DAOS Performance Expands Beyond Intel Optane and Into the Google Cloud

By Rob Farber

October 17, 2022

Distributed Asynchronous Object Storage (DAOS) continues to define high storage performance while transitioning into the cloud and beyond current Intel Optane storage technology devices. Google’s work on DAOS in the cloud, for example, helps users realize the benefits of HPC in the cloud. This includes the flexibility to create and tear down entire HPC clusters within minutes. With the Google HPC-toolkit, of which DAOS is a part, users can transition their on-premises HPC workloads to the Google Cloud. As HPC users know quite well, tying together numerous computational devices drives the need for high-bandwidth, low-latency, and high I/O operations per second HPC storage — else storage becomes the bottleneck. With the Google announcement, capabilities for DAOS are now possible in the Cloud HPC Toolkit for a fully automated user experience.

Andrey Kudryavtsev (DAOS product manager at Intel) recalls, “Google and Intel have been collaborating for almost two years to provide an easy path for HPC workloads to migrate to the cloud. CSPs appreciate the DAOS storage capabilities because it expands their service offerings, delivers high performance on Ethernet communications fabrics and, because it runs in user space, can be easily containerized. DAOS naturally supports the cloud instance use case of spin up the storage server, perform fast IO, then migrate the data to a less expensive storage tier — a use case that is also becoming more common in on-premises datacenters. ”

CSPs appreciate the DAOS storage capabilities because it expands their service offerings, delivers high performance on Ethernet communications fabrics (which are commonly used by ISPs) and, because it runs in user space, can be easily containerized. DAOS naturally supports the cloud instance use case of spin up the storage server, perform fast IO, then migrate the data to a less expensive storage tier — a use case that is also becoming more common in on-premises datacenters. — Andrey Kudryavtsev

Of course, Intel discontinuing Intel Optane Persistent Memory (Intel Optane Pmem) development in 2025 has left many wondering how this affects the overall DAOS story and performance capability. DAOS running on Intel Optane Pmem devices has received much press including how this combined software and memory-bus-based hardware solution set new world records in storage performance.

The Intel team’s planning for life after Intel Optane Pmem included how to meet a significant customer demand to make DAOS cloud friendly and leverage the forthcoming generations of new, multi-vendor storage technologies. The Intel team had to keep the best of the DAOS record-setting performance while eliminating two key barriers to Pmem adoption: (1) the need to purchase these specialized devices (along with a CPU that could support them) and (2) the necessary loss of a DIMM slot for each Pmem device installed on the motherboard. Both imposed unsurmountable barriers to adoption by many, including most ISPs.

The key technology advance being implemented by the Intel team allows DAOS to store its metadata in DRAM and a Write Ahead Log (WAL) on NVMe SSD (an operation analogous to a journaling filesystem) that preserve the advantages of DAOS as a distributed filesystem. “WAL”, according to Kudryavtsev, “gives DAOS the ability to build the persistent metadata store by using volatile DRAM and NVMe SSD. This way DAOS can reach a high level of parallelism by committing more operations at the same time, which wasn’t possible with the PMEM synchronous model since it holds the CPU until the data is persisted. The changes to support this new operation are localized to the management of the backend storage and has no impact on the overall DAOS protocol / existing API/software ecosystem that is already well advanced.”

WAL gives DAOS the ability to build the persistent metadata store by using volatile DRAM and NVMe SSD. This way DAOS can reach a high level of parallelism by committing more operations at the same time, which wasn’t possible with the Pmem synchronous model since it holds the CPU until the data is persisted. The changes to support this new operation are localized to the management of the backend storage and has no impact on the overall DAOS protocol / existing API/software ecosystem that is already well advanced. — Andrey Kudryavtsev

Kudryavtsev continues, “The performance advantages of DAOS extend to a distributed environment as every node has the keys to the data it owns. This makes DAOS fully distributed, with no single point metadata bottleneck and gives the node full control per dataset over how the metadata is updated.”

On-node and Distributed Cluster Performance Proves Success

The storage of choice for many HPC installations is Lustre. At ISC this year, the Intel DAOS demo included TensorFlow integration for AI applications. The results reported by Intel show that DAOS outperforms Lustre when loading a large AI dataset into TensorFlow in the Cosmoflow application. The beauty is that DAOS and Tensorflow I/O integration was achieved entirely in user space. No kernel modifications were needed, which is a boon to both ISPs and on-premises datacenters.

Another example At the University of Cambridge, DAOS has moved from a research project in the University of Cambridge Open Zettascale Lab to a production testbed on the Cumulus supercomputer run within the University of Cambridge’s Service for Data Driven Discovery (CSD3). Results presented at ISC’22 illustrated the advantages of DAOS over Lustre in a distributed HPC environment. (See DAOS Momentum Demonstrated with New IO500 Rankings and Community DAOS Traction.)

Figure 1. DAOS production installation at the University of Cambridge. Performance comparison conducted on the same storage nodes, detailed configuration available at io500.org.

Users running HPC workloads on Google will appreciate how easy it is to the spin up an instance using the Google HPC-Toolkit. All the HPC components including DAOS are made available automatically. (See the Cloud HPC Toolkit.)

Figure 2. Configuring, building, and deploying an HPC environment in Google Cloud. Credit: Google Cloud

Ivan Poddubnyy (software engineering manager, Intel)  notes this simplicity and automation reflects the success of the Google and Intel collaboration, “Automation is quite important for Google and highlights the benefits of their pipeline plus the ease of implementation for users. The Google use case demonstrates the success of incorporating DAOS into an HPC Toolkit. Simplified DAOS installation leads to easy migration of HPC workloads to the cloud, and between on-premises systems. DAOS is what makes the toolkit run better by delivering the storage performance HPC workloads require. The Google and Intel collaboration was a smart move by both companies because it closes the storage performance gap — even in the TCP-based cloud computing environments.”

DAOS Program Changes

Figure 3. DAOS to introduce two development paths: 1. Stay on Pmem with transitioning to CLX 2.0 persistent memory third party products; 2. Implement DRAM/CXL.mem 1.0 + NVMe Write Ahead Log for metadata.

For those users who are committed to Intel Optane Pmem, Intel still plans to release the next generation Intel Optane Pmem devices codenamed Crows Pass. Intel is committed to supporting customers and ecosystem partners and will continue to provide support for existing memory and storage products through end of life. The company will also support development of Compute Express Link (CXL) on future processors and platforms as they believe it will be the future and standard of tiered-memory solutions.

For more information, see:

Rob Farber is a global technology consultant and author with an extensive scientific and technical background in HPC and machine learning technology.


Notices & Disclaimers

Performance varies by use, configuration and other factors. Learn more on the Performance Index site.

Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available ​updates.  See backup for configuration details.  No product or component can be absolutely secure.

Intel technologies may require enabled hardware, software or service activation.

No product or component can be absolutely secure.

Your costs and results may vary.

© Intel Corporation.  Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries.  Other names and brands may be claimed as the property of others.

 

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…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from 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…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire