top of page

Baum’s Power Analysis Suite is Adopted by SiliconArts



Baum Used to Optimize Power in a ray-tracing GPU.

Baum, electronic design automation (EDA) company, today announced that SiliconArts, a fabless semiconductor company focusing on ray-tracing GPU architecture, used Baum’s power analysis suite to lower both peak power and average power consumption of their GPU, RAIV, and plan to continue applying it in the design of next-generation chips.

SiliconArts’s RAIV is a General-Purpose GPU (GPGPU) that accelerates data processing and plays a key role in the development of major industries in the fourth industrial revolution, such as autonomous driving, IoT, VR/AR, and data centers.

SiliconArts’ own graphics technology, the RayCore MC-Series, enables a scalable 3D GPU rendering solution from 1 Grays/sec to up to 10Grays/sec for multi-core solutions that can be integrated into thin GPU platform to provide futuristic capabilities for next-generation visualization. Higher performance rendering platform for dedicated and professional use cases can be scaled to 100’s fo Grays/sec performance with multi-chip board-level design. SiliconArts CEO Hyung-min Yoon says “ The ray tracing movement is so critical to our computing platform’s user interface. We can expect ray tracing to be a core function of all GPUs in the future.”

“Baum power model enables fast and accurate transient power analysis for long usage scenarios, which is very difficult with other power simulation solutions,” said Mr. Yoon. “Transient analysis is a key in detecting the scenarios causing abrupt power changes and in extracting detailed power breakdown which our designers heavily rely on for power optimization.

“Baum is delighted to provide SiliconArts with the power analysis tools they need to optimize the power and thermal efficiency of their current and future RAIV product lines,” remarked Youngsoo Shin, co-CEO of Baum. “PowerBaum’s very high-speed and assurance of implementation accuracy combine for an ideal power analysis solution across SiliconArts’s entire development phase. Large and complex chip designers can benefit even more by using Baum's technology, especially through operational clock gating ratio (OCGR) results, which analyze the clock gating efficiency of RTL input in the early design stages."

Baum will showcase its technology at Design Automation Conferences, San Francisco, CA, USA, on July 10-14.

The Latest Version of PowerBaum

PowerBaum automatically generates high-level power models from design sources and applies advanced learning techniques of gate-level behavior to achieve very high accuracy. Baum power models run in higher abstraction environments, such as RTL simulation, ESL (virtual prototypes), and hardware emulation to achieve orders of magnitude performance improvement compared to competing solutions in the market. Baum’s automated power analysis and power modeling solutions support both dynamic and leakage power, taking in register transfer level (RTL) and netlist descriptions of the design.

PowerBaum 2022.06 is shipping today and available globally. Pricing is available upon request.

About Baum

Baum provides electronic design automation (EDA) solutions for very fast yet accurate power analysis through power modeling technology. Founded in 2016 by seasoned semiconductor professionals with technical, R&D, and business development expertise, Baum is privately held and funded.

About SiliconArts

Siliconarts, a leading provider of the world’s first real-time ray/path tracing GPU (Graphics Processing Unit) technology, has expanded its product line to AI GPU with a product titled 'RAIV'. It is a General-Purpose GPU (GPGPU) that quickly calculates and processes data used in deep learning or machine learning. It can play a key role in the development of major industries in the Industrial Revolution such as autonomous driving, IoT, games(VR/AR), and data centers. With its ray-tracing GPU technology and RAIV, Siliconarts continues to grow as a next-generation innovative hardware solution development company.

bottom of page