AMD released the Bulldozer architecture processors late last year. In this review we will take a look at the AMD FX-8120 processor, which is the slowest octacores AMD has, but the differences between it and faster models are only in the operating frequency. Let’s see how well it performs and if it can hold the line against Intel Core I5 processors.
Check their webpages at www.amd.com for information about the company.
- Family: AMD FX-Series
- Model number: FX-8120
- Frequency: 3100MHz
- Turbo Frequency: 3400MHz (8cores), 4000MHz (4cores)
- Socket: Socket AM3+
- Microarchitecture: Bulldozer
- Platform: Scorpius
- Processor core: Zambezi
- Stepping: B2
- Manufacturing process: 0.032 micron
- Number of Cores/Threads: 8C/8T
- Level 1 cache: 4x64KB + 8x16KB
- Level 2 cache: 4x2MB
- Level 3 cache: 8MB
- Power Consumption: 125W
- AES Instructions
- Advanced Bit Manipulation
- AMD64 Technologi
- Advanced Vector extensions
In this review we will take a look at the FX-8120 processor from AMD. FX-8120 is based on AMD:s newest Bulldozer architecture and has eight cores. Unlike traditional multi-core cpu:s, the Bulldozer architecture implements the cores a bit differently. Each Bulldozer module has two cores; and therefore FX-8120 has four modules. Each module has its dedicated cache memory and each core in a module shares some resources of the module. So unlike traditional multicore processors, Bulldozer processors do not have dedicated resources for each core. Basically a two core bulldozer module can be seen as one core which can do two threads at the same time plus it has an extra L1-cached integer module. So in fact eight core FX-processor isn’t a true eight core but rather four cores + four integer modules.
The sample unit we have is a FX-8120 processor, which has 4x64kb L1 cache (each module has its own) and 8x16kb extra L1 cache (each core has its own). Added to that there is a total of 8MB of L2 cache and the same amount of L3 cache. Now let’s see how well our unit performs in tests!
Pages: 1 2