AMD and Nvidia quit SySmark benchmark group

Summary:The chip leaders have pulled out of the BapCo industry benchmark, after years of complaints by AMD that the application performance evaluation favours Intel

AMD and Nvidia have withdrawn from the SySmark application performance benchmark programme, leaving Intel as the only major chip participant.

The resignation of AMD from the BapCo group, which administers the SySmark benchmark, was revealed on Tuesday. Its announcement came two weeks after the release of the 2012 version of the benchmark, which measures the performance of applications in a variety of hardware environments.

In its statement, AMD said it believes the benchmark does not accurately reflect the performance of its chips.

"The SySmark benchmark is not only comprised of unrepresentative workloads (workloads that ignore the importance of heterogeneous computing and, frankly, favour our competitor's designs), but it actually generates misleading results," Nigel Dessau, the company's chief marketing officer, wrote in a blog post.

Both AMD and Nvidia design processors with a heterogeneous architecture, in which large and small cores are distributed across the same die. By contrast, Intel's chips are homogenous and do not mix differently sized cores on the same die.

Nvidia confirmed to ZDNet UK that it has withdrawn from the SySmark benchmark, but declined to explain why.

BapCo is disappointed that a former member of the consortium has chosen once more to violate the confidentiality agreement they signed, in an attempt to dissuade customers from using SySmark.

– BapCo

The BapCo consortium includes chipmaker Intel, along with Dell, HP, Hitachi, Microsoft, Samsung, Toshiba and others. On Tuesday, the group defended its 2012 benchmark, noting that AMD participated in its two-year development.

"AMD voted in support of over 80 percent of the SySmark 2012 development milestones, and were supported by BapCo in 100 percent of the SySmark 2012 proposals they put forward to the consortium," the group said in a statement.

"BapCo is disappointed that a former member of the consortium has chosen once more to violate the confidentiality agreement they signed, in an attempt to dissuade customers from using SySmark to assess the performance of their systems," it added.

SySmark benchmark

According to BapCo, the SySmark benchmark is used to measure performance by PC makers, hardware and software developers, IT departments, system integrators, publishers and testing labs. It also argued that the applications used in the 2012 package give a "real-world" assessment, as they include Microsoft Office, Adobe Acrobat and Autodesk AutoCad.

The AMD-BapCo row over SySmark has been running for at least a decade. In 2002, AMD accused BapCo of removing applications from the benchmark that showed a performance advantage for AMD over Intel.

In 2005, reviewers at ZDNet found that "Intel's processors perform better under BapCo's SySmark thanks to its greater prevalence of Pentium 4-optimised applications".

In 2010, AMD criticised the SySmark 2007 benchmark, the last one before the 2012 version. The chip designer said the 2007 package did not easily allow IT decision makers to break down the figures by user roles and argued this would lead to a bias toward Intel architecture chips.


Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.

Topics: Processors

About

Jack Clark has spent the past three years writing about the technical and economic principles that are driving the shift to cloud computing. He's visited data centers on two continents, quizzed senior engineers from Google, Intel and Facebook on the technologies they work on and read more technical papers than you care to name on topics f... Full Bio

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Related Stories

The best of ZDNet, delivered

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
Subscription failed.