by Bubba “MasterFung” Wolford
Introduction
AMD is launching their second Athlon64 (A64) CPU today. This 3400+ is running at 2.20GHz and comes on a 1600MHz (1.60GHz) Hypertransport connection between the memory and CPU or a 2.20GHz connection between the memory controller (integrated) and CPU. You might recall we reviewed the FX-51 just a few months ago and found it’s performance to be spectacular. The 3XXX+ series have now become AMD’s “mainstream” part and thus should not be considered in the same class as the FX-51. Intel should be countering the 3XXX series with their upcoming Prescott CPU core but we have yet to see Prescott make its introduction. We believed that we might see Prescott about when the FX-51 came to market, but Intel has been rumored to be having heat issues with Prescott and has delayed it’s launch to sometime later this quarter. In the meantime, AMD is running right along with their A64 core and seeing wonderful results. Rumor on the AMD side is that we might be seeing a newer FX chip sometime soon. I’m sure it’s only a matter of time.
AMD Athlon™ 64 3400+
Some of you might be wondering what the differences are between the Athlon64 3XXX series (like our 3400+) and the FX series (like the FX-51). The main architecture differences revolve around the integrated memory controller. The FX-51 uses a 128-bit wide bus to transfer data while our Athlon64 3400+ uses a 64-bit interface. Essentially the FX-51 has twice the memory bandwidth from information being sent to and from the integrated memory controller interface.
AMD Athlon™ Processors Technical Specifications
(click on the image for an enlarged version)
As you can see in the chart above, there are few other differences between the FX-51 and A64. The biggest physical difference is a major issue if your looking for a new machine, The FX-51 and A64 use different socket sizes. They are not pin compatible. Both CPUs take advantage of 1MB of L2 cache. They share the same die size and both support the same 3D instructions, including SSE2.
A new feature added to the 3XXX series is what AMD calls “Cool’ n ‘Quiet”. When your desktop CPU is up and running, it runs at full speed 100% of the time. It does not matter if you’re playing LOMAC or typing a review in Microsoft Word, the CPU is always throttled up to maximum. On mobile CPUs, the chips will scale up and down as per use. If you’re watching a DVD or typing in Microsoft Word, the CPU might only be running at 50% utilization. If you switch off the DVD and start a game, the CPU recognizes (instantaneously) that there are more demands being made and it throttles up to 100%.
AMD has taken those same features and implemented them on a desktop CPU. When your CPU needs to be at full speed, it will operate at (in this case) 2.20GHz. If your typing in Microsoft Word and your CPU has decided to throttle down (say to 1GHz) two things happen.
You’re not producing as much heat. Your system runs cooler and keeps lockups from occurring. Now your fan will also slow down to create less noise because less heat is being generated from a CPU that is operating at a lower CPU speed.
This is all good unless the CPU is deciding to throttle down on it’s own for other reasons. We will have to get with AMD and find out exactly what point the CPU determines that it’s “all clear” to throttle down. We certainly would not want the CPU to throttle down suddenly in the middle of a LOMAC mission!
