If it wasnt bad enough that Moores Law improvements in the density and cost of transistors is slowing. At the same time, the cost of designing chips and of the factories that are used to etch them is also on the rise. Any savings on any of these fronts will be most welcome to keep IT innovation leaping ahead.
One of the promising frontiers of research right now in chip design is using machine learning techniques to actually help with some of the tasks in the design process. We will be discussing this at our upcoming The Next AI Platform event in San Jose on March 10 with Elias Fallon, engineering director at Cadence Design Systems. (You can see the full agenda and register to attend at this link; we hope to see you there.) The use of machine learning in chip design was also one of the topics that Jeff Dean, a senior fellow in the Research Group at Google who has helped invent many of the hyperscalers key technologies, talked about in his keynote address at this weeks 2020 International Solid State Circuits Conference in San Francisco.
Google, as it turns out, has more than a passing interest in compute engines, being one of the large consumers of CPUs and GPUs in the world and also the designer of TPUs spanning from the edge to the datacenter for doing both machine learning inference and training. So this is not just an academic exercise for the search engine giant and public cloud contender particularly if it intends to keep advancing its TPU roadmap and if it decides, like rival Amazon Web Services, to start designing its own custom Arm server chips or decides to do custom Arm chips for its phones and other consumer devices.
With a certain amount of serendipity, some of the work that Google has been doing to run machine learning models across large numbers of different types of compute engines is feeding back into the work that it is doing to automate some of the placement and routing of IP blocks on an ASIC. (It is wonderful when an idea is fractal like that. . . .)
While the pod of TPUv3 systems that Google showed off back in May 2018 can mesh together 1,024 of the tensor processors (which had twice as many cores and about a 15 percent clock speed boost as far as we can tell) to deliver 106 petaflops of aggregate 16-bit half precision multiplication performance (with 32-bit accumulation) using Googles own and very clever bfloat16 data format. Those TPUv3 chips are all cross-coupled using a 3232 toroidal mesh so they can share data, and each TPUv3 core has its own bank of HBM2 memory. This TPUv3 pod is a huge aggregation of compute, which can do either machine learning training or inference, but it is not necessarily as large as Google needs to build. (We will be talking about Deans comments on the future of AI hardware and models in a separate story.)
Suffice it to say, Google is hedging with hybrid architectures that mix CPUs and GPUs and perhaps someday other accelerators for reinforcement learning workloads, and hence the research that Dean and his peers at Google have been involved in that are also being brought to bear on ASIC design.
One of the trends is that models are getting bigger, explains Dean. So the entire model doesnt necessarily fit on a single chip. If you have essentially large models, then model parallelism dividing the model up across multiple chips is important, and getting good performance by giving it a bunch of compute devices is non-trivial and it is not obvious how to do that effectively.
It is not as simple as taking the Message Passing Interface (MPI) that is used to dispatch work on massively parallel supercomputers and hacking it onto a machine learning framework like TensorFlow because of the heterogeneous nature of AI iron. But that might have been an interesting way to spread machine learning training workloads over a lot of compute elements, and some have done this. Google, like other hyperscalers, tends to build its own frameworks and protocols and datastores, informed by other technologies, of course.
Device placement meaning, putting the right neural network (or portion of the code that embodies it) on the right device at the right time for maximum throughput in the overall application is particularly important as neural network models get bigger than the memory space and the compute oomph of a single CPU, GPU, or TPU. And the problem is getting worse faster than the frameworks and hardware can keep up. Take a look:
The number of parameters just keeps growing and the number of devices being used in parallel also keeps growing. In fact, getting 128 GPUs or 128 TPUv3 processors (which is how you get the 512 cores in the chart above) to work in concert is quite an accomplishment, and is on par with the best that supercomputers could do back in the era before loosely coupled, massively parallel supercomputers using MPI took over and federated NUMA servers with actual shared memory were the norm in HPC more than two decades ago. As more and more devices are going to be lashed together in some fashion to handle these models, Google has been experimenting with using reinforcement learning (RL), a special subset of machine learning, to figure out where to best run neural network models at any given time as model ensembles are running on a collection of CPUs and GPUs. In this case, an initial policy is set for dispatching neural network models for processing, and the results are then fed back into the model for further adaptation, moving it toward more and more efficient running of those models.
In 2017, Google trained an RL model to do this work (you can see the paper here) and here is what the resulting placement looked like for the encoder and decoder, and the RL model to place the work on the two CPUs and four GPUs in the system under test ended up with 19.3 percent lower runtime for the training runs compared to the manually placed neural networks done by a human expert. Dean added that this RL-based placement of neural network work on the compute engines does kind of non-intuitive things to achieve that result, which is what seems to be the case with a lot of machine learning applications that, nonetheless, work as well or better than humans doing the same tasks. The issue is that it cant take a lot of RL compute oomph to place the work on the devices to run the neural networks that are being trained themselves. In 2018, Google did research to show how to scale computational graphs to over 80,000 operations (nodes), and last year, Google created what it calls a generalized device placement scheme for dataflow graphs with over 50,000 operations (nodes).
Then we start to think about using this instead of using it to place software computation on different computational devices, we started to think about it for could we use this to do placement and routing in ASIC chip design because the problems, if you squint at them, sort of look similar, says Dean. Reinforcement learning works really well for hard problems with clear rules like Chess or Go, and essentially we started asking ourselves: Can we get a reinforcement learning model to successfully play the game of ASIC chip layout?
There are a couple of challenges to doing this, according to Dean. For one thing, chess and Go both have a single objective, which is to win the game and not lose the game. (They are two sides of the same coin.) With the placement of IP blocks on an ASIC and the routing between them, there is not a simple win or lose and there are many objectives that you care about, such as area, timing, congestion, design rules, and so on. Even more daunting is the fact that the number of potential states that have to be managed by the neural network model for IP block placement is enormous, as this chart below shows:
Finally, the true reward function that drives the placement of IP blocks, which runs in EDA tools, takes many hours to run.
And so we have an architecture Im not going to get a lot of detail but essentially it tries to take a bunch of things that make up a chip design and then try to place them on the wafer, explains Dean, and he showed off some results of placing IP blocks on a low-powered machine learning accelerator chip (we presume this is the edge TPU that Google has created for its smartphones), with some areas intentionally blurred to keep us from learning the details of that chip. We have had a team of human experts places this IP block and they had a couple of proxy reward functions that are very cheap for us to evaluate; we evaluated them in two seconds instead of hours, which is really important because reinforcement learning is one where you iterate many times. So we have a machine learning-based placement system, and what you can see is that it sort of spreads out the logic a bit more rather than having it in quite such a rectangular area, and that has enabled it to get improvements in both congestion and wire length. And we have got comparable or superhuman results on all the different IP blocks that we have tried so far.
Note: I am not sure we want to call AI algorithms superhuman. At least if you dont want to have it banned.
Anyway, here is how that low-powered machine learning accelerator for the RL network versus people doing the IP block placement:
And here is a table that shows the difference between doing the placing and routing by hand and automating it with machine learning:
And finally, here is how the IP block on the TPU chip was handled by the RL network compared to the humans:
Look at how organic these AI-created IP blocks look compared to the Cartesian ones designed by humans. Fascinating.
Now having done this, Google then asked this question: Can we train a general agent that is quickly effective at placing a new design that it has never seen before? Which is precisely the point when you are making a new chip. So Google tested this generalized model against four different IP blocks from the TPU architecture and then also on the Ariane RISC-V processor architecture. This data pits people working with commercial tools and various levels tuning on the model:
And here is some more data on the placement and routing done on the Ariane RISC-V chips:
You can see that experience on other designs actually improves the results significantly, so essentially in twelve hours you can get the darkest blue bar, Dean says, referring to the first chart above, and then continues with the second chart above. And this graph showing the wireline costs where we see if you train from scratch, it actually takes the system a little while before it sort of makes some breakthrough insight and was able to significantly drop the wiring cost, where the pretrained policy has some general intuitions about chip design from seeing other designs and people that get to that level very quickly.
Just like we do ensembles of simulations to do better weather forecasting, Dean says that this kind of AI-juiced placement and routing of IP block sin chip design could be used to quickly generate many different layouts, with different tradeoffs. And in the event that some feature needs to be added, the AI-juiced chip design game could re-do a layout quickly, not taking months to do it.
And most importantly, this automated design assistance could radically drop the cost of creating new chips. These costs are going up exponentially, and data we have seen (thanks to IT industry luminary and Arista Networks chairman and chief technology officer Andy Bechtolsheim), an advanced chip design using 16 nanometer processes cost an average of $106.3 million, shifting to 10 nanometers pushed that up to $174.4 million, and the move to 7 nanometers costs $297.8 million, with projections for 5 nanometer chips to be on the order of $542.2 million. Nearly half of that cost has been and continues to be for software. So we know where to target some of those costs, and machine learning can help.
The question is will the chip design software makers embed AI and foster an explosion in chip designs that can be truly called Cambrian, and then make it up in volume like the rest of us have to do in our work? It will be interesting to see what happens here, and how research like that being done by Google will help.
Read the rest here:
Google Teaches AI To Play The Game Of Chip Design - The Next Platform
- Are We Overly Infatuated With Deep Learning? - Forbes [Last Updated On: August 18th, 2024] [Originally Added On: December 28th, 2019]
- CMSWire's Top 10 AI and Machine Learning Articles of 2019 - CMSWire [Last Updated On: August 18th, 2024] [Originally Added On: December 28th, 2019]
- Can machine learning take over the role of investors? - TechHQ [Last Updated On: August 18th, 2024] [Originally Added On: December 28th, 2019]
- Pear Therapeutics Expands Pipeline with Machine Learning, Digital Therapeutic and Digital Biomarker Technologies - Business Wire [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Dell's Latitude 9510 shakes up corporate laptops with 5G, machine learning, and thin bezels - PCWorld [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Limits of machine learning - Deccan Herald [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Forget Machine Learning, Constraint Solvers are What the Enterprise Needs - - RTInsights [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Tiny Machine Learning On The Attiny85 - Hackaday [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Finally, a good use for AI: Machine-learning tool guesstimates how well your code will run on a CPU core - The Register [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- How Will Your Hotel Property Use Machine Learning in 2020 and Beyond? | - Hotel Technology News [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Technology Trends to Keep an Eye on in 2020 - Built In Chicago [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- AI and machine learning trends to look toward in 2020 - Healthcare IT News [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- The 4 Hottest Trends in Data Science for 2020 - Machine Learning Times - machine learning & data science news - The Predictive Analytics Times [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- The Problem with Hiring Algorithms - Machine Learning Times - machine learning & data science news - The Predictive Analytics Times [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Going Beyond Machine Learning To Machine Reasoning - Forbes [Last Updated On: August 18th, 2024] [Originally Added On: January 11th, 2020]
- Doctor's Hospital focused on incorporation of AI and machine learning - EyeWitness News [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Being human in the age of Artificial Intelligence - Deccan Herald [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Raleys Drive To Be Different Gets an Assist From Machine Learning - Winsight Grocery Business [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Break into the field of AI and Machine Learning with the help of this training - Boing Boing [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- BlackBerry combines AI and machine learning to create connected fleet security solution - Fleet Owner [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- What is the role of machine learning in industry? - Engineer Live [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Seton Hall Announces New Courses in Text Mining and Machine Learning - Seton Hall University News & Events [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Christiana Care offers tips to 'personalize the black box' of machine learning - Healthcare IT News [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Leveraging AI and Machine Learning to Advance Interoperability in Healthcare - - HIT Consultant [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Essential AI & Machine Learning Certification Training Bundle Is Available For A Limited Time 93% Discount Offer Avail Now - Wccftech [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Educate Yourself on Machine Learning at this Las Vegas Event - Small Business Trends [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- 2020: The year of seeing clearly on AI and machine learning - ZDNet [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- How machine learning and automation can modernize the network edge - SiliconANGLE [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Five Reasons to Go to Machine Learning Week 2020 - Machine Learning Times - machine learning & data science news - The Predictive Analytics Times [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Don't want a robot stealing your job? Take a course on AI and machine learning. - Mashable [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Adventures With Artificial Intelligence and Machine Learning - Toolbox [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Optimising Utilisation Forecasting with AI and Machine Learning - Gigabit Magazine - Technology News, Magazine and Website [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Machine Learning: Higher Performance Analytics for Lower ... [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Machine Learning Definition [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Machine Learning Market Size Worth $96.7 Billion by 2025 ... [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Difference between AI, Machine Learning and Deep Learning [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Machine Learning in Human Resources Applications and ... [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Pricing - Machine Learning | Microsoft Azure [Last Updated On: August 18th, 2024] [Originally Added On: January 19th, 2020]
- Looking at the most significant benefits of machine learning for software testing - The Burn-In [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- New York Institute of Finance and Google Cloud Launch A Machine Learning for Trading Specialization on Coursera - PR Web [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- Uncover the Possibilities of AI and Machine Learning With This Bundle - Interesting Engineering [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- Red Hat Survey Shows Hybrid Cloud, AI and Machine Learning are the Focus of Enterprises - Computer Business Review [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- Machine learning - Wikipedia [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- Vectorspace AI Datasets are Now Available to Power Machine Learning (ML) and Artificial Intelligence (AI) Systems in Collaboration with Elastic -... [Last Updated On: August 18th, 2024] [Originally Added On: January 22nd, 2020]
- Learning that Targets Millennial and Generation Z - HR Exchange Network [Last Updated On: August 18th, 2024] [Originally Added On: January 23rd, 2020]
- Machine learning and eco-consciousness key business trends in 2020 - Finfeed [Last Updated On: August 18th, 2024] [Originally Added On: January 24th, 2020]
- Jenkins Creator Launches Startup To Speed Software Testing with Machine Learning -- ADTmag - ADT Magazine [Last Updated On: August 18th, 2024] [Originally Added On: January 24th, 2020]
- Research report investigates the Global Machine Learning In Finance Market 2019-2025 - WhaTech Technology and Markets News [Last Updated On: August 18th, 2024] [Originally Added On: January 25th, 2020]
- Expert: Don't overlook security in rush to adopt AI - The Winchester Star [Last Updated On: August 18th, 2024] [Originally Added On: January 25th, 2020]
- Federated machine learning is coming - here's the questions we should be asking - Diginomica [Last Updated On: August 18th, 2024] [Originally Added On: January 25th, 2020]
- I Know Some Algorithms Are Biased--because I Created One - Scientific American [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Iguazio Deployed by Payoneer to Prevent Fraud with Real-time Machine Learning - Business Wire [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Want To Be AI-First? You Need To Be Data-First. - Forbes [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- How Machine Learning Will Lead to Better Maps - Popular Mechanics [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Technologies of the future, but where are AI and ML headed to? - YourStory [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- In Coronavirus Response, AI is Becoming a Useful Tool in a Global Outbreak - Machine Learning Times - machine learning & data science news - The... [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- This tech firm used AI & machine learning to predict Coronavirus outbreak; warned people about danger zones - Economic Times [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- 3 books to get started on data science and machine learning - TechTalks [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- JP Morgan expands dive into machine learning with new London research centre - The TRADE News [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Euro machine learning startup plans NYC rental platform, the punch list goes digital & other proptech news - The Real Deal [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- The ML Times Is Growing A Letter from the New Editor in Chief - Machine Learning Times - machine learning & data science news - The Predictive... [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Top Machine Learning Services in the Cloud - Datamation [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Combating the coronavirus with Twitter, data mining, and machine learning - TechRepublic [Last Updated On: August 18th, 2024] [Originally Added On: February 1st, 2020]
- Itiviti Partners With AI Innovator Imandra to Integrate Machine Learning Into Client Onboarding and Testing Tools - PRNewswire [Last Updated On: August 18th, 2024] [Originally Added On: February 2nd, 2020]
- Iguazio Deployed by Payoneer to Prevent Fraud with Real-time Machine Learning - Yahoo Finance [Last Updated On: August 18th, 2024] [Originally Added On: February 2nd, 2020]
- ScoreSense Leverages Machine Learning to Take Its Customer Experience to the Next Level - Yahoo Finance [Last Updated On: August 18th, 2024] [Originally Added On: February 2nd, 2020]
- How Machine Learning Is Changing The Future Of Fiber Optics - DesignNews [Last Updated On: August 18th, 2024] [Originally Added On: February 2nd, 2020]
- How to handle the unexpected in conversational AI - ITProPortal [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- SwRI, SMU fund SPARKS program to explore collaborative research and apply machine learning to industry problems - TechStartups.com [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- Reinforcement Learning (RL) Market Report & Framework, 2020: An Introduction to the Technology - Yahoo Finance [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- ValleyML Is Launching a Series of 3 Unique AI Expo Events Focused on Hardware, Enterprise and Robotics in Silicon Valley - AiThority [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- REPLY: European Central Bank Explores the Possibilities of Machine Learning With a Coding Marathon Organised by Reply - Business Wire [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- VUniverse Named One of Five Finalists for SXSW Innovation Awards: AI & Machine Learning Category - PRNewswire [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- AI, machine learning, robots, and marketing tech coming to a store near you - TechRepublic [Last Updated On: August 18th, 2024] [Originally Added On: February 5th, 2020]
- Putting the Humanity Back Into Technology: 10 Skills to Future Proof Your Career - HR Technologist [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]
- Twitter says AI tweet recommendations helped it add millions of users - The Verge [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]
- Artnome Wants to Predict the Price of a Masterpiece. The Problem? There's Only One. - Built In [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]
- Machine Learning Patentability in 2019: 5 Cases Analyzed and Lessons Learned Part 1 - Lexology [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]
- The 17 Best AI and Machine Learning TED Talks for Practitioners - Solutions Review [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]
- Overview of causal inference in machine learning - Ericsson [Last Updated On: August 18th, 2024] [Originally Added On: February 6th, 2020]