GUNNIR Arc B580 Photon OC vs NVIDIA B200

Comparison GUNNIR Arc B580 Photon OC with 12 GB GDDR6 and 2,560 cores vs NVIDIA B200 with 90 GB HBM3e and 18,944 cores.

Loading...

Performance Rating

NVIDIA B200 outperforms GUNNIR Arc B580 Photon OC by 642.78% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

GUNNIR Arc B580 Photon OC

8.9

GUNNIR Arc B580 Photon OC

8.9
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA B200

65.8

NVIDIA B200

65.8

Expert Comparison

GUNNIR Arc B580 Photon OC лучше по энергоэффективности и для низкобюджетных задач благодаря более низкому TDP (190 Вт) и высокому GPI (8.9). NVIDIA B200 же с гораздо большей пропускной способностью памяти (4.10 TB/s) и вычислительной мощностью (74.45 TFLOPS) подходит для тяжелых задач, таких как обработка больших объемов данных или машинное обучение.

Contents:

Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks Additional

Memory

Memory Size

12 GB
🔥 +650% 90 ГБ ×2 (180 ГБ)

Memory Type

GDDR6 HBM3e

Memory Bandwidth

456.0 GB/s
🔥 4.10 TB/s ×2 (8.2 TB/s)

Memory Bus Width

192 бит 4,096 бит ×2 (8192 бит)

ML Performance

FP16 (Half Precision)

🔥 +2,350% 29.18 TFLOPS
1.1912 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

14.59 TFLOPS
🔥 +410% 74.45 TFLOPS

FP64 (Double Precision)

0.912 TFLOPS
🔥 +3,981% 37.22 TFLOPS

CUDA Cores

2,560
🔥 +640% 18,944 ×2 (37888)

RT Cores

🔥 20
No

Architecture & Compatibility

GPU Architecture

Xe2-HPG Blackwell

SM (Streaming Multiprocessor)

No
🔥 148

PCIe Version

PCIe 4.0 x8 PCIe 5.0 x16

ML Software Support

CUDA Version

No 10.0

CUDA Toolkit (first supported)

v12

CUDA Toolkit status

Supported Supported

Clocks & Performance

Base Clock

🔥 +281% 2,670
700

Boost Clock

🔥 +45% 2,850
1,965

Memory Clock

🔥 +19% 2,375
2,000

Power Consumption

Recommended PSU

🔥 -68% 450 W
1400 W

Power Connector

2x 8-pin No

TDP/TGP

🔥 -81% 190 W
1000 W

Rendering

Texture Units (TMU)

160
🔥 +270% 592 ×2 (1184)

ROP

🔥 20
No

L2 Cache

18 MB
🔥 +178% 50 MB

Benchmarks

llama.cpp, llama 7B Q4_0

315.9 tokens/s

Additional

Slots

Dual-slot
🔥 SXM Module

Release Date

Dec. 13, 2024 March 13, 2024

Display Outputs

1x HDMI 2.1a
3x DisplayPort 2.1
No outputs

Renting is cheaper than buying