Arteris Articles

SemiWiki: ML and Memories: A Complex Relationship

Kurt Shuler, VP Marketing at Arteris IP, helped Bernard Murphy (SemiWiki) learn the multiple ways that different types of memory need to connect to these accelerators in the latest SemiWiki blog:

ML and Memories: A Complex Relationship

March 13th, 2019 - By Bernard Murphy

How do AI architectures connect with memories? The answer is more complex than in conventional SoC architectures.

No, I’m not going to talk about in in-memory-compute architectures. There’s interesting work being done there but here I’m going to talk here about mainstream architectures for memory support in Machine Learning (ML) designs. These are still based on conventional memory components/IP such as cache, register files, SRAM and various flavors of off-chip memory, including not yet “conventional” high-bandwidth memory (HBM). However, the way these memories are organized, connected and located can vary quite significantly between ML applications.

For more information, please visit the Arteris IP AI package webpage: http://www.arteris.com/flexnoc-ai-package

Topics: semiconductor artificial intelligence semiwiki kurt shuler flexnoc ai package noc interconnect cache coherence

SemiWiki: Segmenting the Machine-Learning Hardware Market

Kurt Shuler, VP Marketing at Arteris IP, shares his 91 entries into finding every company and product that is active in the AI hardware space in this latest SemiWiki blog:

Segmenting the Machine-Learning Hardware Market

March 13th, 2019 - By Bernard Murphy

Machine learning is everywhere, but it can be difficult at times to understand what that really means. Bernard Murphy (SemiWiki) talked to Kurt Shuler and dug through a very detailed spreadsheet Kurt developed to understand better better what is being used where in the ML market.

One of the great pleasures in what I do is to work with people who are working with people in some of the hottest design areas today. A second-level indirect to be sure but that gives me the luxury of taking a broad view. A recent discussion I had with Kurt Shuler (VP Marketing at Arteris IP) is in this class. As a conscientious marketing guy, he wants to understand the available market in AI hardware because they have quite a bit of activity in that space – more on that later. So Kurt put a lot of work into finding every company and product he could that is active in this space, 91 entries in his spreadsheet. This he broke down by company, territory (eg China or US), product, target market (eg vision or speech), implementation (eg FPGA or ASIC), whether the product is used in datacenters or at the edge and whether it is being used for training or inference. 


 For more information, download this FlexNoC AI Package datasheet; http://www.arteris.com/flexnoc-ai-package

Topics: FPGA semiconductor edge computing semiwiki inference kurt shuler flexnoc ai package AI training noc interconnect

SemiWiki: Safety: Big Opportunity, A Long and Hard Road

Kurt Shuler, VP Marketing at Arteris IP, explains the support and business cycle from the vendor to the integrator in the latest SemiWiki blog written by Bernard Murphy:

Safety: Big Opportunity, A Long and Hard Road

February 27th, 2019 - By Bernard Murphy

Still think you want to sell IP into the automotive chain? Bernard Murphy (SemiWiki) distills Kurt Shuler insights into what this takes. There’s certainly a lot of promise. More big chips in the central brain and in intelligent sensors together offer a lot of opportunity. The US, Europe and Israel markets are all very aggressive in developing ADAS and ML. China has been a laggard but is coming on strong and is not held back by legacy so much. They also see a big tie-in with AI where they are very strong. Kurt says there are more than a couple of hundred funded startups in automotive and AI in China.

That said, this is not an easy way to get rich. You’ll have to put a lot of investment into supporting your customers, supporting their customers and so on up to the top. The market is very dynamic, so what “done” means may not always be clear. You may not be paid for quite a long time. But if you have the grit to hang on and keep your customer happy the whole way through, you might just be successful!


 For more information, download this FlexNoC AI Package datasheet; http://www.arteris.com/flexnoc-ai-package

Topics: ISO 26262 semiconductor semiwiki kurt shuler flexnoc ai package noc interconnect ML-centric design

Semiconductor Engineering: The Race To Multi-Domain SoCs

 Arteris IP's CEO looks at how automotive and AI are Altering chip design in this article in Semiconductor Engineering;

The Race To Multi-Domain SoCs

February 7th,  2019 - By Ed Sperling

K. Charles Janac, president and CEO of Arteris IP, sat
down with Semiconductor Engineering to discuss the impact of automotive and AI on chip design. What follows are excerpts of that conversation.

SE: What do you see as the biggest changes over the next 12 to 24 months?
Janac: There are segments of the semiconductor market that are shrinking, such as DTV and simple IoT. Others are going through an investment phase, including automotive, AI/machine learning and China. You really want to be focused on those segments. 

SE: So does IP that’s being developed today look radically different than it did five years ago?
Janac:
Yes, everything is getting amazingly complex. What people are building right now are multi-domain SoCs. The CPU, which used to do all the work, does relatively less work. There are accelerators for vision and data analysis outside of the CPU subsystem. There are machine learning sections, some general-purpose, some very specific, all on-chip. There is a memory subsystem with very high-bandwidth memory and low latency. There also is functional safety. You need tremendous performance because a car is a supercomputer on wheels. The car has to be very efficient, because you need to deliver that compute power without water cooling. Power management becomes very sophisticated. And then there are functional safety and security subsystems to keep these safe from environmental and man-made issues.

SE: Where does the network on chip (NoC) fit into all of this?
Janac: All data goes through the NoC of the chip. There are opportunities for generating value from that. But the increase in complexity is increasing the number and sophistication of the interconnect parts of the chip. Before, you may have had networks on chip. Now you may have 20 or 30.

Topics: semiconductor automotive ADAS neural networks AI LIDAR flexnoc ai package noc interconnect ML AI SoC Designers chiplets