News

HP Pushes Windows Server 2003 Scalability Higher

A day after Microsoft and NEC Corp. reached the top of the Transaction Processing Performance Council's OLTP benchmark raw performance list, Microsoft and Hewlett-Packard smashed that record.

Both the NEC and HP results are based on Intel's unreleased third-generation of the Itanium processor, code-named "Madison."

The announcements were timed to coincide with the launch of Microsoft's Windows Server 2003, the first general release of Windows server that supports the 64-bit Itanium processor.

The record-breaking HP system contained 64 processors, twice as many as have ever been used in a Windows-based TPC-C benchmark.

The HP/Microsoft/Intel test achieved 658,277 tpmC (transactions per minute on the TPC-C test). That's a 28 percent performance boost over the 514,035 tpmC submitted to the TPC by NEC/Microsoft/Intel on Wednesday.

The result is also a 45 percent bump in performance over a 128-processor result run by Fujitsu on Sparc processors and Solaris back in 2001, which was the No. 1 result until Wednesday.

The 64-processor HP Superdome server, the backbone of a $6.4-million configuration, did well on Windows servers' traditional stronghold of price performance as well, registering a price per tpmC of $9.82.

During his Windows Server 2003 launch keynote Thursday, Microsoft CEO Steve Ballmer said the NEC and HP results together validate a promise he made many years ago when hiring David Cutler, the architect of Windows NT.

"When we originally hired Dave Cutler, he said, 'Look, I don't want to work on toy operating systems. I want to work on real operating systems.' I know I can now look David in the eye and say, we're past that for sure," Ballmer said.

The system maxed out the capabilities of the 64-bit version of Windows Server 2003 Datacenter Edition, using 64 processors and 512 GB of RAM. The benchmark also used the 64-bit version of SQL Server 2000 Enterprise Edition, which was also launched on Thursday.

The new "Madison" Itanium processors are due out in the middle of the year. They have a clockspeed of 1.5 GHz and 6 MB of cache.

Intel president and chief operating officer Paul Otellini said the processors are only a start for Intel Itanium-based scalability in the next few years.

While the follow-on to "Madison" will be a lower-end Itanium chip called "Deerfield," the one after that slated for 2005 will bring multiple-core technology to the processor. "These chips will take the absolute performance another two to three times higher," Otellini said. A few years after that, Otellini promised, "We have in development an Itanium processor that will take this up to a factor of 10 of where this [pointing to the HP Superdome on the stage] is today."

About the Author

Scott Bekker is editor in chief of Redmond Channel Partner magazine.

Featured

  • Closeup of the new Copilot keyboard key

    Microsoft Updates Copilot To Add Context-Sensitive Agents to Teams, SharePoint

    Microsoft has rolled out a new public preview for collaborative "always on" agents in Microsoft 365 Copilot, bringing enhanced, context-aware tools into Teams channels, meetings, SharePoint sites, Planner workstreams and Viva Engage communities.

  • Windows 365 Cloud Apps Now Available for Public Preview

    Microsoft announced this week that Windows 365 Cloud Apps are now available for public preview. This aims to allow IT administrators to stream individual Windows applications from the cloud, removing the need to assign Cloud PCs to every user.

  • Report: Security Initiatives Can't Keep Pace with Cloud, AI Boom

    The increasingly fast adoption of hybrid, multicloud, and AI systems is easily outgrowing existing security measures, according to a recent global survey by the Cloud Security Alliance (CSA) and exposure management firm Tenable.

  • World Map Image

    Microsoft Taps Nebius in $17B AI Infrastructure Deal To Alleviate Cloud Strain

    Microsoft has signed a five-year, $17.4 billion agreement with Amsterdam-based Nebius Group to expand its AI computing capabilities through third-party GPU infrastructure.