| 
      
      
    
    
    AMD Athlon XP 2200+ Overview 
    
    
     With 
    the recent extreme frequency increase Pentium 
    
    4 
    processors have shown, AMD had to reply to Intel by unveiling processors 
    capable of matching the latest Pentium 
    
    4 
    processors. To do so, AMD was forced to revamp their CPU architecture since 
    the ‘old’ Palomino Athlon core, engraved in 
    
    0,18µ, 
    wasn’t really able to support frequencies higher than 1.73 GHz (AMD Athlon 
    XP 
    
    2100+). 
      
    
    That’s 
    why AMD’s engineers have designed a brand new CPU core, known as 
    Thoroughbred, that is engraved in 
    
    0,13µ, 
    a real premiere for AMD processors. The main advantage of this new engraving 
    process is, in theory, the fact the CPU will emit less heat than before thus 
    allowing AMD to crank up frequencies in the future while power users will be 
    able to overclock their processors in all peace of mind. This is true for 
    low range Athlon XP processors with smaller clock frequencies: with the 
    Thoroughbred core an Athlon XP 
    
    1800+ 
    will produce 
    
    20% 
    less heat than an Athlon XP 
    
    1800+ 
    using the Palomino core. Unfortunately since the Athlon XP 
    
    2200+ 
    requires 1.65 Volts of power, the heat generated is still quite considerable 
    (around 
    
    52° 
    C [that is to say 125.60 fahrenheit] at full charge) putting the processor 
    in a hot spot. To have an idea the Athlon XP 
    
    2200+ 
    T-Bred heat dissipation is only 
    
    4.1 
    Watts less than the Athlon XP 
    
    2100+ 
    Palomino. The Athlon XP 
    
    2200+ 
    includes a built-in thermal diode so software can accurately monitor the 
    temperature of the CPU. 
    
      
    AMD Thorougbred Die 
    
    In 
    reality, the main advantage of the 
    
    0,13µ 
    technology lies somewhere else: using this manufacturing process, the 
    processor’ die size is reduced from 
    
    128 
    to 
    
    80mm² 
    thus AMD could produce more CPUs per 
    
    200mm 
    wafer (~322 
    against 
    
    201 
    previously) which implies a noticeable cost reduction both for AMD and the 
    end user. The reduced size of the processor’s die generates a casualty: 
    being 
    
    40% 
    smaller the die emits more heat that cannot be dissipated very well since 
    the surface in contact with the heatsink is also smaller. That’s why AMD now 
    only recommends high end CPU cooler with a copper plate that makes contact 
    with the surface of the CPU.  
    
      
    AMD Athlon XP 2200+ CPU 
    
    Despite 
    its name, the AMD Athlon XP 
    
    2200+ 
    runs in reality at 1.8 GHz. Yes AMD has still not jumped over the 2.0 GHz 
    barrier… The name of the Athlon XP 
    
    2200+ 
    is the fruit of the Performance Rating, a marketing trick that consists of 
    naming the processor on the performance it equals rather than its real 
    frequency. Personally I think this naming convention will most likely 
    swindle consumers by letting them think the AMD Athlon XP 
    
    2200+ 
    is clocked at 
    
    2.2GHz 
    while it’s not the case.   
    
    Apart 
    from the finer engraving process and the frequency increase; this Athlon 
    ‘T-bred’ built on the K7 architecture has nothing new in terms of 32-bit 
    instructions and still hosts 
    
    37.5 
    million transistors. The AMD Athlon XP Thoroughbred features a 266 MHz front 
    side bus which compares to the Pentium 4 533 MHz front side bus. This 266 
    MHz only front side bus is kind of disappointing when you consider a 333 MHz 
    front side bus would have fully exploited the recent DDR333 
    memory. 
    
    
    Physically the processor still uses the old good Socket A and doesn’t differ 
    from previous Athlon XP except for the presence of more bridges. The core 
    also looks rectangular while it was looking more like a square before. The 
    Socket A architecture means you’ll be able to use the processor with every 
    Athlon XP Slot A compatible motherboard (most motherboards will require a 
    BIOS update). 
    
    CPU Architecture 
    
       
    The Athlon XP 
    
    2200+ 
    comes with 
    
    128 
    KB of L1 
    cache memory and 
    
    256 
    KB of full speed L2 
    cache memory. The L1 
    Data translation look aside buffer (TLB) can contain 
    
    40 
    entries that can be written speculatively. The regrettable thing is that the 
    Athlon XP ‘Thoroughbred’ doesn’t offer as much cache memory as its 
    equivalent the Pentium 
    
    
    4 
    ‘Northwood’ which comes with 
    
    512 
    KB of embedded cache memory. The x86 
    compliant architecture of the Athlon XP 
    
    2200+ 
    includes a full implementation of Intel SSE instructions set. AMD markets 
    this feature as ‘3D 
    now! Professional’. It enables SSE compatible applications to recognize the 
    Athlon XP as a SSE-capable processor which will provide a significant 
    performance boost.  
    
    Overclocking 
    
       
    This is probably the most surprising thing for an AMD processor, but the 
    Athlon XP 
    
    2200+ 
    isn’t really overclockable at least in major proportions. Natively the CPU 
    uses a 
    
    13.5 
    clock multiplier (13.5*133=1800 
    MHz). During our extensive testing we weren’t able to safely push the CPU 
    over 1890 MHz (with 
    
    13.5*140) 
    which represents a 
    
    5% 
    only overclocking! 
    
      
    AMD Athlon XP 2200+ 0.13µ 
    
      
     |