Enabling 3X Verification Performance on HPE GreenLake Flex Solutions with Synopsys FlexEDA

The semiconductor industry is undergoing a period of dramatic transformation. Designs are getting larger, and with multi-die architectures, design complexity is increasing even further. To ensure robust design quality, the required verification is increasing exponentially. Engineering teams are under constant pressure to deliver quality designs under shrinking time to market, yet many organizations still rely on EDA licensing and compute models created decades ago—models that often slow down rather than accelerate chip development.

The combination of Synopsys FlexEDA Pay-Per-Use (PPU) licensing and HPE GreenLake Flex Solutions changes this paradigm. Together, they deliver a modern, secure, on-premises environment with the elasticity and efficiency of cloud consumption. This blog explores how this hybrid approach accelerates regressions, improves design quality, and meaningfully reduces time-to-market.

ASIC development is inherently iterative and compute-intensive. Every design cycle involves specification, RTL creation, verification, synthesis, physical design, signoff, and Engineering Change Order (ECO) loops. Each stage places different, often unpredictable, demands on compute and EDA license resources.

Traditional approaches create several persistent challenges:

• EDA License Bottlenecks: Fixed quantities of term licenses mean engineers often wait for access to tools instead of running regressions or validating fixes. By the time additional licenses are procured and installed, precious engineering time is already wasted. In addition, due to the spiky nature of the regression workloads, EDA tool licenses are either under or over provisioned, resulting in sub-optimal utilization of precious resources.
• Underutilized Resources: To avoid compute bottlenecks, companies frequently overprovision compute hardware, which remains idle when workloads fluctuate.
• Slow Scalability: Traditional CAPEX procurement cycles make it difficult to respond quickly to market shifts or new internal schedule pressures.
• Security and Compliance Constraints: From RTL to verification environments to design databases, sensitive IP may not be able to leave on-premises environments, potentially making public cloud adoption challenging.
• Operational Overhead: Managing hybrid environments, multiple vendors, and distributed teams increases operational complexity.

Fueled by rising demand for software-defined hardware such as AI accelerators, as designs expand into technology domains like 2nm/3nm, chiplets (multi-die), limitations of traditional models amplify schedule risk and inhibit engineering efficiency.

The Synopsys FlexEDA licensing model represents a fundamental shift from traditional static licensing. In addition to being able to purchase a fixed number of licenses, engineering teams can now leverage highly scalable EDA tools on-demand by the minute.

Key benefits of FlexEDA Pay-Per-Use (PPU) include:

• On-Demand Scaling: Teams can instantly ramp up to dozens, hundreds, or even thousands of EDA licenses during peak periods such as nightly regressions, coverage analysis, or tape-out crunch time.
• Faster Time-to-Market: By removing licensing constraints, engineers can execute more tests in parallel, perform broader what-if analyses, and validate fixes immediately. This leads to engineering productivity improvements up to 40%.
• Automated License Management: Cloud-native licensing infrastructure enables frictionless automatic scaling of license servers without involvement from CAD or IT teams. This ensures nightly regressions run smoothly and engineers get the results on time, every time.
• Flexible Deployment: Synopsys FlexEDA supports two models: SaaS (Software as a Service) and BYOC (Bring-Your-Own-Cloud or Bring-Your-Own-Compute). BYOC enables organizations to keep their sensitive IP data on-prem while benefiting from cloud-like elasticity.

The result is an EDA licensing model aligned with the dynamic nature of semiconductor design development and one that supports scalability and maximizes efficiency.

While FlexEDA PPU removes EDA licensing limitations, compute capacity may emerge as the next constraint. HPE GreenLake Flex Solutions addresses this by delivering on-premises infrastructure with cloud-like elasticity.

HPE GreenLake Flex Solutions uses a two-tier capacity model:

• Reserve Capacity: Always-on resources to support baseline workloads.
• Variable Capacity: Elastic additional compute that can be activated instantly when workloads spike.

The combination of FlexEDA licensing and elastic on-premises compute unlocks new workflows that were impossible under traditional models. When the HPE Slingshot ASIC team deployed FlexEDA PPU on HPE GreenLake Flex Solutions, the impact on regression and coverage acceleration was measurable and immediate.

Regressions were run using fixed (TSL) and PPU for three unique design blocks. As seen in all three runs, with PPU EDA licenses, the duration of regression runs reduced substantially.

Adopting Synopsys FlexEDA PPU and HPE GreenLake Flex Solutions introduces new opportunities, but also new operational considerations.

Compute Sizing

Teams should find the right balance of HPE GreenLake Flex Solutions reserve and variable capacity based on regression counts, number of testcases, and acceptable completion windows.

Hardware Performance

EDA license cost typically outweighs compute cost. Faster compute reduces total license minutes consumed, making modern and scalable hardware critical when using FlexEDA PPU licensing.

Job Launch Throughput

To fully use high compute slot counts, the job scheduling infrastructure must be tuned to launch hundreds or thousands of jobs rapidly. Bottlenecks here can undermine the benefits of PPU parallelization.

Cost Governance

On-demand, Pay-Per-Use licensing introduces a need for:

• Job tagging
• Usage monitoring
• Spending guardrails

Different tests consume dramatically different simulation minutes, making visibility essential for budget planning. Synopsys FlexEDA provides a GUI-based platform to monitor EDA PPU usage, along with governance capabilities, to ensure access and budget is controlled tightly.

Despite its advantages, FlexEDA PPU isn't meant to replace term-based EDA licensing entirely. Term-based Cloud Subscription Licenses (CSL) are also available with FlexEDA. Similarly for compute, HPE GreenLake Flex Solutions is not intended to replace on-prem compute infrastructure. Instead, it augments compute capacity when there is an on-demand need. Many organizations will benefit from a hybrid model.

• Fixed quantity FlexEDA CSLs for predictable, everyday workloads
• FlexEDA PPU licenses for high-intensity and high-impact phases
• Reserve and Variable capacities from HPE GreenLake Flex Solutions

This combination ensures cost-effective baseline capacity while enabling unlimited elasticity during peak periods.

As software-defined hardware continues to evolve, traditional fixed EDA licenses and static on-premises compute infrastructure are not well-suited to meet the demands of scale, speed, and agility. As chip designs become more complex and time-to-market windows shrink, these established models create bottlenecks that slow the pace of innovation and introduce schedule risk.

Synopsys FlexEDA Pay-Per-Use licensing, combined with HPE GreenLake Flex Solutions, provides a secure, flexible, consumption-based design environment that unlocks massive parallelization capabilities, resulting in faster regressions, shorter coverage cycles, and more agile engineering workflows.

Chip design teams striving to improve productivity, optimize costs, accelerate time-to-market, and ensure first pass silicon success, while keeping all proprietary design IP in-house, will find this solution offers a compelling path forward.