JANUARY TALK: ANALYZING MILLIONS OF LINES OF LOGFILES WITH CHEW_LOGS
There is a lot of glorious high-tech science in modernsilicon IC design, but someone still needs to take out the trash. Yes, chip design is hardware engineering, but with billions of gates on a chip, software is involved at every step. Mundane tasks for the engineer include analyzing log files from EDA (Electronic Design Automation) tools, which can spew out millions of lines with each run. How do you know which errors and warnings really matter? And you can’t just ignore them – an undetected error might mean a respin of masks, costing millions of dollars and inestimable loss due to schedule slip. This talk describes how Marvell deals with large amounts of EDA tool logfile reports on a daily basis, using data analysis techniques in Chew_Logs (Check Errors & Warnings in Logfiles).
Date and Time
Location
Hosts
Registration
- Date: 24 Jan 2023
- Time: 12:30 AM UTC to 01:45 AM UTC
-
Add Event to Calendar
- 300 3rd Ave SW
- Rochester, Minnesota
- United States 55902
- Building: Medical Sciences Building
- Room Number: Mann Hall
Speakers
Thomas Arneberg of Marvell Technology, Inc.
Biography:
A dinosaur of the chip design industry, Tom Arneberg earned his Electrical Engineering degree from the University of Minnesota way back in 1982. (His first chip design at Sperry Univac was pretty much a one-man show with lots of graph paper and a new program called SPICE.) But it was his post-graduate classes in Computer Science that have helped him the most during the bulk ofhis 40-year career in integrated circuits.
Tom’s chip design career took him from Sperry Univac in Minnesota to Honeywell in Colorado, a small startup in Oregon called BIT (Bipolar Integrated Technology), then 21 years in Chippewa Falls, Wisconsin, at Cray Research and its descendants, before Intel bought Cray’s chip design group in 2012. He joined the GlobalFoundriesASIC division in Rochester in 2017, which got bought by Marvell in 2019.
Tom has served his sentence in management, but mostly enjoys being an “individual contributor” solving problems as they come up for anyone who wants help, especially by applying computer science concepts to large amounts of data. He has two patents in chip design, and presented a poster for Chew_Logs at DAC 2022, the Design Automation Conference in San Francisco.