40 server vendors to use Intel ‘Tulsa’ chips

At least 40 server server manufacturers announced plans to incorporate Intel Corp.’s new dual-core 64-bit Xeon 7100 processor, dubbed “Tulsa,” into upcoming servers.

Hewlett-Packard Co., Dell Inc., and Unisys Corp. are among 40 system manufacturers designing Tulsa into their new server offerings.

Tulsa is Intel’s response to the Advanced Micro Devices Inc.’s (AMD) Opteron 800-series processors, the most recent challenger to Intel’s dominance in the server processor market.

HP Tuesday announced that Tulsa will go into upgraded versions of its x86 platform ProLiant 500 series servers. The ProLiant DL580 rack-mounted server carries a list price of US$6,649, while the ML570 G4 tower server starts at $5,799. HP says that offering ProLiants with the Tulsa chip will strengthen its market share lead in the x86 processor segment, which stood at 34.5 percent in the second quarter, according to IDC research.

Dell is offering Tulsa in new versions of its PowerEdge 6800 and 6850 servers, starting at a list price of $6,900. Dell reports performance gains of up to 123 percent and performance per watt gains of up to 129 percent, compared to servers running Intel’s “Paxville-MP” chip line, the predecessor to Tulsa.

Unisys said it plans to use Tulsa in its next generation ES7000 sometime in the fourth quarter, and later in its ClearPath server lines. Unisys does not use AMD chips, said Mark Feverston, director of enterprise servers and storage at Unisys.

“We know our market, we know their buying patterns. They’re happy with Intel,” Feverston said.

But Intel lost significant market share to customers who weren’t happy. In April, AMD cited Mercury Research numbers that showed its share of the x86 processor market grew to 22 percent in June from just 7 percent in June 2005.

“It certainly added to our sense of urgency,” said Tom Kilroy, vice president of Intel’s digital enterprise group. But he also said Intel launched Tulsa to meet customer demand for processors that increased performance while reducing energy consumption and enabling virtualization.

Tulsa, based on a 65-nanometer chip design, is built to run on servers with four or more processors, Intel said. Tulsa’s cores run 13 percent faster than Paxville, while using 20 to 40 percent fewer watts, the company said. Intel builds each Tulsa chip by combining two 3.4GHz Pentium 4 cores on a single die, and will deliver both 150-watt and 95-watt versions. Tulsa supports four threads per processor and has a 16M-byte cache, compared to a 4M-byte cache in Intel’s Woodcrest processor, another member of the Xeon family introduced in June.

Intel will use the Tulsa chip to fill a price and performance gap in its line between the “Woodcrest” Xeon 5100 and the “Montecito” Dual-Core Itanium 2, said Shane Rau, an analyst with IDC.

“With Itanium, users are looking primarily at performance, and with Woodcrest they are looking at price. Something like (Tulsa) will go right down the middle,” he said.

Intel is selling the “Tulsa” dual-core Xeon 7100 series chips in eight models, including two with caches of 16M bytes, two with 8M-byte caches, and the rest with 4M bytes.

Intel charges $1,980 for either the 3.40GHz Xeon 7140M or 3.33GHz Xeon 7140N, $1,391 for either the 3.20GHz Xeon 7130M or 3.16GHz Xeon 7130N, $1,177 for either the 3.00GHz Xeon 7120M or 3.00GHz Xeon 7120N, and $856 for either the 2.60GHz Xeon 7110M or 2.50GHz Xeon 7110N.

Related Download
3 reasons why Hyperconverged is the cost-efficient, simplified infrastructure for the modern data center Sponsor: Lenovo
3 reasons why Hyperconverged is the cost-efficient, simplified infrastructure for the modern data center
Find out how Hyperconverged systems can help you meet the challenges of the modern IT department. Click here to find out more.
Register Now