Skip to main content

Intel Arc B570 Graphics Card Coming Soon

·332 words·2 mins
Arc B570 GPU

Since its release, the Intel Arc B580 has been well-received, and it was once hard to find on many retail platforms. Recently, the more affordable Intel Arc B570 has also made an appearance in the Geekbench OpenCL database. As a lower-priced version of the Arc B580, the Arc B570 is equipped with 10 GB of GDDR6 memory. Its performance is about 12% lower than the Arc B580, and it’s also priced about 12% cheaper.

Arc B580

Data shows that the Arc B570 scored 86,716 in the OpenCL test, while the Arc B580 typically scores between 95,000 and 100,000.

In terms of hardware specifications, the Arc B570 features 144 compute units and a boost clock speed of up to 2.75 GHz. While official documentation lists a base frequency of 2500 MHz, the 2.75 GHz shown in this test might be a characteristic of a custom AIB (Add-in Board) version. The Arc B570 uses 10 GB of GDDR6 memory and a 160-bit memory bus, achieving a memory bandwidth of 380 GB/s. This is a reduction in specifications compared to the Arc B580’s 12 GB of GDDR6 and 192-bit memory bus. Both GPUs use the BMG-G21 chip, manufactured with TSMC’s 5nm process. The chip has a die area of 272 mm² and contains 19.6 million transistors.

Arc B580

Regarding power consumption, the Arc B570 has a maximum power draw of 150 watts, compared to the Arc B580’s 190 watts. The Arc B570 is priced at $219, with limited edition and custom-designed versions from Intel’s board partners potentially costing more. Intel hopes to meet the needs of different users with the Arc B570 and plans to release more models in the future to continue supporting the growth of the discrete graphics card market.

The launch of the Intel Arc B570 “Battlemage” GPU provides consumers with more choices, especially for users on a budget who still require reliable performance. As the official release date approaches, the market will continue to pay close attention to the Arc B570’s performance and user feedback.

Related

Types of NVIDIA GPUs and Their Applications in Large-Scale Model Training and Inference
·899 words·5 mins
NVIDIA GPU LLM AI Training Inference
美国将限制中国经第三国购买GPU AI芯片
·21 words·1 min
GPU AI
什么是CPU、GPU、TPU
·23 words·1 min
CPU GPU TPU