GSI Technology’s Gemini-I® APU Showcased in “In-Memory Acceleration for Big Data”
July 21 2020 - 10:52AM
GSI Technology, Inc. (Nasdaq: GSIT), a leading
provider of high-performance memory solutions for the networking,
telecommunications and military markets, and developer of the
Gemini® Associative Processing Unit (APU), announces the
publication of the whitepaper “In-Memory Acceleration For Big Data”
by The Linley Group, the industry's leading source for independent
technology analysis of semiconductors.
The Linley Group whitepaper provides an overview of the Gemini
APU’s unique, patented technology, its advantages over other
processors, and provides examples of its performance. The Gemini
APU is based upon a new processor architecture that combines memory
and compute units on a single chip, which is optimized for large
amounts of data. GSI has demonstrated Gemini's performance in
similarity searches run on datasets as large as one billion
records. A unique benefit of Gemini’s design and software
capabilities allows linear scalability to manage 40 billion records
and beyond by combining multiple systems with multiple cards.
Gemini delivers performance orders of magnitude better than a
single processor with cost and power savings relative to the number
of other types of processors required to match its performance.
Lee-Lean Shu, Chairman and Chief Executive Officer of GSI
Technology, commented, "As a leading expert in the semiconductor
industry, the Linley Group's analysis of the Gemini APU expands the
visibility and awareness of our revolutionary technology, and
supports our sales activities. Companies are increasingly analyzing
vast data sets to enhance business outcomes, increase security and
surveillance, and improve R & D returns. Gemini's
outperformance in searching large datasets is optimal for multiple
applications, from image-recognition and drug discovery to
e-commerce, natural-language processing, and visual search. We look
forward to participating in The Linley Group’s Fall Processor
Conference this October."
Linley Gwennap, principal analyst at The Linley Group and
editor-in-chief of Microprocessor Report, commented, "GSI
Technology’s advanced processor architecture allows businesses to
keep pace with the need to analyze large datasets, which are
rapidly growing and seeing broader applications across industries.
By specializing its design and dividing the computation among many
small cores, Gemini easily outperforms other types of
processors."
Gemini APU's patented design includes millions of processors
that can load data directly from the on-chip memory. This unique
design intersperses more than two million small processing units
among 48 million memory cells, dividing this memory into enough
subunits to feed each processor, thereby allowing for much higher
data flow. As a result, Gemini outperforms standard processors by
100x or more on big-data workloads while reducing power by 70%.
The table below, which was excerpted from The Linley Group
whitepaper, shows that the APU’s memory bandwidth exceeds other
processors, allowing it to quickly analyze large datasets, and
lower total power dissipation (TDP), which reduces total cost of
ownership.
|
Gemini APU |
Xeon 8280 |
Nvidia A100 |
Graphcore |
Compute Cores |
2 million x 1 bit |
28 x 2x512 bits |
104 x 4,096 bits |
1,216 x 64 bits |
Compute Speed |
400MHz |
2.7GHz |
1.4GHz |
1.6GHz |
Peak Compute* |
25 TOPS |
10 TOPS |
75 TOPS |
16 TOPS |
On-Chip Memory |
12MB L1 |
38.5MB L3 |
40MB L2 |
300MB L1 |
Mem Bandwidth |
26 TB/s |
1 TB/s |
7 TB/s |
16 TB/s |
Power |
60W TDP |
205W TDP |
400W TDP |
150W TDP |
*Trillions of 8-bit ADD operations per second (TOPS). (Source:
vendors)
The whitepaper, released on July 17, 2020, explains the Gemini
APU technology in detail and how it might apply to various
applications. GSI Technology sponsored the whitepaper, but the
opinions and analysis are those of Linley Gwennap and his team of
analysts.
To view the whitepaper “In-Memory Acceleration for Big Data” by
Linley Gwennap, Principal Analyst at The Linley Group, visit the
GSI Technology website at
https://www.gsitechnology.com/In-Memory-Acceleration-for-Big-Data.
Subscribers to The Linley Group can access the whitepaper on their
website at www.linleygroup.com.
ABOUT GSI TECHNOLOGY
Founded in 1995, GSI Technology, Inc. is a leading provider of
semiconductor memory solutions. GSI’s resources are currently
focused on bringing new products to market that leverage existing
core strengths, including radiation-hardened memory products for
extreme environments, and Gemini, the APU designed to deliver
performance advantages for diverse artificial intelligence
applications. GSI Technology is headquartered in Sunnyvale,
California and has sales offices in the Americas, Europe, and
Asia. For more information, please
visit www.gsitechnology.com.
About the Linley Group
The Linley Group is the industry's leading source for
independent technology analysis of semiconductors for a broad range
of applications including networking, communications, data-center
applications, mobile, and embedded. The company provides strategic
consulting services, in-depth analytical reports, and conferences
focused on advanced technologies for chip and system design. The
Linley Group is the publisher of the noted Microprocessor
Report, a weekly publication. For insights on recent industry news,
subscribe to the company's free email newsletter: Linley
Newsletter.
Contacts:Investor Relations:Hayden IR Kim
Rogers 385-831-7337
Company:GSI Technology, Inc. Douglas M. Schirle
Chief Financial Officer 408-331-9802
GSI Technology (NASDAQ:GSIT)
Historical Stock Chart
From Apr 2024 to May 2024
GSI Technology (NASDAQ:GSIT)
Historical Stock Chart
From May 2023 to May 2024