AMD's Athlon 64 FX-51

The Athlon 64 FX-51 Processor
AMD Drops the Hammer, On Your Desktop!

By, Marco Chiappetta
And Dave Altavilla
September 23, 2003

How We Configured Our Test Systems:

We tried to ensure that all of our test systems were configured as similarly as possible.  Both of the Pentium 4 systems and the Athlon XP system were equipped with identical hardware, with the obvious exceptions being the motherboards and processors.  The same applied to the Athlon 64 FX-51 system, but because it required registered DIMMs, the memory was different as well.  The video cards, hard drives, driver versions (where applicable) and OS configurations were identical.  Before we started benchmarking these systems, we enter the system BIOS and set each board to their "Optimized Defaults".  We then configured our RAM to run at 200MHz (DDR400), with the timings set by the SPD.  The hard drives were then formatted, and Windows XP Professional (SP1) was installed.  When the installation was complete, we hit the Windows Update site and downloaded all of the available updates, with the exception of the ones related to Windows Messenger and Media Player 9. 

Then we installed all of the necessary drivers, and removed Windows Messenger from the system altogether.  Auto-Updating and System Restore were disabled as well, and we setup a 768MB permanent page file on the same partition as the Windows installation.  Lastly we set Windows XP's Visual Effects to "best performance", installed all of our benchmarking software, defragged the hard drives and ran all of the tests.

HotHardware's Test Systems A Prosumer's Fantasy

We should take time to note, that Intel made a proverbial "dive and catch" play, in an effort to take the steam out of AMD's launch today.  Late last week, we were rushed an early release sample of Intel's new Pentium 4 Extreme Edition chip, which has an additional 2MB of L3 Cache on board.  This Northwood based core, runs at 3.2GHz as well but with the extra on board high speed cache, should have significantly better through-put in memory bandwidth sensitive situations, like gaming, audio and video encoding and 3D rendering.

Our contacts at Intel claim that the Athlon 64's real advantage is its large on chip cache, HyperTransport link and integrated 128 bit memory controller, rather than its 64-bit architecture.  We'll see in the benchmarks ahead but it sure was interesting to see Intel actually "sweat" a bit, in an effort to react to what they knew would be a truly compelling product from their strong competitor in the Desktop space.

The Athlon 64 FX-51: Synthetic Benchmarks SiSoft SANDRA | PCMark2002 | 3DMark03 CPU

We began our testing with a few synthetic tests.  First up, we have some results from SiSoftware's SANDRA 2003 MAX.  SANDRA, the System ANalyzer, Diagnostic and Reporting Assistant, consists of a set of informational and diagnostic utilities.  It isn't only a tool for benchmarking though, SANDRA also provides a host of other useful information about your hardware and operating system. We ran four of the built-in sub-system tests that comprise the SANDRA 2003 suite (CPU, Multimedia, Memory and Cache).  All of these tests were run with the Athlon 64 FX-51 set to its default clock speed of 2.2GHz.
 

SANDRA CPU ATHLON 64 FX-51 2200MHZ

SANDRA MULTIMEDIA ATHLON 64 FX-51 2200MHZ

SANDRA CACHE ATHLON 64 FX-51 2200MHZ

SANDRA MEMORY ATHLON 64 FX-51 RAM @ 200MHZ

The SANDRA scores don't show the Athlon 64 FX-51 in a very positive light.  However, scores from these types of synthetic tests are difficult to translate into "real world" performance.  SANDRA's CPU test puts the Athlon 64 FX-51 ahead of the similarly clocked Athlon XP 3200+, but the Pentium 4 3.2GHz, with Hyperthreading enabled, ran away with this one.  The results were the same in the Multimedia test, with the FX-51 besting every other reference system except for the P4 3.2GHz.  SANDRA's cache benchmark also shows the P4 with a significant advantage, especially with smaller block sizes.  The FX-51 didn't pull ahead until the block sized broke the 512k mark.  On the other hand, the Athlon 64 FX-51 rocked SANDRA's memory bandwidth test, crushing a Dual-DDR i875 rig by roughly 11%.  This memory bandwidth and latency advantage is due most likely to the integrated memory controller onboard the Athlon 64 FX-51.

Next, we ran PCMark2002's "CPU" and "Memory" performance modules on four different test systems, again with the FX-51 clocked at its default 2.2GHz clock speed.  From this point forward, we'll be directly comparing the performance on the Athlon 64 FX-51 test system to three other high-end machines, that were similarly configured.  PCMark2002's CPU module incorporates the following types of tests:

CPU Tests:

• JPEG decompression

• Zlib compression & decompression

• Text search

• MP3 Audio Conversion

• 3D Vector Calculation

Like SANDRA, PCMark2002's CPU performance module puts the Athlon 64 FX-51 ahead of the Athlon XP 3200+, but somewhat behind the 3.2GHz P4s.  The Extreme Edition P4's L3 cache doesn't help it much in this test, but the FX-51 does pull well ahead of 3200+.  Although they are both clocked at 2.2GHz, based on the results of the SANDRA and PCMark2002 CPU tests, it seems like the Athlon 64 XP-51 is faster than an Athlon XP clock-for-clock.

PCMark2002 Memory Test technical background: (Quote Taken From Futuremark)
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.

The integrated memory controller and other architectural enhancements really helped the Athlon 64 FX-51 in PCMark2002's Memory performance module.  Had Intel not launched their sneak attack with the P4 Extreme Edition, the Athlon 64 FX-51 would have walked away with a victory here.  The P4 Extreme Edition's huge L3 cache gave it a big boost in this test, in the neighborhood of 41%.  Looking at these results tells us PCMark2002's Memory Performance module benefits more from decreased latency, than increased bandwidth.  Both of the P4s were running with the same system bus and with the same memory configuration, yet the performance delta between the two CPUs was dramatic.
 

We also ran the default CPU tests that are built into Futuremark's much maligned 3DMark03.  These tests consist of two different 3D scenes, that are generated with a software renderer, which is dependant on the host CPU's computational ability.

The Athlon 64 FX-51 pulled way ahead of all of the other systems in this test.  The 3.2GHz Pentium 4 schooled the Athlon XP 3200+ by about 11%, with the 3.2GHz P4 Extreme Edition besting the standard P4 by 14%.  The Athlon 64 FX-51 however, outperformed Intel's best by 7.5% and smoked the 3200+ by roughly 37%.  AMD is claiming the Athlon 64 FX-51 is a gamer's CPU.   If 3DMark03's CPU results are indicative of real world gaming scenarios, AMD is definitely going to make hardcore gamers giddy with delight.

3D Modeling & Animation