Compare XPUs

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

Select XPUs to Compare

Showing 128 XPUs8 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

Meta

MTIA v1

Microsoft

Maia 100

700 TFLOPs

Moore Threads

MTT S80

NVIDIA

A10

125 TFLOPs

NVIDIA

A100 SXM

312 TFLOPs

NVIDIA

A10G

125 TFLOPs

NVIDIA

A16

71.2 TFLOPs

NVIDIA

A30

165 TFLOPs

NVIDIA

A40

150 TFLOPs

NVIDIA

A800

312 TFLOPs

NVIDIA

B100

1,800 TFLOPs

NVIDIA

B200

2,250 TFLOPs

NVIDIA

GB200

2,500 TFLOPs

NVIDIA

GeForce GT 710

0.366 TFLOPs

NVIDIA

GeForce GT 730

0.693 TFLOPs

NVIDIA

GeForce GTX 1050 Ti

2.138 TFLOPs

NVIDIA

GeForce GTX 1060

4.375 TFLOPs

NVIDIA

GeForce GTX 1070

6.463 TFLOPs

NVIDIA

GeForce GTX 1070 Ti

8.186 TFLOPs

NVIDIA

GeForce GTX 1080

8.873 TFLOPs

NVIDIA

GeForce GTX 1080 Ti

11.34 TFLOPs

NVIDIA

GeForce GTX 1650

2.984 TFLOPs

NVIDIA

GeForce GTX 1660

5.027 TFLOPs

NVIDIA

GeForce GTX 1660 Super

5.027 TFLOPs

NVIDIA

GeForce RTX 2060

6.5 TFLOPs

NVIDIA

GeForce RTX 2060 Super

7.2 TFLOPs

NVIDIA

GeForce RTX 2070

7.5 TFLOPs

NVIDIA

GeForce RTX 2080

10.1 TFLOPs

NVIDIA

GeForce RTX 2080 Ti

13.4 TFLOPs

NVIDIA

GeForce RTX 3050

18.2 TFLOPs

NVIDIA

GeForce RTX 3060

25.4 TFLOPs

NVIDIA

GeForce RTX 3060 Ti

33.2 TFLOPs

NVIDIA

GeForce RTX 3070

40.5 TFLOPs

NVIDIA

GeForce RTX 3070 Ti

43.2 TFLOPs

NVIDIA

GeForce RTX 3080

59.2 TFLOPs

NVIDIA

GeForce RTX 3080 Ti

68.4 TFLOPs

NVIDIA

GeForce RTX 3090

71 TFLOPs

NVIDIA

GeForce RTX 3090 Ti

80 TFLOPs

NVIDIA

GeForce RTX 4060

30 TFLOPs

NVIDIA

GeForce RTX 4060 Ti

44 TFLOPs

NVIDIA

GeForce RTX 4060 Ti 16GB

44 TFLOPs

NVIDIA

GeForce RTX 4070

58.92 TFLOPs

NVIDIA

GeForce RTX 4070 Super

70 TFLOPs

NVIDIA

GeForce RTX 4070 Ti

80.16 TFLOPs

NVIDIA

GeForce RTX 4070 Ti Super

88 TFLOPs

NVIDIA

GeForce RTX 4080

97.76 TFLOPs

NVIDIA

GeForce RTX 4080 Super

104 TFLOPs

NVIDIA

GeForce RTX 4090

165.2 TFLOPs

NVIDIA

GeForce RTX 4090 D

165 TFLOPs

NVIDIA

GeForce RTX 5060

60 TFLOPs

NVIDIA

GeForce RTX 5060 Ti

88 TFLOPs

NVIDIA

GeForce RTX 5070

110 TFLOPs

NVIDIA

GeForce RTX 5070 Ti

133 TFLOPs

NVIDIA

GeForce RTX 5080

171 TFLOPs

NVIDIA

GeForce RTX 5090

318 TFLOPs

NVIDIA

GH200

989 TFLOPs

NVIDIA

H100 SXM

1,979 TFLOPs

NVIDIA

H200 SXM

1,979 TFLOPs

NVIDIA

K80

8.73 TFLOPs

NVIDIA

L4

121 TFLOPs

NVIDIA

L40

181.05 TFLOPs

NVIDIA

L40S

733 TFLOPs

NVIDIA

P100

9.3 TFLOPs

NVIDIA

P4

5.5 TFLOPs

NVIDIA

P40

12 TFLOPs

NVIDIA

Quadro K620

0.768 TFLOPs

NVIDIA

Quadro M4000

2.57 TFLOPs

NVIDIA

Quadro M60

4.825 TFLOPs

NVIDIA

Quadro P1000

1.894 TFLOPs

NVIDIA

Quadro P4000

5.304 TFLOPs

NVIDIA

Quadro P5000

8.873 TFLOPs

NVIDIA

Quadro P600

1.117 TFLOPs

NVIDIA

Quadro P6000

12.63 TFLOPs

NVIDIA

Quadro P620

1.386 TFLOPs

NVIDIA

Quadro RTX 6000

130.5 TFLOPs

NVIDIA

RTX 4000

57.6 TFLOPs

NVIDIA

RTX 4000 SFF Ada Generation

52 TFLOPs

NVIDIA

RTX 4500 Ada Generation

97 TFLOPs

NVIDIA

RTX 5000

89.2 TFLOPs

NVIDIA

RTX 5000 Ada Generation

130 TFLOPs

NVIDIA

RTX 5000 Blackwell

242 TFLOPs

NVIDIA

RTX 5880 Ada Generation

167 TFLOPs

NVIDIA

RTX 6000 Ada Generation

182.5 TFLOPs

NVIDIA

RTX 6000 Blackwell

362 TFLOPs

NVIDIA

RTX A2000

16 TFLOPs

NVIDIA

RTX A4000

38.7 TFLOPs

NVIDIA

RTX A4500

46.6 TFLOPs

NVIDIA

RTX A5000

54.2 TFLOPs

NVIDIA

RTX A6000

77.9 TFLOPs

NVIDIA

T1000

2.5 TFLOPs

NVIDIA

T4

65 TFLOPs

NVIDIA

T4G

65 TFLOPs

NVIDIA

V100S

16.4 TFLOPs

Rebellions

ATOM

SambaNova

SN40L

Tenstorrent

Grayskull

200 TFLOPs

Tenstorrent

Wormhole

364 TFLOPs

Maximum of 5 XPUs can be compared at once. Deselect one to add another.

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

SpecificationTenstorrent WormholeEnflame Technology CloudBlazer T20Google TPU v5pNVIDIA RTX 5000 Ada GenerationNVIDIA T4GNVIDIA Quadro P6000NVIDIA GeForce RTX 3090NVIDIA A100 SXM
ArchitectureTensix CoreGCU (AI Graphics Unit)TPU v5Ada LovelaceTuringPascalAmpereAmpere
Form FactorMezzaninePCIePCIePCIePCIeSXM
VRAM24 GB32 GB95 GB32 GB16 GB24 GB24 GB80 GB
Memory Bandwidth576 GB/s320 GB/s432 GB/s936 GB/s2,039 GB/s
TFLOPs (FP32)65.38.112.6335.619.5
TFLOPs (FP16)312
TFLOPs3644591306512.6371312
TFLOPs (FP8)
TDP160 W250 W400 W250 W70 W250 W350 W400 W
Launch DateOct 2023Nov 2021Aug 2023Mar 2023May 2020Oct 2016Sep 2020May 2020

Efficiency Metrics

MetricWormholeCloudBlazer T20TPU v5pRTX 5000 Ada GenerationT4GQuadro P6000GeForce RTX 3090A100 SXM
TFLOPs per Watt (FP32-eq)1.140.570.260.460.050.100.39
Memory Bandwidth per GB18.0 GB/s20.0 GB/s18.0 GB/s39.0 GB/s25.5 GB/s

Performance Equivalence

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

To match 1x Tenstorrent Wormhole

Enflame Technology CloudBlazer T20
VRAM
0.75x
CloudBlazer T20 has 1.33x more
Google TPU v5p
Compute (FP32-eq)
0.79x
TPU v5p is 1.26x faster
VRAM
0.25x
TPU v5p has 3.96x more
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
2.80x
Need 2.80x RTX 5000 Ada Generation
VRAM
0.75x
RTX 5000 Ada Generation has 1.33x more
NVIDIA T4G
Compute (FP32-eq)
5.60x
Need 5.60x T4G
VRAM
1.50x
Need 1.50x T4G
NVIDIA Quadro P6000
Compute (FP32-eq)
14.41x
Need 14.41x Quadro P6000
VRAM
1.00x
Quadro P6000 has 1.00x more
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
5.13x
Need 5.13x GeForce RTX 3090
VRAM
1.00x
GeForce RTX 3090 has 1.00x more
NVIDIA A100 SXM
Compute (FP32-eq)
1.17x
Need 1.17x A100 SXM
VRAM
0.30x
A100 SXM has 3.33x more

To match 1x Enflame Technology CloudBlazer T20

Tenstorrent Wormhole
VRAM
1.33x
Need 1.33x Wormhole
Google TPU v5p
VRAM
0.34x
TPU v5p has 2.97x more
NVIDIA RTX 5000 Ada Generation
VRAM
1.00x
RTX 5000 Ada Generation has 1.00x more
NVIDIA T4G
VRAM
2.00x
Need 2.00x T4G
NVIDIA Quadro P6000
VRAM
1.33x
Need 1.33x Quadro P6000
NVIDIA GeForce RTX 3090
VRAM
1.33x
Need 1.33x GeForce RTX 3090
NVIDIA A100 SXM
VRAM
0.40x
A100 SXM has 2.50x more

To match 1x Google TPU v5p

Tenstorrent Wormhole
Compute (FP32-eq)
1.26x
Need 1.26x Wormhole
VRAM
3.96x
Need 3.96x Wormhole
Enflame Technology CloudBlazer T20
VRAM
2.97x
Need 2.97x CloudBlazer T20
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
3.53x
Need 3.53x RTX 5000 Ada Generation
VRAM
2.97x
Need 2.97x RTX 5000 Ada Generation
NVIDIA T4G
Compute (FP32-eq)
7.06x
Need 7.06x T4G
VRAM
5.94x
Need 5.94x T4G
NVIDIA Quadro P6000
Compute (FP32-eq)
18.17x
Need 18.17x Quadro P6000
VRAM
3.96x
Need 3.96x Quadro P6000
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
6.46x
Need 6.46x GeForce RTX 3090
VRAM
3.96x
Need 3.96x GeForce RTX 3090
NVIDIA A100 SXM
Compute (FP32-eq)
1.47x
Need 1.47x A100 SXM
VRAM
1.19x
Need 1.19x A100 SXM

To match 1x NVIDIA RTX 5000 Ada Generation

Tenstorrent Wormhole
Compute (FP32-eq)
0.36x
Wormhole is 2.80x faster
VRAM
1.33x
Need 1.33x Wormhole
Enflame Technology CloudBlazer T20
VRAM
1.00x
CloudBlazer T20 has 1.00x more
Google TPU v5p
Compute (FP32-eq)
0.28x
TPU v5p is 3.53x faster
VRAM
0.34x
TPU v5p has 2.97x more
NVIDIA T4G
Compute (FP32-eq)
2.00x
Need 2.00x T4G
FP32 Compute
8.06x
Need 8.06x T4G
VRAM
2.00x
Need 2.00x T4G
Memory Bandwidth
1.80x
Need 1.80x T4G
NVIDIA Quadro P6000
Compute (FP32-eq)
5.15x
Need 5.15x Quadro P6000
FP32 Compute
5.17x
Need 5.17x Quadro P6000
VRAM
1.33x
Need 1.33x Quadro P6000
Memory Bandwidth
1.33x
Need 1.33x Quadro P6000
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
1.83x
Need 1.83x GeForce RTX 3090
FP32 Compute
1.83x
Need 1.83x GeForce RTX 3090
VRAM
1.33x
Need 1.33x GeForce RTX 3090
Memory Bandwidth
0.62x
GeForce RTX 3090 has 1.63x more
NVIDIA A100 SXM
Compute (FP32-eq)
0.42x
A100 SXM is 2.40x faster
FP32 Compute
3.35x
Need 3.35x A100 SXM
VRAM
0.40x
A100 SXM has 2.50x more
Memory Bandwidth
0.28x
A100 SXM has 3.54x more

To match 1x NVIDIA T4G

Tenstorrent Wormhole
Compute (FP32-eq)
0.18x
Wormhole is 5.60x faster
VRAM
0.67x
Wormhole has 1.50x more
Enflame Technology CloudBlazer T20
VRAM
0.50x
CloudBlazer T20 has 2.00x more
Google TPU v5p
Compute (FP32-eq)
0.14x
TPU v5p is 7.06x faster
VRAM
0.17x
TPU v5p has 5.94x more
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
0.50x
RTX 5000 Ada Generation is 2.00x faster
FP32 Compute
0.12x
RTX 5000 Ada Generation is 8.06x faster
VRAM
0.50x
RTX 5000 Ada Generation has 2.00x more
Memory Bandwidth
0.56x
RTX 5000 Ada Generation has 1.80x more
NVIDIA Quadro P6000
Compute (FP32-eq)
2.57x
Need 2.57x Quadro P6000
FP32 Compute
0.64x
Quadro P6000 is 1.56x faster
VRAM
0.67x
Quadro P6000 has 1.50x more
Memory Bandwidth
0.74x
Quadro P6000 has 1.35x more
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
0.92x
GeForce RTX 3090 is 1.09x faster
FP32 Compute
0.23x
GeForce RTX 3090 is 4.40x faster
VRAM
0.67x
GeForce RTX 3090 has 1.50x more
Memory Bandwidth
0.34x
GeForce RTX 3090 has 2.92x more
NVIDIA A100 SXM
Compute (FP32-eq)
0.21x
A100 SXM is 4.80x faster
FP32 Compute
0.42x
A100 SXM is 2.41x faster
VRAM
0.20x
A100 SXM has 5.00x more
Memory Bandwidth
0.16x
A100 SXM has 6.37x more

To match 1x NVIDIA Quadro P6000

Tenstorrent Wormhole
Compute (FP32-eq)
0.07x
Wormhole is 14.41x faster
VRAM
1.00x
Wormhole has 1.00x more
Enflame Technology CloudBlazer T20
VRAM
0.75x
CloudBlazer T20 has 1.33x more
Google TPU v5p
Compute (FP32-eq)
0.06x
TPU v5p is 18.17x faster
VRAM
0.25x
TPU v5p has 3.96x more
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
0.19x
RTX 5000 Ada Generation is 5.15x faster
FP32 Compute
0.19x
RTX 5000 Ada Generation is 5.17x faster
VRAM
0.75x
RTX 5000 Ada Generation has 1.33x more
Memory Bandwidth
0.75x
RTX 5000 Ada Generation has 1.33x more
NVIDIA T4G
Compute (FP32-eq)
0.39x
T4G is 2.57x faster
FP32 Compute
1.56x
Need 1.56x T4G
VRAM
1.50x
Need 1.50x T4G
Memory Bandwidth
1.35x
Need 1.35x T4G
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
0.36x
GeForce RTX 3090 is 2.81x faster
FP32 Compute
0.35x
GeForce RTX 3090 is 2.82x faster
VRAM
1.00x
GeForce RTX 3090 has 1.00x more
Memory Bandwidth
0.46x
GeForce RTX 3090 has 2.17x more
NVIDIA A100 SXM
Compute (FP32-eq)
0.08x
A100 SXM is 12.35x faster
FP32 Compute
0.65x
A100 SXM is 1.54x faster
VRAM
0.30x
A100 SXM has 3.33x more
Memory Bandwidth
0.21x
A100 SXM has 4.72x more

To match 1x NVIDIA GeForce RTX 3090

Tenstorrent Wormhole
Compute (FP32-eq)
0.20x
Wormhole is 5.13x faster
VRAM
1.00x
Wormhole has 1.00x more
Enflame Technology CloudBlazer T20
VRAM
0.75x
CloudBlazer T20 has 1.33x more
Google TPU v5p
Compute (FP32-eq)
0.15x
TPU v5p is 6.46x faster
VRAM
0.25x
TPU v5p has 3.96x more
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
0.55x
RTX 5000 Ada Generation is 1.83x faster
FP32 Compute
0.55x
RTX 5000 Ada Generation is 1.83x faster
VRAM
0.75x
RTX 5000 Ada Generation has 1.33x more
Memory Bandwidth
1.63x
Need 1.63x RTX 5000 Ada Generation
NVIDIA T4G
Compute (FP32-eq)
1.09x
Need 1.09x T4G
FP32 Compute
4.40x
Need 4.40x T4G
VRAM
1.50x
Need 1.50x T4G
Memory Bandwidth
2.92x
Need 2.92x T4G
NVIDIA Quadro P6000
Compute (FP32-eq)
2.81x
Need 2.81x Quadro P6000
FP32 Compute
2.82x
Need 2.82x Quadro P6000
VRAM
1.00x
Quadro P6000 has 1.00x more
Memory Bandwidth
2.17x
Need 2.17x Quadro P6000
NVIDIA A100 SXM
Compute (FP32-eq)
0.23x
A100 SXM is 4.39x faster
FP32 Compute
1.83x
Need 1.83x A100 SXM
VRAM
0.30x
A100 SXM has 3.33x more
Memory Bandwidth
0.46x
A100 SXM has 2.18x more

To match 1x NVIDIA A100 SXM

Tenstorrent Wormhole
Compute (FP32-eq)
0.86x
Wormhole is 1.17x faster
VRAM
3.33x
Need 3.33x Wormhole
Enflame Technology CloudBlazer T20
VRAM
2.50x
Need 2.50x CloudBlazer T20
Google TPU v5p
Compute (FP32-eq)
0.68x
TPU v5p is 1.47x faster
VRAM
0.84x
TPU v5p has 1.19x more
NVIDIA RTX 5000 Ada Generation
Compute (FP32-eq)
2.40x
Need 2.40x RTX 5000 Ada Generation
FP32 Compute
0.30x
RTX 5000 Ada Generation is 3.35x faster
VRAM
2.50x
Need 2.50x RTX 5000 Ada Generation
Memory Bandwidth
3.54x
Need 3.54x RTX 5000 Ada Generation
NVIDIA T4G
Compute (FP32-eq)
4.80x
Need 4.80x T4G
FP32 Compute
2.41x
Need 2.41x T4G
VRAM
5.00x
Need 5.00x T4G
Memory Bandwidth
6.37x
Need 6.37x T4G
NVIDIA Quadro P6000
Compute (FP32-eq)
12.35x
Need 12.35x Quadro P6000
FP32 Compute
1.54x
Need 1.54x Quadro P6000
VRAM
3.33x
Need 3.33x Quadro P6000
Memory Bandwidth
4.72x
Need 4.72x Quadro P6000
NVIDIA GeForce RTX 3090
Compute (FP32-eq)
4.39x
Need 4.39x GeForce RTX 3090
FP32 Compute
0.55x
GeForce RTX 3090 is 1.83x faster
VRAM
3.33x
Need 3.33x GeForce RTX 3090
Memory Bandwidth
2.18x
Need 2.18x GeForce RTX 3090

Pricing

Price TypeWormholeCloudBlazer T20TPU v5pRTX 5000 Ada GenerationT4GQuadro P6000GeForce RTX 3090A100 SXM
CAPEX (Street Price)$15,000
OPEX (per hour)$5.00/hr$0.42/hr$1.10/hr$0.11/hr$4.05/hr
Price per TFLOPs (FP32-eq)$96