We took a very logical approach to configuring our test systems. The first thing we did was enter the system BIOS and set the memory timings for optimal performance. In the case of our Asus A7N8X, this meant setting our memory to run synchronously with the processor FSB at 166MHz, with the timings set to 2-4-2-2 with a CAS latency of 2. The memory was similarly configured on our Dual-DDR, Granite Bay powered Intel test bed, but the memory was set for 133MHz operation. After configuring the BIOS, we booted the systems, formatted the hard drives and installed a fresh copy of Windows XP with Service Pack 1. When our Windows installation was complete, we hit the Windows Update site and installed all of the updates currently available (Except for Windows Media Player 9 and the updates associated with Windows Messenger). We then disabled Auto-Updating and System Restore and disabled, then removed Windows Messenger from the system. All of the necessary drivers were installed, and we then defragged the hard drive. Lastly, we installed all of the necessary benchmarking software and ran our tests.

SANDRA (the System ANalyzer, Diagnostic and Reporting Assistant) is an information and diagnostic utility developed by the folks at SiSoftware. Beyond benchmarking, SANDRA can also provide a host of other information about your hardware and operating system. We began our testing with four of the built-in tests that partially comprise the SANDRA 2002 suite (CPU, Multimedia, Memory and Cache sub-system). We ran these tests at the Athlon XP 3000+ CPU's default clock speed of 2.16GHz and compared it to a similarly clocked Athlon XP 2700+, a P4 2.8 and a P4 3.06 with Hyper-Threading enabled.

ATHLON XP 3000+ 2.16GHZ (13X166) CPU	ATHLON XP 2700+ 2.16GHZ (13X166) CPU	PENTIUM 4 2.8GHz 2800MHZ (21X133) CPU	PENTIUM 4 3.06GHz 3066MHZ (23X133) CPU
ATHLON XP 3000+ 2.16GHZ (13X166) Multimedia	ATHLON XP 2700+ 2.16GHZ (13X166) Multimedia	PENTIUM 4 2.8GHz 2800MHZ (21X133) Multimedia	PENTIUM 4 3.06GHz 3066MHZ (23X133) Multimedia
ATHLON XP 3000+ 2.16GHZ (13X166) Memory	ATHLON XP 2700+ 2.16GHZ (13X166) Memory	PENTIUM 4 2.8GHz 2800MHZ (21X133) Memory	PENTIUM 4 3.06GHz 3066MHZ (23X133) Memory
ATHLON XP 3000+ 2.16GHZ (13X166) Cache	ATHLON XP 2700+ 2.16GHZ (13X166) Cache	PENTIUM 4 2.8GHz 2800MHZ (21X133) Cache	PENTIUM 4 3.06GHz 3066MHZ (23X133) Cache

The Athlon XP 3000+ performed very well across the board, coming in second only to the Pentium 4 3.06GHz with Hyper-Threading enabled. ALU and FPU performance was excellent, especially considering this CPU is running a full 900MHz lower than Intel's top-of-the-line model. In fact, if you take SSE2 optimizations out of the FPU performance equation, the Athlon XP 3000+ handily outperforms the P4 3.06GHz. We see a similar situation in the Multimedia tests, but in the Memory bandwidth tests, the P4 coupled with the Granite Bay chipset really shines. The Athlon / nForce combo is no slouch however, breaking the 2.5GB/sec barrier. SANDRA's Cache and Memory test show the Athlon XP 3000+ falling behind Intel's offerings in most instances, except when chunk sizes are between 8kb and 32kb.

Next we ran some tests with Futuremark's (formerly MadOnion) PCMark2002 benchmarking suite. Like other synthetic benchmarks, it's difficult to translate PCMark2002 scores into "real world" performance. However, because it is very easy to run, and produces repeatable, comparable results, PCMark2002 has become a staple here in the H.H. labs. We ran PCMark2002's "CPU" and "Memory" performance modules on all four test systems. The CPU module incorporates the following tests:

The extra cache found on the Athlon XP 3000+ doesn't help it perform too much higher than the similarly clocked 2700+. The 3000+ was only able to best its older cousin by 82 points, which falls well within the margin of error in this test. The Pentium 4s dominated here, with the 3.06GHz CPU with HT enabled smoking the competition.

Raw read, write, and read-modify-write operations are performed starting from a 3072 kilobytes array decreasing in size to 1536 KB, 384 KB, 48 KB and finally 6 KB. Each size of block is tested two second and the amount of accessed data is given as result. In the STL container test a list of 116 byte elements is constructed and sorted by an integer pseudo-random key. The list is then iterated through as many times as possible for 2 seconds and the total size of the accessed elements is given as result. There are 6 runs of this test, with 24576 items in the largest run corresponding to a total data amount of 1536 KB, decreasing in size to 12288 items (768 KB), 6144 items (384 KB), 1536 items (96 KB), 768 items (48 KB) and 96 items in the smallest run corresponding to 6 KB of total data.

Memory bandwidth tests really show the benefits of the Pentium 4's quad-pumped bus architecture. In PCMark2002's memory tests, the P4 2.8GHz and 3.06GHz dominated, besting the Athlon systems by a minimum of 47%. Don't read into this performance difference too much though. There is only so much stock you can place in a synthetic test like this one. It's time we move on to something more relative.