Is the Athlon 64 3500+ the chip to have for budget-conscious enthusiasts? Is its 3500+ performance rating well-deserved? Read on to find out.
A closer look at the 3500+
Before we dive into the benchmarks, let's have a look at the 3500+ up close.
AMD's current Athlon 64 lineup is made up of chips with different clock speeds, memory interfaces, and L2 cache sizes, making assigning arbitrary performance ratings a dubious task. With a little help from our handy CPU Decoder Ring, here's a cheat sheet that sums up where the 3500+'s clock speed, memory interface, and L2 cache size fit in with the rest of the Athlon 64 family.
|Socket||Memory interface||L2 cache||1.8GHz||2.0GHz||2.2GHz||2.4GHz|
As you can see, the Athlon 64 3500+ actually shares the same 2.2GHz clock speed as the 3400+. So why the higher performance rating for the 3500+? Because the 3400+'s on-die memory controller can only address a single channel of DDR memory. The 3500+'s memory controller is a dual-channel design that offers twice the peak theoretical memory bandwidth of the 3400+, provided you use two DIMMs. So with clock speeds being equal, the 3500+'s extra memory channel earns it a higher performance rating. Simple, right?
Not so fast. The Athlon 64 3500+ and 3400+ also differ in L2 cache size, and when it comes to cache, the 3400+ is better-equipped. The 3400+ is based on an older Clawhammer Athlon 64 core that AMD endowed with 1MB of L2 cache, while the 3500+ is based on a more recent Newcastle core with only 512KB of L2. The 3500+ may have twice the memory channels of the 3400+, but it only has half as much L2 cache.
Based on their Athlon 64 performance ratings, AMD apparently feels that the 3500+'s dual-channel memory controller can more than compensate for the chip's smaller L2 cache. They may be right, but it's hard to believe that the 3400+'s larger cache won't give it the upper hand in at least a couple of tests. We have plenty of benchmark results to sift through to find out.
|The TR Podcast is live, so come ask us stuff!||0|
|AMD shows off DirectX 12 performance with new 3DMark benchmark||20|
|Intel and Micron sampling 3D NAND based on floating gates||13|
|Report: Microsoft to build an Intel-powered, non-Pro Surface tablet||45|
|Toshiba's 3D flash spreads 16GB over 48 layers||6|
|Cougar's 300M gaming mouse looks awfully familiar||14|
|Wednesday Night Shortbread||69|
|EVGA Hybrid liquid-cools the GeForce GTX 980||20|
|Rowhammer attack exploits shrinking process size in DRAM||40|