Get email delivery of the Cadence blog featured here
In the SPECTRE 20.1 base release, we released Spectre® XDP-HB as part of the new Spectre X-RF simulation technology. Spectre XDP-HB uses a highly distributed multi-machine multi-core simulation technology to perform HB and HB small-signal analyses. In this blog, we introduce the Spectre XDP-HB technology.
Spectre XDP-HB Overview
Spectre XDP-HB targets the most challenging RF simulation problems, such as HB simulations that require large memory or very long simulation times, or both. By distributing the simulation to multiple machines, Spectre XDP-HB enables you to run HB simulations that could never complete on a single machine before.
Multi-tone advanced node post-layout designs often require huge memory. At times, even small designs may need more memory because of a large number of harmonics required for the simulation. By using distributed memory, the capacity of Spectre XDP-HB has been greatly expanded to cover these applications. In addition, the simulation speed is also accelerated because more cores and machines are involved in the simulation. Spectre XDP-HB can perform simulation beyond 128 cores at the same time.
Currently, Spectre XDP-HB supports only the simulation of driven designs. Support for oscillators, HBSTB, and Envelope analyses will be added in a later release.
The use model of Spectre XDP-HB is the same as distributed Spectre XDP for transient analysis.
Spectre XDP-HB Performance and Memory Scaling
Spectre XDP-HB uses a cloud infrastructure to perform scalable and massively parallel HB simulation. It enables you to improve the capacity and memory utilization using the existing hardware. The CPU scaling can be beyond 128 cores on the same computing host, or across different hosts.
The following figure shows the performance gains for a representative large post-layout advanced node design. An eight-thread simulation is used as the performance reference.
The following figure illustrates the memory reduction for the same representative post-layout design. With two hosts, the simulation needs 65% of the memory per host that would be needed for a single-host simulation run. With four hosts, it needs about 25% of the memory per host. While the total memory consumption of Spectre XDP-HB over all the hosts may be larger, the memory consumption per host is significantly reduced. This reduced memory usage per host enables the higher capacity of Spectre XDP-HB.
In this blog, we introduced Spectre XDP-HB and observed the performance and memory scaling it provides compared to Spectre X-RF on a single host.
You may also contact your Cadence support AE for guidance.
For more information on Cadence products and services, visit www.cadence.com.
About Spectre Tech Tips
Spectre Tech Tips is a blog series aimed at exploring the capabilities and potential of Spectre. In addition to providing insight into the useful features and enhancements in Spectre, this series broadcasts the voice of different bloggers and experts, who share their knowledge and experience on all things related to Spectre. Enter your email address in the Subscriptions box and click SUBSCRIBE NOW to receive notifications about our latest Spectre Tech Tips posts.