AMD's Athlon XP 2200+ Processor

AMD's Athlon XP 2200+ Processor - Page 2


AMD's Athlon XP 2200+ Processor
The "Thoroughbreds" at .13 Debut!

By, Marco Chiappetta
June 10, 2002

We tested our AMD Athlon XP 2200+ using a VIA KT333 based motherboard from Gigabyte, the GA-7VRX, with 512MB (2x256MB) of TwinMos PC2700 DDR memory.  The rest of the hardware used in our test system is outlined on the next page.  For now let's take quick look at the motherboard and memory.

The Supporting Cast Members
What's Under the Hood?

The Gigabyte GA-7VRX is a feature rich board powered by the VIA KT333 chipset.  The GA-7VRX has an on-board ATA/133 capable Promise RAID controller, Creative Labs on-board sound, an on-board NIC and USB 2.0.  This board also proved to be very stable, only exhibiting instability when overclocked well beyond specifications (although not as high as we would have liked).  Also visible in the above picture, is the Taisol CPU cooler we used with the Athlon XP 2200+.  Due to the smaller surface area of the core, there is less contact area between the CPU and heatsink.  To help insure proper cooling, AMD will be recommending better coolers with higher contact pressure for use with the "Thoroughbred" based Athlons.

The two TwinMos 256MB sticks of PC2700 DDR RAM we used, were populated by eight 6ns WinBond chips, rated to run at 2.5-3-3.  Our particular modules ran at CAS 2 without a problem though, so that's where we left them set throughout testing.


Overclocking The 2200+
An AMD .13 core...Very Interesting!

I'm sure all of you overclockers out there will no doubt be familiar with what's visible in the above picture.  The Athlon XP 2200+ we tested came with all of the L1 bridges left intact.  Like all of the older socketed Athlons, having connected L1 bridges should unlock the processor and allow for altering of the multiplier.  Unfortunately, the Gigabyte motherboard we used for testing our CPU would not post if we made any adjustments to the multiplier.  We then set out to overclock our processor by raising the FSB, but were only able to hit a maximum of 142MHz at default voltage, which equated to a clock speed of 1917MHz.  We tried raising the core voltage and going higher, but our CPU would not go any higher and remain stable.  The highest speed we were able to boot into Windows was 1984MHz (13.5x147), but running any program crashed the system.  We tried to hit 2GHz, but the system would not POST, sorry folks!  Frankly, we were expecting more out of this CPU but suspect that because it was an early sample, and the Taisol cooler we used wasn't exactly a high-performance model, there wasn't much headroom.

While overclocking, we paid close attention to CPU temperatures and noticed that the "Thoroughbred" did not run much cooler than a similarly clocked "Palomino".  As an experiment, we installed an Athlon XP 2100+ onto the Gigabyte motherboard we used, and raised the FSB to 138MHz to bring the clockspeed up to 1800MHz.  Using the same Taisol cooler, we saw temperatures ranging from 47 - 55 degrees Celsius on the "Palomino" based 2100+.  With the "Thoroughbred' based 2200+, at it's default clock speed, we saw temperatures ranging from 44 - 53 degrees Celsius, a difference of approximately 7%. We were hoping to see much lower temperatures than this but with the much smaller contact area, dissipating heat efficiently is going to be the challenge.  It's going to be interesting to see what type of cooler AMD ultimately ships with their boxed processors.  Perhaps a nice all Copper model will absorb heat from the core a little better?

The Athlon XP 2200+ Exposed
What Makes This Baby Tick?

Using WCPUID, you can see the AMD Athlon 2200+ is running at an actual clock speed of 1800MHz (1807MHz in our screenshot due to the aggressive timings of the GA-7VRX motherboard).  This clock speed is attained using a multiplier of 13.5 and a front side bus of 133Mhz. (13x133=1800).  We also took a snapshot of the CacheID information and the Standard and Enhanced feature flags.  Athlon XPs are equipped with 64K of 2-Way set associative Instruction L1 cache, 64K of 2-Way set associative data L1 cache and 256K of full speed, 16-Way set associative L2 cache, for a grand total of 384K of effective on chip cache.  Later in the year AMD will be introducing the "Barton" core, which will increase on-die cache to 512K, which should give them a nice boost in performance.

On to the Benchmarks!

Related content