Technological advancements, market trends, and evolving consumer demands shape the semiconductor industry’s future. At the intersection of all this is Hanny Lohchab, who, with his decade of experience in the essential industry of semiconductor tech, is making an impact in the Physical Design domain. Here are some key industries that are and will continue to reap the most benefits of his breakthrough work:

Impact on Cloud Computing

Today, all cloud computing happens on robust data centers that support the infrastructure of modern computing and digital services. Data centers utilize a variety of CPU (Central Processing Unit) options based on their specific needs, workload requirements, and budget considerations. Hanny is a pioneer in exploring the ARM (Advanced RISC machines) architecture-based server CPU. During his work as the lead of the memory system design of ThunderX products, he came up with the idea of noise-canceling Bus Routing Topology. This resulted in a double-digit latency improvement of the memory ring, speeding up the overall throughput. While doing the composition of the blocks of the memory subsystem, he came up with the brilliant idea of inverted flops at the ports. 

This approach was never tried across the industry, but being the brain behind the design, he could control the whole inverted logic flow, pass the equivalency, and reduce the gate count demand. This approach helped save leakage power loss, which is a dominant component on server chips. These great ideas and execution got him noticed at the top management level, and he was the youngest to be given the Staff designation and memory system lead title. His crucial work on the ThunderX product line at Marvell helped spearhead the push to ARM servers, which are now becoming the main architecture for cloud computing, and he achieved that three years before the market noticed the advantages of ARM architecture on server chips.

Optimizing Graphics Processing Unit (GPU) 

GPUs are specialized electronic circuits designed to accelerate the creation of images, videos and animations in computers. While at AMD (Advanced Micro Devices), his work directly contributed to bringing out the latest Radeon Gaming cards in the market. He was the key member in the backend design of cache coherency blocks. These blocks are the key components in the overall GPU command processing time. He created a novel approach for memory placement based on data flow in the blocks. This critical idea was replicated in other block logic of graphics and SoC, which improved overall timing significantly, resulting in a frequency uplift. 

This improvement helped in the critical component of shader processing and gave enhanced images for EKG, digital X-ray and other medical imaging needs. Because of his key leadership and critical thinking, he was also tasked with enabling Samsung to adopt AMD RDNA graphic architecture. This made him the single point of contact for data exchange between the teams. He showcased leadership skills by assisting Samsung with early adoption and contributed to the development of initial designs that surpassed the projected benchmarks. Throughout his time at AMD, the company’s stock value quadrupled, leading to a significant increase in AMD’s total market capitalization. This achievement positioned AMD ahead of Intel as a vital provider of semiconductor chips to medical and defense sectors.

Leading on-device AI in the mobile compute ecosystem

The majority of smartphone users worldwide utilize Android devices, with a significant portion powered by Snapdragon chips. Being a lead engineer in the Graphics team of Snapdragon, he was tasked with the execution of the next-gen chipset. These presented challenges in terms of density on wafer and this is where his brilliant idea of a gate selection algorithm with emphasis on a good mix of complex cells and NAND/NOR gates saved the day. This helped in meeting the high density demand and achieving the higher throughput requirements of the advanced tech nodes. 

Because of his curious mind, he let his ideas take a bigger form for the whole GPU hierarchy, and all blocks deployed that approach. He was also tasked with the Look Ahead responsibility of early design analysis for the latest technology and fabrication trends. His data flow analysis and memory arrangement came to use here and helped relieve congestion on a lot of critical blocks. During his tenure Qualcomm has been able to explore the Auto and Compute markets with Hanny tasked with the key role.

The demand for semiconductor chips powering artificial intelligence (AI) and machine learning (ML) applications is growing rapidly. Hanny’s proven track record, makes him a great engineering mind to keep an eye on as an expert in this industry.

Published by: Martin De Juan