Azure NetApp Recordsdata is redefining what’s doable for EDA within the cloud—delivering scalable, high-performance storage that helps large concurrency, low latency, and constant manufacturing efficiency. With unbiased benchmark validation and real-world adoption, organizations can now run EDA workloads at scale with out conventional storage bottlenecks.
Final 12 months, we outlined how Azure NetApp Recordsdata helped reshape silicon design by delivering the low-latency, high-throughput storage required for Digital Design Automation (EDA) workloads at cloud scale. Since then, we now have continued to increase efficiency and scalability. At this time, we’re advancing that progress with one other vital step ahead.
Trendy semiconductor design is outlined by scale. 1000’s of concurrent EDA jobs spanning simulation, synthesis, and verification run constantly in opposition to shared datasets, the place even small variations in storage latency can ripple throughout total design cycles. For a lot of groups, this has traditionally restricted how far EDA workflows may scale within the cloud.
That constraint is now altering.
Azure NetApp Recordsdata (ANF) is redefining what is feasible for EDA within the cloud by delivering predictable, high-performance shared storage at large concurrency. With new unbiased benchmark outcomes and rising adoption by main semiconductor corporations, Azure is establishing itself as a viable—and in lots of circumstances superior—platform for contemporary EDA environments.
Why EDA storage has been troublesome to scale within the cloud
EDA workloads mix three traits which have historically challenged cloud storage architectures:
Extraordinarily excessive concurrency, with 1000’s of jobs accessing shared file techniques concurrently.
Strict latency sensitivity, the place even minor delays scale back compute effectivity and prolong runtimes.
Intensive shared information entry patterns, creating rivalry below load.
Whereas cloud compute scales simply, shared storage has usually launched variability that limits general system effectivity. As concurrency will increase, storage turns into the bottleneck, impacting regression cycles, rising device license prices, and slowing time to tape-out.
For EDA groups evaluating cloud transformation, the central query has remained constant: can storage scale with compute whereas sustaining predictable efficiency?
A contemporary strategy: Azure NetApp Recordsdata for EDA at scale
Azure NetApp Recordsdata is designed particularly to handle this problem. Its structure aligns on to the necessities of extremely parallel, shared workloads like EDA.
At its core, ANF permits unbiased scaling of compute and storage, so EDA clusters can develop with out storage turning into the constraint, and extra compute nodes don’t introduce hotspots or rivalry on the storage layer. It natively helps concurrent metadata operations at scale, dealing with the tens of millions of small file interactions typical of EDA workflows with out degradation. And its service-level efficiency mannequin ensures that throughput and IOPS scale predictably with capability, eliminating the necessity for advanced tuning.
Extra not too long ago, improvements equivalent to massive volumes and huge volumes breakthrough mode have expanded the concurrency envelope even additional. These capabilities enable 1000’s of parallel jobs to share a single storage surroundings whereas sustaining constant latency below sustained load.
This delivers what cloud-based EDA techniques have lengthy struggled to supply: constant, repeatable efficiency, not solely at low utilization, but in addition below full manufacturing load.
Unbiased validation: SPECstorage® Answer 2020 benchmark outcomes
To validate these capabilities in a real-world context, Azure NetApp Recordsdata was measured utilizing the industry-standard SPECstorage® Answer 2020 EDA_BLENDED benchmark. This benchmark simulates life like EDA workflows by combining metadata-intensive frontend operations with throughput-heavy backend processing, all below strict latency necessities.
These outcomes reveal a number of essential traits:
The flexibility to maintain very excessive ranges of concurrent EDA workloads.
Constantly low response occasions below load.
Linear scaling habits as concurrency will increase.
No requirement for overprovisioning.
Traditionally, prime benchmark outcomes on this class have been related to tightly built-in on-premises techniques. This validation underscores a broader shift within the {industry}: when architected accurately, cloud-based EDA infrastructure cannot solely match on-premises approaches, however in some situations surpass them in each scale and operational effectivity.
Confirmed in manufacturing: EDA workloads already operating on ANF
This efficiency shouldn’t be restricted to benchmarks. Organizations equivalent to AMD and ASML are already utilizing Azure NetApp Recordsdata to run EDA and high-performance design workloads in manufacturing environments.
These corporations function at the forefront of semiconductor innovation, the place infrastructure should assist each excessive scale and exact predictability. Their adoption of ANF displays a broader {industry} development: shifting EDA workloads to the cloud is now not experimental, it’s turning into a strategic benefit.
These prospects, together with others, persistently report the identical operational advantages:
The flexibility to extend regression concurrency with out efficiency degradation.
Improved utilization of compute assets and decreased EDA device license charges.
Better predictability in design cycles, enabling extra assured scheduling of key milestones.
On this context, storage is now not the limiting issue—it turns into an enabler of scale.
How Azure helps EDA groups scale with confidence
Organizations have flexibility in how they deploy EDA environments with Azure NetApp Recordsdata, relying on workload traits and operational priorities.
Some groups select a centralized mannequin constructed round a single massive quantity to maximise throughput and tightly management latency. Others undertake a multi-volume strategy to distribute workloads and scale concurrency throughout completely different job varieties. Many enterprises prolong present on-premises environments into Azure, utilizing cloud capability to soak up peak demand with out everlasting infrastructure growth.
Throughout all of those patterns, one precept stays constant: storage efficiency should scale predictably alongside compute. Azure NetApp Recordsdata gives that basis.
Azure NetApp Recordsdata delivers the constant, excessive‑throughput NFS efficiency that trendy EDA workloads demand, shrinking runtimes, accelerating tape‑out schedules, and giving chip designers the boldness that storage won’t ever be the bottleneck.
Srikanth Gubbala, Head of International HPC Infrastructure, Utilized Supplies
Bringing all of it collectively
The evolution of cloud storage for EDA marks an essential inflection level for the semiconductor {industry}. What was as soon as thought-about a tradeoff—scale versus predictability—is now not a constraint.
With Azure NetApp Recordsdata, organizations can confidently run extremely concurrent EDA workloads within the cloud, supported by structure designed for his or her particular calls for and validated by unbiased benchmarking.
For groups exploring the way to modernize their EDA infrastructure, the trail ahead is more and more clear. Cloud-based storage can now meet the necessities of even probably the most demanding design environments, whereas providing the pliability to scale as workloads proceed to develop.
For a deeper technical exploration of the benchmark configuration and design issues, see the companion Azure Tech Group technical weblog: “From scale to breakthrough: Azure NetApp Recordsdata units a brand new cloud benchmark for EDA.”
For additional data, discover the Azure NetApp Recordsdata documentation or e-mail askanf@microsoft.com.
Scale high-performance EDA workloads with Azure NetApp Recordsdata
Uncover how Azure NetApp Recordsdata delivers predictable, high-performance storage for EDA workloads, enabling large concurrency, low latency, and constant scaling in manufacturing.
