- Date:
- Monday , June 27, 2005
- Author:
- Kyle Bennett
- Google +1

AMD Athlon 64 FX-57
The AX8 is ABIT’s latest VIA-based Athlon 64 board. Overclocking quirks aside, this board performed surprisingly well. Should you consider VIA as an alternative to NVIDIA for your next AMD motherboard?
Introduction
AMD’s Athlon 64 FX processor series is far from new, but there are a few new features in their FX-57 CPU that we are looking at today. First, for those of you that are not “in the know” about the FX processor, here is a bit of basic background.
When the K8 generation AMD cores started being thrown into the mix, AMD decided to segment their market space into very different defined areas. The one interesting thing about these market definitions is that they have never been very well defined, at least to the outsider looking in. In fact, the only static difference in AMD processors over the last 18 months has been in their pricing levels. Still, the “FX” line of CPUs have headlined as AMD’s “gaming” processor since the introduction of the FX-51. Rounding out AMD’s lineup is the Athlon 64 X2 Dual-Core processor, the Athlon 64, the Sempron, and finally the aging Athlon XP.
The FX processors stay seated firmly as AMD’s gaming processor even though it no longer touts both superior clock speeds and superior L2 cache. Nowadays, the FX series is simply defined by a healthier clock speed as 1MB L2 caches have seeped down into the vanilla Athlon 64 line of CPUs. In fact, if you look hard, you will see that today the Athlon 64 4000+ processor is identical to the now discontinued Athlon FX-53.
My point here is not to beat AMD about the head with negative comments, but rather to explain to our readers that only a single factor differentiates the Athlon FX and many of the Athlon 64 series of products, and that is clock speed.
The AMD Athlon FX-57
What’s New?
As usual, we have the obligatory speed bump with the launch of a new FX. This FX-57 moves along at a very healthy 2.8GHz on the now solidified 939-pin socket architecture. Most notably, probably even more notable than the 200MHz bump, is the new core on which the processor is based. The FX-57 uses AMD’s new 90nm core, codenamed “San Diego.”
The San Diego core brings with it some very important things. Primarily, it has what AMD terms as a “more flexible memory controller.” We at HardOCP would prefer to call it a “fixed memory controller.” “Fixed” as in the older one was broken. The older memory controller had substantial issues with low latency RAM and certain configurations over 2GB. Given the nature of this high end CPU, it settles better with us now that it is able to support high-end memory configurations correctly. This is not a moment too soon as we are starting to see games that might very well benefit from memory configurations up to 2GB.

The next most important feature that comes with this new San Diego core is power utilization. Even though the clock speed has scaled 200MHz, AMD has kept the power envelope virtually the same as its predecessor. Therefore, we are not looking at having to dissipate any more heat from this CPU than the previous FX-55. This can be contributed to a maturing Silicon on Insulator (SOI) process that prevents current leakage on the silicon substrate of the CPU.

My mistake here to show the CPU clocked at 2.4GHz when it fact it does run at 2.8GHz Stock Clock. All FX-57 benchmarks to follow were run at 2.8GHz. Also, the "FX-25" noted above is due to a BIOS glitch.
Lastly, we have AMD adopting Intel’s SSE3 Single Instruction-Multiple Data (SIMD) instruction set. While this is not really a big thing overall, you might see it pick up a ½% in some benchmarks due to this, but I think it would be a stretch to say that you will notice a tangible difference in your computing experience because of it.
What’s Old?
AMD is once again breaking their market segmentation rules with the launch of the FX-57. Until now, the launch of a new FX series processor meant that previous FX model was now discontinued. This time around, AMD is keeping both the FX-55 and the FX-57 alive simultaneously. While their public explanation is product demand, we have to believe that AMD is finally getting better yields as we think the demand has been there all along. In addition, with the dual core Athlon X2 stepping up to the plate to become a mainstream part over the next 12 months, we can be assured there are going to be some changes once again in the Athlon product line. Therefore, it is anyone’s guess as to exactly how the entire lineup will shake out. It seems to us that AMD production capabilities mold their marketing strategy instead of the other way around as many other companies do it.
Pricing
AMD Athlon 64 FX-55 processor $827 each (PIB / 1KU pricing)
AMD ATHLON 64 FX-57 PROCESSOR TECH SPECS
Cache Sizes: 64KB L1 Data + 64KB L1 Instruction Cache / 1MB L2 Cache (Exclusive)
CPU to MC: 2.8GHz
Memory Controller: Integrated 128-bit memory controller
Supported Memory: PC1600, PC2100, PC2700 and PC3200 DDR memory
HyperTransport Links: 1
HyperTransport Spec: 2GHz (2x 1000MHz / DDR)
Effective data bandwidth: 14.4 GB/sec [8GB/sec x1 HyperTransport link + 6.4GB/sec memory bandwidth]
Packaging: 939-pin organic micro-PGA
Fab location: AMD's Fab 30 wafer fabrication facility in Dresden, Germany
Process Technology: 90nm (.09-micron) Silicon on Insulator (SOI)
Approx. Transistor count: 114 million
Approx. Die Size: 115 mm2
Nominal Voltage: 1.35-1.40V
Max Thermal Power: 104 W
Max Icc (processor current): 74.9A
Max Ambient Case Temp: 49-63 degrees Celsius
