From critical memory components to data center infrastructure management, explore our leadership in business-critical continuity on Semiconductor Industry issues and topics. We help our clients take their intellectual capital to market, with in-house research and content-rich analysis. Below is a comprehensive list of our publicly available technical white papers, regularly updated with our latest offerings.
|Title||Download Format||Summary||Date Published|
|Designing a Memory Benchmark (Netlist)||Analyzing the performance of a memory subsystem, as a component of a full system, can be difficult. Most benchmarks are a scaled down version of some real tool, leading to difficulties isolating the performance of the memory from the rest of the server, such as CPU, chipsets, and storage. In addition, the memory’s footprint is often a pitfall. Some tests from many years ago are still utilized, when gigabytes-worth of memory were only a dream, and only exercise a small percentage of RAM. However, modern systems have large caches with multiple levels and benchmarks need to use a LOT of memory to avoid skewing results in the event of a fast on-chip cache. To meet these requirements, the open-source tool, Ramspeed, was selected.||May 2011|
|Mastering EDA Environments with High Performance Memory Technology (Netlist)||Constraints in advancing memory design are the result of many factors. We will explore how memory technology is currently designed, how its cost is derived, and the barriers that exist in designing larger, and more cost effective, capacity. Throughout this process, the inherent need for denser, and faster memory will become clear. Moreover, this paper shows how HyperCloud™ memory from Netlist, Inc., can help overcome these constraints by adding 288GB DRAM running at 1333 MT/s. This can enable 15% improvement in an EDA job runtime.||April 2011|
|Grid Engine Case Study: The Power of Commercial Support for Open Core Products (Univa)||Examine TSAE’s process of evolving from an open source solution to a commercial product (with commercial support and an open source core) from the Univa Corporation. In addition, this case study illustrates the valuable experience that led to an even deeper relationship between TSAE and Univa. Like most business savvy customers, Tata Steel Automotive Engineering continuously monitors operational costs in an effort to guarantee they are receiving quality solutions (at cost-effective prices) for the services and functions necessary to run their business. Therefore, when Tata Steel (i.e., TSAE) originally sought a workload management solution, Grid Engine (at the time offered from Sun Microsystems) was the perfect fit. As an open-source solution, Sun Grid Engine (i.e., SGE) limited initial investment, making it a cost-effective direction. Grid Engine was also selected for a reputation as a very mature, full-featured product, with a long history and proven track record - lending credibility to its reliability. However, self-support solutions, while possessing tremendous advantages, often contain hidden costs and risks.||April 2011|
|Increasing Revenue Through Continuously Optimized Data Center Management (Univa)||Just as it no longer makes sense for us to write programs in machine code, optimizing a technology’s environment is not something that can be accomplished through manual manipulation. Fortunately, a high level of optimization can be accomplished by defining the policy that automates the process - continuously reassessing the total workload, environment resources, and policy criteria for optimal configuration and workload placement. This optimization and reassessment are accomplished through solving numerous, very complex matrices of multi-variable equations. Alone, humans are very poor at solving these types of equations, but with the assistance of computers, they can be solved at speeds faster than the mind can comprehend. Thus, enabling programmed policy to execute the appropriate events in the right sequence, at the right time, and only when necessary, is the best method for optimizing our business technology.|
|This paper will discuss a framework for setting up a scalable / extensible monitoring infrastructure using opensource tools, starting out very simple and providing options to extend the sophistication. It will talk about the policy, the different components, the particulars of how to design, and things to look out for in your implementation of an environment monitoring system.||February 2004|