vacancies advertise contact news tip The Vault
facebook rss twitter

Review: AMD Athlon XP2200+ (.13 Micron)

by Ryszard Sommefeldt on 10 June 2002, 00:00


Quick Link:

Add to My Vault: x


AMD watchers have been waiting for this processor for what seems like forever and it was certainly an event for me when it finally turned up after waiting a day longer than I should have (thanks RoyalFail!). On the outside, looking at the name, it seems like just another processor to add to the Athlon XP line. It's the next model up from XP2100+ and adds the customary 66MHz of CPU clock speed over the outgoing flagship CPU but underneath the hood and as part of the bigger picture, it's a lot more significant.

Underneath the name it's quite a different beast from XP2100+ and represents a big change for AMD, namely their success in bringing a 0.13u process desktop processor to market. It's been quite a wait with it being expected for quite a while before now, but when shifting to an inherently more complex method of manufacturing a processor, some delays were almost inevitable.

Don't forget, this is AMD dipping their toes into the water of 0.13u CPU production. If you remember, Intel did the same with the 0.13u Tualatin Pentium III core before migrating Pentium 4 to the same process. XP2200+ serves as AMD's confirmation that they can execute at that process size and feel confident about moving future CPU production wholesale to 0.13u.

In AMD's case, this means a lot more than it does to Intel. AMD don't have the resources to tool up as many fabs for 0.13 production as Intel so switching over is a gamble for them. Re-tooling for a different process is a massive undertaking so it's quite understandable to see a delay or two from AMD in bringing XP2200+ to market.

So we know that this new processor uses a different process size than all previous Athlon XP processors, so what does this bring to the table? We know by looking at Intel's switch to 0.13 with Northwood that it brought a smaller die size, lower operating voltage and temperature, less power consumption and from the die shrink, more room on the wafer for extra's like HT logic and the much trumpeted 512kb of L2 cache.

Have AMD done the same? Well no, not quite to the same degree as Intel with Northwood. We have the die shrink (it's physically tiny compared to older AMD processors, including older Athlon XP's) and along with it a drop in operating voltage, temperature and power consumption (down from 1.75V Vcore to 1.65V) but there is no extra L2 cache on this processor.

Also, unlike Intel with Northwood 'B', this Athlon XP hasn't seen a front side bus increase. One of the current limiting factors with Athlon XP as speeds get higher and multiplier gets bigger is that it's memory bandwidth limited to around 2.1GB/sec between CPU and memory controller.

We've seen with Northwood 2.4B and 2.53B that given a memory technology to feed the CPU like DDR333 SDRAM or PC800/1066 RDRAM, the P4 enjoys a healthy performance boost at 533MHz FSB (133MHz QDR, up from 100MHz) making them currently the quickest desktop processors that money can buy.

So without a boost in L2 cache size and without a boost in front side bus and memory bandwidth between CPU and memory controller, making the CPU no more than an Athlon XP at 0.13u, can XP2200+ hope to compete with the new Northwood's and wrestle back the consumer processor performance crown? This is the major question that this review and many others will be trying to answer.

So I'll take a look at performance over a decent range of benchmarks using a number of different scenarios, talk about specific cases within those benchmarks and hopefully talk a bit more about AMD's current processor platform and what XP2200+ means for the future for AMD and you, the consumer. For the performance junkies among you, I'll also take a brief look at overclocking the new chip.

So onto the benchmarks where XP2200+ goes head to head with 2.4B, both using DDR333 memory on the latest supporting chipset for each platform.