Specifications

9
Technology Overview
Xserve
Performance-per-watt
Performance-per-watt is a measure of power efficiency, calculated by dividing the
SPECint_rate_base2000 results by total TDP watts. With the 2.0GHz and 2.66GHz
processors running at 65W per socket and the 3.0GHz processor running at 80W
per socket, the Dual-Core Intel Xeon 5100 processor provides nearly four times the
performance-per-watt versus the Xserve G5.
Benchmark performance: Industry-leading performance-per-watt
The new dual-core Xeon processors in Xserve are over three times more power
efficient than previous Intel Xeon processors and nearly two times more efficient
than the AMD Opteron 285.
Intel Xeon 5100
3.0GHz
AMD Opteron 285
2.6GHz
3.77x
2.17x
PowerPC G5
2.3GHz
Intel Xeon
3.6GHz
1.03x
Baseline
Testing conducted by Apple in October 2006 using preproduction quad 3.0GHz Xeon-based Xserve units; Xserve G5 systems were shipping units.
Estimated SPECint_rate_base2000 score: 112.0 for quad 3.0GHz Xeon system and 21.1 for dual 2.3GHz G5 system. SPEC® and SPECint® are registered
trademarks of the Standard Performance Evaluation Corporation (SPEC); see www.spec.org for more information. AMD Opteron and 3.6GHz Intel
Xeon comparisons based on the best published SPECint_rate_base2000 results for each processor type as of November 9, 2006, and TDP (thermal
design power) values available from public sources. Performance tests are conducted using specific computer systems and reflect the approximate
performance of Xserve.
LINPACK
Computers use double-precision floating-point mathematics to perform calculations
requiring great numerical magnitude or extremely high decimal accuracy. Apple used
the LINPACK benchmark to illustrate the benefits of the Dual-Core Intel Xeon processor’s
two floating-point units. LINPACK measures both single- and double-precision floating-
point performance running a program that solves a dense system of linear equations.
Benchmark performance: Three times the floating-point performance of Xserve G5
Xserve is 3.1 times faster than Xserve G5 with single-precision floating-point
calculations and three times faster with double-precision floating-point calculations.
Xserve G5
Dual 2.3GHz G5
Xserve
Quad 3.0GHz Xeon
Baseline
3.0x
Double-precision results
Xserve G5
Dual 2.3GHz G5
Xserve
Quad 3.0GHz Xeon
Baseline
3.1x
Single-precision results
Testing conducted by Apple in October 2006 using preproduction quad 3.0GHz Xeon-based Xserve units; Xserve G5 systems were shipping
units. All systems were configured with 4GB of RAM. Results are based on the 1000s and 1000d LINPACK benchmarks available from
www.netlib.org/benchmark using a matrix size of 20,000 for the single-precision results and 16,000 (for quad Xeon-based Xserve) and 15,000
(for Xserve G5) for the double-precision results. All systems were tested using the Accelerate Framework. Performance tests are conducted
using specific computer systems and reflect the approximate performance of Xserve.