OpenPOWER Reboot – New Director, New Silicon Partners, Leveraging Linux Foundation Connections

By John Russell

July 2, 2020

Earlier this week the OpenPOWER Foundation announced the contribution of IBM’s A21 Power processor core design to the open source community. Roughly this time last year, IBM announced open sourcing its Power instruction set (ISA) and Open Coherent Accelerator Processor Interface (OpenCAPI) and Open Memory Interface (OMI). That’s also when IBM said OpenPOWER would become a Linux Foundation entity. Then a few weeks ago, OpenPOWER named a new executive director, James Kulina.

Change is afoot at the OpenPOWER Foundation. Will it be enough to prompt wider (re)consideration and adoption of the OpenPOWER platform and ecosystem?

James Kulina, executive director, OpenPOWER Foundation

The big gun here historically has been IBM, which with Google, Tyan, Nvidia, and Mellanox, founded the OpenPOWER Foundation in 2013. It was meant to challenge Intel’s processor dominance in servers and HPC and to offer an alternative. Not surprisingly, creating a new ecosystem turns out to be hard. For one thing, IBM was the lone Power chip provider sometimes making OpenPOWER seem like just another IBM venture. That’s probably unfair and certainly understates the formidable capabilities of IBM’s Power microprocessor (think Summit supercomputer) and the technical strengths of OpenCAPI and OMI (Open Memory Interface).

One has to admire IBM’s pluck in dumping virtually all of its x86-based systems businesses for a big bet on IBM Power microprocessors and OpenPOWER generally.

That was then. This is now. For many reasons – let’s just lump them together as insufficient market traction – IBM now looks to be deemphasizing its bet on OpenPOWER and passing the torch or at least a big chunk of the torch to others. A strong piece of evidence is that a successor to the Power9 processor (Power10) seems long overdue. Indeed, Kulina has written in a blog introducing himself, “It’s my goal to make OpenPOWER one of the easiest platforms to go from an idea to a silicon chip.”

That means getting more Power chips from more silicon providers as has sometimes been discussed if not aggressively pursued in the past. The sudden emergence of processor diversity (AMD, ARM, RISC-V) and slight loosening of Intel’s grip on the market suggests the window for alternatives is open. Performance and price will be critical.

The question is: Can what often seemed like an IBM-plus-a-few-friends club transform into a thriving ecosystem that captures meaningful pieces of the HPC, mainstream server, cloud, edge, and perhaps other market segments. OpenPOWER currently reports it has 350-plus members, 150 OpenPOWER-ready certified products and 40 OpenPOWER systems shipping or in development. Kulina has high hopes for expansion and recently discussed his plans for growth with HPCwire. Presented here are a few of his comments.

HPCwire: OpenPOWER generally been seen as IBM centric and mostly a way to push its chips and systems. Is that wrong?

James Kulina: I actually would agree with you that historically the connotation has been this is an IBM pet project. If you talk to IBM, they don’t want that.  I think the announcement last year actually sets us up for that – the next chapter where we can actually go and be this fully independent entity, where we can build out an ecosystem, both hardware, silicon hardware, systems and software. Being under the Linux Foundation gives us access to the latest and greatest open source software technologies and knowledge of how to actually build out really sustainable ecosystems.

So that’s kind of where I come from anyway. We want to be viewed as an independent entity. We want people to come to the OpenPOWER Foundation to adopt the technology, the Power ISA, and all the other peripheral technologies that can spin out of that. We don’t want it viewed as, “we’re coming to IBM first and OpenPOWER is just this thing that sits off the side.” That’s not how it’s going to be moving forward.

We’re actually going to be proactive in our approach towards building out an ecosystem. To get to the idea that I framed around silicon, that’s part of it. We need to have more people actually building chips. I think before a lot of companies were just waiting to see what IBM would do with its silicon with the power architecture. But now, including geopolitical [forces] and whatnot, you see a lot of interest in having a fully open source, architecture and ISA to adopt and actually develop [their] own domestic chips. [For example], a lot of companies out of China and other areas are already asking about power, and I’m very interested in it. [That] will lead to really good things because anybody can get involved in taking the Power ISA and customize as long as they stay within compliance.

HPCwire: That seems like a lot to tackle. What makes you think OpenPOWER can pull off a successful restart that avoids past missteps?

James Kulina: I’ve been following OpenPOWER for a while now, actually, since it started. I’m very interested in hardware in general and early in my career I did hardware design. I always wanted to get back into it and have always thought the success of open source software could now lead to [success in] open source hardware. The software folks have [built] roadmaps. It is different because hardware is different. The cost structure is different – how you sell and buy and all that stuff – but there’s a lot of learnings that we can utilize in terms of talking to actual silicon vendors.

I’ve had preliminary talks about who we might want to go after, and what we might want to want to do with them. But it’s really going to be around showing the value first of the architecture showing the use cases that it can benefit from Power. As well as I think software is a very key component in making sure that people don’t have any reservations about adopting a new platform. So our goal, my view is our goal as a foundation is to de-risk as much as possible, to think three to four steps ahead, and to try to take out any of those barriers that might be preventing people from adopting the technology.

HPCwire. What do you see as the advantages the Power ISA set brings to the market. What distinguishes it from existing architectures, particularly given we seem be entering an emerging era of processor diversity with many aspirants?

James Kulina: The first [advantage] is that it is a mature technology. It’s not still being ratified, like RISC-V. Now RISC-V is a great project and we’re happy to be involved with them, but it’s still nascent. The Power architecture is, you know, running the two fastest supercomputers in the world already. It has a proven software ecosystem, although I still think that that can be drastically enhanced in terms of what type of software is running. We need to cozy up more towards the things that a lot of your average developers would be interested in, not just HPC developers. There’s a lot of room for growth there and a lot of interest.

So OpenPOWER is a mature ecosystem. It’s the fact that we have the OpenCAPI initiative. Hopefully that will be merging with the other 410 or however many other [standards efforts] that are basically doing the same thing – CXL (Compute Express Link) and all that. It’s things that we have that are already fairly mature and production grade and enterprise grade, and fully open source. One great thing we have is that you are patent-protected under the structure we have. That makes it a lot easier for people to actually adopt and consume [OpenPOWER] technology and to add extensions.

HPCwire: Can you give us a few examples of specific use case areas that you think Power is well suited for?

James Kulina: The low hanging fruit still is HPC and enterprise but there are others we want to understand better. I’m still coming up to speed on them – where do we actually stand in other use cases and other segments, such as telecommunications, such as networking, such as edge computing, AI, and, and the like. This is where by being part of the Linux Foundation we get an inside track into all of those segments because they’re leaders in driving software that’s running these [segments]. We’re already starting conversations with top leaders and those projects to see how we can create a feedback loop between us so we know what we need to start building into our technology so that their members can actually adopt.

HPCwire: How about nearer term plans for the Power10 processor? We’d been expecting either more Power9 options or the Power10 introduction by now. Arm and AMD have been gaining momentum while the Power chip line seems stuck.

IBM Power9 microprocessor die

James Kulina: I’m not really plugged into what IBM is doing there. You’d have to talk to IBM directly about that. I will say it’s going to be a fantastic chip. But you know, this is why we want multiple vendors for silicon. I think the Googles of the world, the hyper scalars of the world, wanted to see multiple vendors as well. And there’s a lot of people in the middle, right, that want that kind of technology that don’t have the resources and this is where I think we can have a groundswell of support around a fully open source platform.

HPCwire: It sounds like you’re expecting the OpenPOWER platform to expand beyond HPC.

James Kulina: We’re not abandoning HPC at all. There’s a lot of work that we can be doing in HPC. I honestly believe a lot of the technologies happening in the cloud space are going to be pushing into the HPC space. You’re already seeing containers and Kubernetes being adopted in HPC. You know, they’re now adopting things that traditional HPC workload managers have been doing. There’s going to be a merging, I think, of these spaces. That’s actually a good thing because then you don’t have silos and can have workloads being fully portable between HPC environments as well as cloud native environments.

So we want to make sure that we have irons in both fires there. But I definitely think there’s a lot of momentum with RISC-V. There’s a lot of momentum for Arm and AMD and that’s because they have platforms that are readily accessible to the developer community. That’s always been a struggle for the OpenPOWER systems because it’s such an enterprise grade system. It’s such a beast of a system and really can do a lot of great things but we need to have multiple silicon vendors so we can actually gain access to a wider swath of developers.

HPCwire: What sorts milestones are you setting for yourself for the next 6-12-18 months. What are you hoping to get done?

James Kulina: I think in the first couple months we need to think through organizational issues in terms of work groups. How do we architect our work groups? I want them to be more around use cases and segments and not just focused specifically on technologies. It makes it a lot easier for members if they can understand what is what’s in it for them, and where they fit and how they can gain value.

The first thing is just getting organized in a way so that we can truly scale it. After that, it’s more about interfacing with silicon providers, IP houses, as well as fab houses. I think at every layer of the stack in the pipeline we want to see what we can do to make it easier to build new silicon as well as systems and getting the right people in the room to talk things through and have those feedback loops be as efficient as possible. My hope is that within a year or a year and a half, we actually have another silicon provider out there. And, you know, to give us some breathing room in terms of showcasing how an ecosystem can then grow even further from that.

HPCwire: Do you hope to include having silicon targeting broader segments?

James Kulina: That’s my goal. Whether it’s going to happen in a year or two years might be too early.

HPCwire: What do you think are the most pressing challenges?

James Kulina: The first thing is showcasing the use cases and the value proposition and getting organized into a state where we can scale. The next thing is more about access. So getting systems to people and getting to developers in particular, getting integrated with all the key open source groups and making sure that Power architecture is a first class citizen; it’s already a first class citizen in the kind of languages and the kind of software that people are actually caring about. Then it’s getting silicon providers on board and investigating power and then hopefully adopting power and producing new chips.

There’s also an education piece, a curriculum piece that we need to focus on as well. We have a fully open source platform, which is ready and it’s very useful in the academia world as well as in the commercial space, you can have people go from the nuts and the bolts all the way up through the sack and see how everything works. We’re working with a number of universities to figure out how we can actually put together a curriculum. And how can we make that accessible not just for the students there, but also globally. There’s a lot of stuff that we can do in terms of, you know, events and meetups, and hackathons and all that.

HPCwire: Are there any specific synergies between Red Hat (IBM) and OpenPOWER?

James Kulina: Well, they (Red Hat) have an internal multi-architecture group and are already testing across the Power architecture. There’s a lot more we can be doing. It boils down, again, to getting access to Power systems. I think that a lot of Red Hatters don’t have access to it. So this is one of the things now that Red Hat’s part of IBM that will help out.

HPCwire: Could share a little bit more about how you came into the position?

James Kulina: I’ve been following OpenPOWER for quite some time and I saw the announcement last year. In my previous role, I co-founded an open source cloud startup called hyper.Sh and we were acquired last year by Ant Financial, which was Alibaba. I was going through that transition and they eventually asked me to relocate to China. I said no for a number reasons and took a couple months off and saw that OpenPOWER had announced it fully open sourced the Power ISA. I reached out to Hugh Blemings, former OpenPOWER executive director and now an OpenPOWER board advisor), who I’ve known for a while through the open source ecosystem, and said, “Do you need any help because this is, this is awesome; it is actually what needed to happen six years ago.” Hugh said, actually, we do and it kind of kicked off from that.

Brief Bio of James Kulina
James is Executive Director of the OpenPower Foundation, with over 10 years of open source experience across hardware, software, and network engineering disciplines. James brings a passion for open source and is committed to growing OpenPower Foundation’s membership, community, and ecosystem. He is a serial entrepreneur with a background in enterprise technology and has worked in roles spanning operations, business development, product management, and engineering.

Previously, James was co-founder and COO at Hyper.sh, an open source cloud-native virtualization startup acquired by Ant Financial. Prior to that, he led product management in Red Hat’s OpenStack group, and was a product lead on AT&T’s first OpenStack Cloud. James graduated from University of Virginia with a degree in Electrical Engineering and is based in New York.

Link to article on the A21 core design just put into open source: https://www.hpcwire.com/off-the-wire/a2i-power-processor-core-contributed-to-openpower-community-to-advance-open-hardware-collaboration/

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

ARM, Fujitsu Targeting Open-source Software for Power Efficiency in 2-nm Chip

July 19, 2024

Fujitsu and ARM are relying on open-source software to bring power efficiency to an air-cooled supercomputing chip that will ship in 2027. Monaka chip, which will be made using the 2-nanometer process, is based on the Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI projects.). Other tools have joined the CUDA castle siege. AMD Read more…

Quantum Watchers – Terrific Interview with Caltech’s John Preskill by CERN

July 17, 2024

In case you missed it, there's a fascinating interview with John Preskill, the prominent Caltech physicist and pioneering quantum computing researcher that was recently posted by CERN’s department of experimental physi Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to communities across the globe. As climate change is warming ocea Read more…

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical claims. A paper published on July 10 by researchers in the U. Read more…

Belt-Tightening in Store for Most Federal FY25 Science Budets

July 15, 2024

If it’s summer, it’s federal budgeting time, not to mention an election year as well. There’s an excellent summary of the curent state of FY25 efforts reported in AIP’s policy FYI: Science Policy News. Belt-tight Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI pro Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to com Read more…

Shutterstock 1886124835

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical c Read more…

Shutterstock 2203611339

NSF Issues Next Solicitation and More Detail on National Quantum Virtual Laboratory

July 10, 2024

After percolating for roughly a year, NSF has issued the next solicitation for the National Quantum Virtual Lab program — this one focused on design and imple Read more…

NCSA’s SEAS Team Keeps APACE of AlphaFold2

July 9, 2024

High-performance computing (HPC) can often be challenging for researchers to use because it requires expertise in working with large datasets, scaling the softw Read more…

Anders Jensen on Europe’s Plan for AI-optimized Supercomputers, Welcoming the UK, and More

July 8, 2024

The recent ISC24 conference in Hamburg showcased LUMI and other leadership-class supercomputers co-funded by the EuroHPC Joint Undertaking (JU), including three Read more…

Generative AI to Account for 1.5% of World’s Power Consumption by 2029

July 8, 2024

Generative AI will take on a larger chunk of the world's power consumption to keep up with the hefty hardware requirements to run applications. "AI chips repres Read more…

US Senators Propose $32 Billion in Annual AI Spending, but Critics Remain Unconvinced

July 5, 2024

Senate leader, Chuck Schumer, and three colleagues want the US government to spend at least $32 billion annually by 2026 for non-defense related AI systems.  T Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock_1687123447

Nvidia Economics: Make $5-$7 for Every $1 Spent on GPUs

June 30, 2024

Nvidia is saying that companies could make $5 to $7 for every $1 invested in GPUs over a four-year period. Customers are investing billions in new Nvidia hardwa Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Leading Solution Providers

Contributors

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

IonQ Plots Path to Commercial (Quantum) Advantage

July 2, 2024

IonQ, the trapped ion quantum computing specialist, delivered a progress report last week firming up 2024/25 product goals and reviewing its technology roadmap. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire