AMD's Athlon 64 FX-53

AMD's Athlon 64 FX-53
The FX Gets Its First Speed Bump...

By, Marco Chiappetta
March 18, 2004

We spent some time exploring the Athlon 64 FX-53's inner workings with the latest version of CPU-Z, before we ran it through the wringer with our suite of benchmarking applications.  The information in the screenshots below show the CPU particulars, cache configuration, and memory timings with the FX-53 installed in an nForce3 powered SK8N motherboard equipped with 1GB of low-latency Mushkin RAM...

AMD Athlon 64 FX-53 Vital Signs Default & Overclocked Speeds and Temps

There are a couple of interesting things to point out with regard to this CPU-Z information.  If you take a look at the CPU information page, the processor is correctly identified as a Socket 940 Athlon 64 FX-53 built on a .13 micron manufacturing process, and as you can see our particular FX-53 is a rev. SH7-CG with an 'A' stepping code.  The 2400MHz (2.4GHz) stock clock speed is achieved by multiplying 12 (the default multiplier) by the 200MHz HyperTransport clock.  The Cache information page reports a full 1024K of full-speed L2 cache, along with the additional 64K of L1 Data cache and 64K of L1 Instruction cache.  We've also included CPU-Z's memory configuration information to show the actual timings of Mushkin's new low-latency PC3200 registered DIMMs.  When the FX first debuted last year, low-latency registered DIMMs were virtually non-existent, but this is no longer the case.  And using this "faster" memory certainly gave the FX a performance boost, as you'll see in the pages ahead.

Athlon 64 FX-53 2.65GHz (13x204MHz) CPU-Z: CPU

Athlon 64 FX-53 2.65GHz (13x204MHz) SANDRA: CPU

Athlon 64 FX-53 2.65GHz (13x204MHz) SANDRA: MM

Athlon 64 FX-53 2.65GHz (13x204MHz) SANDRA: Memory

When we initially reviewed the Athlon 64 FX-51, and tried to overclock it, our results weren't exactly spectacular.  We were able to take our FX-51 sample up from its default 2.2GHz clock speed to only 2.34GHz.  Back then, we speculated that a switch to a .09 micron manufacturing process was probably necessary for AMD to push the FX's clock speed much higher.  So, when the FX-53 arrived with a default clock speed of 2.4GHz built using the same .13 micron process as the FX-51, we suspected there wouldn't be much overclocking headroom, but we were pleasantly surprised.

With a modest .15v bump in voltage, up to 1.65v from the default 1.5v, we were able to push our particular Athlon 64 FX-53 all the way up to 2.652GHz - roughly a 10.5% overclock and a full 300+MHz higher than our FX-51.  It seems like AMD has been hard at work refining their .13 micron Silicon on Insulator (SOI) manufacturing process.  We got to 2.652GHz by first raising the CPUs multiplier to 13, and raising the 200MHz HT clock until the system was no longer stable.  You should also know that our particular SK8N motherboard has a penchant for corrupting the BIOS when the HT clock is set too high, so we focused on multiplier overclocking here.  And remember, you won't need an engineering sample to alter the FX's multiplier, should you opt for one of these CPUs, as AMD is shipping all Athlon 64 FXs unlocked.

Also note that while overclocking the Athlon 64 FX-53, all we used was the stock heatsink / fan combo supplied by AMD.  Even with this modest heatsink, the FX-53 ran relatively cool throughout all of our testing.  At default clock speeds, we never saw the FX-53 break the 43°C mark, and this was after hours of gaming and benchmarking in a closed mid-tower case.  At a max temp of 43°C with the FX-53 at default clock speeds, it's running about 20% - 25% cooler than Intel's high-end CPUs at the moment.  When we over-volted and overclocked the processor, temperatures obviously went up, but they still remained manageable.  After running for about an hour at 2.65GHz, the FX-53 peaked at "only" 55°C.  This temperature is much lower than what we saw with a 3.2GHz "Prescott" at default speeds, and only a couple of degrees higher than the 3.4GHz P4 Extreme Edition.  Kudos to AMD for know how to keep things cool under pressure and for engineering a CPU core with more than reasonable thermal characteristics.

HotHardware's Test Systems Not all are created equal...

How we configured our test systems:
We made an effort to ensure that all of our test systems were configured as similarly as possible for this review.  All three of the Athlon 64 systems and the P4 systems were equipped with identical hardware, with the obvious exceptions being the motherboards and processors.  The same applied to the Athlon 64 FX 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, we entered 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.  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.

If you've read our recent articles covering the official launches of Intel's new "Prescott" core and 3.4GHz Extreme Edition CPU, and AMD's Athlon 64 3400+, you may notice that we've made some subtle, yet significant changes to our test systems.  For this review, we've replaced the GeForce FX 5900 Ultra used in those articles with an ATi Radeon 9800 Pro, and as we mentioned earlier, our Athlon 64 FX test system has also been upgraded with some new low-latency (2-3-2-6) Mushkin RAM.  The Asus SK8N nForce3 based motherboard we used to test the Athlon 64 FX CPUs has gotten a BIOS upgrade (v1004) as well.  These changes to our test beds had a marginal effect on performance in some tests, but some others show a more significant difference (some for the better, some for the worse).

Synthetics - SANDRA and PC Mark 2004