W	400 W	125 W
Launch Date	Jan 2025	May 2020	Oct 2019

Efficiency Metrics

Metric	RTX 6000 Blackwell	A100 SXM	GeForce GTX 1660 Super
TFLOPs per Watt (FP32-eq)	0.60	0.39	0.04
Memory Bandwidth per GB	16.0 GB/s	25.5 GB/s	56.0 GB/s

Performance Equivalence

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

To match 1x NVIDIA RTX 6000 Blackwell

NVIDIA A100 SXM

Compute (FP32-eq)

1.16x

Need 1.16x A100 SXM

FP32 Compute

9.28x

Need 9.28x A100 SXM

VRAM

1.20x

Need 1.20x A100 SXM

Memory Bandwidth

0.75x

A100 SXM has 1.33x more

NVIDIA GeForce GTX 1660 Super

Compute (FP32-eq)

36.01x

Need 36.01x GeForce GTX 1660 Super

FP32 Compute

36.01x

Need 36.01x GeForce GTX 1660 Super

VRAM

16.00x

Need 16.00x GeForce GTX 1660 Super

Memory Bandwidth

4.57x

Need 4.57x GeForce GTX 1660 Super

To match 1x NVIDIA A100 SXM

NVIDIA RTX 6000 Blackwell

Compute (FP32-eq)

0.86x

RTX 6000 Blackwell is 1.16x faster

FP32 Compute

0.11x

RTX 6000 Blackwell is 9.28x faster

VRAM

0.83x

RTX 6000 Blackwell has 1.20x more

Memory Bandwidth

1.33x

Need 1.33x RTX 6000 Blackwell

NVIDIA GeForce GTX 1660 Super

Compute (FP32-eq)

31.03x

Need 31.03x GeForce GTX 1660 Super

FP32 Compute

3.88x

Need 3.88x GeForce GTX 1660 Super

VRAM

13.33x

Need 13.33x GeForce GTX 1660 Super

Memory Bandwidth

6.07x

Need 6.07x GeForce GTX 1660 Super

To match 1x NVIDIA GeForce GTX 1660 Super

NVIDIA RTX 6000 Blackwell

Compute (FP32-eq)

0.03x

RTX 6000 Blackwell is 36.01x faster

FP32 Compute

0.03x

RTX 6000 Blackwell is 36.01x faster

VRAM

0.06x

RTX 6000 Blackwell has 16.00x more

Memory Bandwidth

0.22x

RTX 6000 Blackwell has 4.57x more

NVIDIA A100 SXM

Compute (FP32-eq)

0.03x

A100 SXM is 31.03x faster

FP32 Compute

0.26x

A100 SXM is 3.88x faster

VRAM

0.07x

A100 SXM has 13.33x more

Memory Bandwidth

0.16x

A100 SXM has 6.07x more

Pricing

Price Type	RTX 6000 Blackwell	A100 SXM	GeForce GTX 1660 Super
CAPEX (Street Price)	—	$15,000	—
OPEX (per hour)	$1.26/hr	$4.05/hr	$0.04/hr
Price per TFLOPs (FP32-eq)	—	$96	—