Silicon IP

מנכ"לית NetSpeed תפתח את יום הטכנולוגיה של החברה בהרצליה

NetSpeed8060

סונדרי, מנכ"לית החברה, היא דמות מוכרת בתעשית הפאבלס העולמית לא רק בגלל המיזמים השונים שהיא הקימה, אלא גם בשל תמיכתה האיתנה במהנדסים צעירים ובנשים בתעשיית המוליכים למחצה.

 

 מנכ"לית חברת NetSpeed, סונדרי מיטרה תגיע לארץ ב-24 לספטמבר כדי להעביר את הרצאת הפתיחה של כנס הטכנולוגיה של NetSpeed : "צעד באומץ למקום בו שום חיבוריות לא היתה לפניך!" אשר יערך ביום שני 25 לספטמבר במלון הרודס, במרינת הרצליה.

יחד עם סונדרי יגיעו אנוש מוהנדאס, סמנכ"ל שיווק ופיתוח עסקי, ועוד חברי סגל בכירים של החברה.

סונדרי היא דמות מוכרת בתעשית הפאבלס העולמית לא רק בגלל המיזמים השונים שהיא הקימה, אלא גם בשל תמיכתה האיתנה במהנדסים צעירים ובנשים בתעשיית המוליכים למחצה.

netspeed ceo

ערכנו ראיון עם סונדרי רגע לפני הגעתה לישראל, בו היא מדברת על תשוקתה ל-NetSpeed, יעוד החברה בעולם הפאבלס העכשווי, ציפיותיה מביקורה השלישי בארץ ההולך וקרב, ועוד.

למידע נוסף ולהרשמה ליום הטכנוליגי של נטספיד בהרצליה לחצו כאן
(יום שני 25 לספטמבר, 2017, מלון הרודס, מרינה הרצליה)

(הראיון עם סונדרי נשמר באנגלית כדי לשמר את תוכנו היחודי)

 

[SemIsrael] What does NetSpeed do?
[S.M.] NetSpeed System is the IP provider of choice for architects who are doing any SoC development. NetSpeed’s solution is an on-chip interconnect that handles heterogeneous architectures seamlessly. We provide a system interface that can bridge coherent and non-coherent traffic on the same fabric with best in class quality of service features, safety features, security features, debuggability of silicon. We are a fully verified IP and we provide all the collateral necessary for customers to integrate our IP into their SoC. We enable OEMs to bring IP together on an SoC in the most efficient manner and with the best performance possible. If you want to make SoC interconnect a problem of the past, you’ve got to work with a company like NetSpeed.
The vision behind NetSpeed is to transform SoC design by introducing a radically new way of designing interconnects. To accomplish this, we’ve been developing machine learning technology over the past few years to identify the ideal topology needed while solving complex SoC issues like QoS and deadlock avoidance. One of the key advantages we’ve discovered is that with machine learning it’s possible to model the system as a whole, taking into account system interactions and understanding how they affect quality-of-service. It also becomes possible to optimize performance and power efficiency broadly across all use cases. In contrast, the conventional approach tends to take a divide and conquer strategy that optimizes individual subsystems in isolation, which tends to create bottlenecks and overdesigned systems to handle worst case conditions.

[SemIsrael] I hear you are planning a trip to Israel; what brings you here?
[S.M.] As you know, Mauro Diamant with IPro represents NetSpeed in Israel. In fact, we were the first IP company to sign with IPro and it has been a fantastic relationship. Mauro is so knowledgeable and well respected and customers really like him; we are really lucky to have him as part of our team. Coming up on the 25th of September, IPro has organized a NetSpeed Technology Day where we will discuss an innovative approach to designing complex chips. We expect 20 to 30 tech companies will join us at the Herods Hotel in Herzelia.
This is my third visit to Israel and I expect to be coming more often. We have both US-based customers with offices in Israel and also Israeli-based customers. The tech industry in Israel is so vibrant and there are so many new and very advanced chip designs; we are excited to be a part of it.

[SemIsrael] Who are some of your customers and why do they choose NetSpeed’s products?
[S.M.] We work with leaders in the automotive, augmented reality, hyperscale SoCs —companies like Mobileye and Intel and others— as well as growing segments in mobile, cloud, networking, storage, digital home and Industrial IoT. Many of our customers are established industry leaders with large engineering teams and who may already have several in-house-developed interconnect solutions that they have finetuned to maximize performance for their application. At the other end of the spectrum, we have new startups who are not attached to doing it themselves and don’t necessarily have the staff or the know-how to do so. In the case of the big players, the pain point is often how to achieve sustainable high performance for their next generation applications in a reasonable footprint and power budget. Once they actually start using our product they become sold on the built-in modeling capability that gives them so much more insight and control over their designs. It makes it easy for them to make changes during development and enables them to reuse their platforms. Startups are not invested in doing things a certain way because they are starting from scratch. For them, using our product is a huge advantage because it kind of jump-starts the platform giving them a proven foundation to build on. They can get up and running fast and with a relatively small team. In all cases, we significantly reduce risks associated with timing closure, deadlock problems, and end-to-end quality of service.

[SemIsrael] What is special about NetSpeed’s solution?
[S.M.] As part of our on-chip interconnect product, we have a whole flow that takes the user through an architecture exploration process where they can actually do all the architectural analysis, run through what-if scenarios. Our platform helps the user explore the architectural tradeoffs and then tailors the interconnect to meet the specific requirements of an individual SoC. We use a highly automated approach that enables us to get the best results in terms of cost, area and power while meeting the bandwidth and latency requirements of the systems. We do this for coherent, streaming, and a wide variety of protocols that exist in the industry today. This is unique because, typically, the architecture effort involves a department of people, with a bunch of engineers who do the modeling. The problem is that whenever any change is required it has to go through a process that could take from hours to days to weeks. But with the NetSpeed platform, an architect can get real-time data for whatever scenario they want to run through without having to rely on a large support team. This is quite powerful and the uniqueness of this platform is the design itself is derived from what the architect has shaped during the exploration process. The nice thing is if your design has to change halfway through because of market requirements, you can dynamically change things. The industry has been trying to do this and you hear about writing System-C models and deriving designs from them that go directly on the chip. NetSpeed’s solution is almost at the point today; we provide a synthesis engine that creates an optimal interconnect based on high level architectural requirements.

[SemIsrael] So, how does machine learning fits into the solution?
[S.M.] The interconnect is controlled by many parameters. If you look at the number of knobs that we need to turn to build an interconnect, it’s probably 2^100 different knobs that we have to manipulate, maneuver and tune to get to the optimal solution. That’s a lot of knobs. At a very high level, to define the space it’s like you have two ports that are connected to each other. You have to find out which is the driver, which is the receiver. Then you have to find out how many bits are on these ports, and you’ve got to see physically whether it is connected this way or that way. Are there congestion points anywhere? What is the optimal way of connecting them? You have to know the bandwidth between these two points. What do you need to service? What is the latency requirement between these two points, and what kind of traffic is originating from these? Is it possible that one of them is dependent on traffic coming from somewhere else? If it is dependent on something else it cannot service that request until the other request is satisfied, so there is a dependency that you need to take care of. If you don’t resolve the dependencies correctly, can this link deadlock? If you haven’t sized it correctly, can it lead to congestion? And what if you’ve worked out all these things, and then you have figured out the route table for it? You know it’s going to go this way, but then when you try to do that there is not enough physical space to fit these two. And this is only two points. Multiply this with N number of IPs with Y number of ports, and now virtualized and real channels. There are so many different knobs that you have to tune/play with to analyze and figure out what that fabric is that you have to use some way of doing it.

[SemIsrael] Is there anything you’d like to add?
[S.M.] Our goal is to give SoC architects a method in which they can do bigger and better things, rather than number crunching and spreadsheets. We are giving them a mathematical model that lets them now innovate beyond that. When you are not constrained with those clusters, horizons expand by taking it outside the chip. Why do I have to have a processor and an FPGA that is providing acceleration, but they still have different memory domains? Why not unify that if I can get them thinking alike? We want to help our customers in Israel to boldly go to new frontiers of technology freed from the limitations of outdated methodology.

הרשמו לכנס הטכנולוגי של NetSpeed כאן
(יום שני 25 לספטמבר, 2017, מלון הרודס, מרינה הרצליה)

מערכת סמישראל