Introduction

AMD has been very popular with the enthusiast community for some time thanks to the success of the Athlon XP line of processors. These processors generally sold for below Intel's competing Pentium line yet allowed for equal or better performance at lower clock speeds. This difference in clock versus performance has even lead Intel to change its marketing of their latest line of processors. The Athlon XP 2500+ earned a reputation as a very strong overclocking processor. Its low cost to performance ratio made it very popular with those on a tight budget as well.

Today we take a look at another AMD processor that is becoming popular with the enthusiast crowd, but it's a processor that is geared towards the mobile market. Let's see how well the AMD Athlon XP-M 2600+ can perform compared to the previous XP 2500+.

XP vs. XP-M

Before going into the details of how the processor performs, let's take a look at what exactly the difference between the XP and XP-M processor. Both processors look identical and share the same pin layout. This allows both processors to run on the Socket A based motherboards that can support their speeds. In fact, each of the processors used in this review are based on the same Barton core from AMD. So what is the difference?

Close Up of the AMD Athlon XP-M 2600+

The difference comes in how the processors are handled post-production. The first difference is in their power usage. The XP-M processors are those that test to run properly at lower voltages (approximately 1.45 volts vs. 1.65 volts) so they consume less power. This is especially important for mobile systems that run on batteries.

Another important difference is in the clock speed and multiplier settings. The XP processors are defined by a specific bus speed and multiplier that sets its particular PR rating. In the case of the XP-M processors, the bus speed is typically fixed, but the multiplier must be able to float between a range of multipliers to help the system conserve power during low CPU usage. This combined with the wider range of voltage tolerances makes them perfect for overclocking.

Benchmarking

Since the XP-M 2600+ doesn't have a defined set of clock speeds in most desktop BIOS, all testing with the processor was done with overclocking. Manual clock and multiplier settings were done and compared to the overclocking results from my Athlon XP 2500+ Review. For additional information on overclocking, check out the What is Overclocking? article.

The hardware and software versions for the benchmarking is as follows:

To get a broader range of tests to show the difference between the processors, I've added some additional benchmarks. The same benchmarks of PCMark2002 and SiSoft Sandra MAX 2004 were used for the CPU and memory data as was used in the XP review. To see how the processor impacts the performance of graphics, both the Future Mark 3DMark2001SE and 3DMark03 DirectX benchmarks were used.

Since the XP-M is multiplier unlocked, testing was done at the same 2.2GHz speed as the overclocked XP 2500+ desktop processors that is equivalent to the standard settings of the XP 3200+ CPU. This translates into a 200MHz bus speed and 11 times multiplier. This was done at the standard voltage rates for a desktop processor of 1.65v. Additional overclocking was done to the limits available in the A7N8X BIOS of 2.5GHz or a 200x12.5 setting. Below is a chart of the relevant information:

In addition to the benchmarks that were carried out, Motherboard Monitor 5 was used to check the maximum temperatures at an idle state and at a full load state to see how hot the processors would run.

PCMark2002

FutureMark's PCMark2002 is an older general PC performance benchmark that has since been replaced by the PCMark2004 version, but since the XP 2500+ was tested with the 2002 version, it was also used for this review. The benchmark returns an aggregated score, but for the sake of this review, the specific numbers for the CPU and memory were taken. Below are the results:

From these scores, the XP-M 2600+ scores very close to the XP 2500+ at the same clock speed. It is interesting that the memory score was a bit higher, but still within the expected range. Increasing the clock speed to the 2.5Ghz speed gives a significant jump of 13% in the CPU score. It also provided a 4% boost in the memory speed even though the memory bus speed did not change.

Sandra MAX 2004

SiSoft's Sandra Max is a very well designed benchmark application that allows the specific components of a computer system to be tested. It also has a database of standard scores for general components for comparison. For the sake of this review, the CPU and memory scores were recorded and listed below:

From the Sandra testing, it is apparent that the scores a bit more accurate to the processors. The XP and XP-M score statistically equivalent for the CPU scores. Once again, the CPU scores show a 13% improvement in the scores for the XP-M by moving from the 2.2GHz speed to the 2.5 GHz speed, also a 13% increase. It is interesting to see that the XP-M seems to score a 6% higher memory score at the same clock speed as the XP 2500+.

3DMark2001SE & 3DMark03

Of course many people who overclock their processors are doing this to try and squeeze a few more frames out of their system for playing games. With this in mind, the FutureMark 3DMark2001SE and 3DMark03 tests were run. All testing was done at the default settings for the tests to show how the difference in CPUs translates into improved frame rates. Below are the results:

Based on the test results, the XP and XP-M processors running at the same 2.2Ghz clock show virtually identical scores. More surprising though is that the higher clocked 2.5GHz speed of the XP-M only generates a 7% speed increase for 3DMark2001SE, about half the clock speed performance increase, and only 1.4% for the 3DMark03. This shows that the tests are much more limited by the graphics card than the processor, especially for the DirectX9 related code.

Temperatures

Heat is always a problem for those looking to overclock a system. Higher temperatures can lead to system instability, so it is important to try and keep the CPU temperatures as low as possible. To see the differences between the processors and their speed settings, Motherboard Monitor 5 was used to record the temperatures from the thermal diode built into the processors. The highest temperature during idle computer usage and full CPU usage was then recorded.

The surprising result from this is that the XP-M 2600+ processor running at the same clock speed at the XP 2500+ processor is 14% cooler! This is very good news for those looking to really push the processor to its limits. Increasing the clock speed to 2.5GHz increases the heat by roughly 9% over the 2.2GHz speed, but they were still below the overclocked XP 2500+ at the 2.2GHz speed.

Conclusions

Overall, the Athlon XP-M processor definitely has the advantage over the standard desktop XP processors. The voltage optimizations allow the processor to run cooler than its desktop sibling and also to have greater potential to run at higher speeds. Performance was strong for the aging 32-bit platform.

There are a few drawbacks to using the XP-M processors though. First, since they are really designed for the mobile market, finding a retailer that carries the OEM XP-M parts is difficult. Supplies are often limited making them more difficult to find. Prices are also higher than equivalent rated desktop versions. Since they are OEM parts, they also don't include a heatsink and fan unit for cooling.

Another major issue is motherboard and BIOS compatibility. In the case of this review, the ASUS A7N8X Deluxe was only able to use a 12.5 multiplier. This restricts the maximum speed the processor could be clocked at even though it had the potential to go further. Because of this, researching a good motherboard combination that allows for the higher multipliers is essential to using this processor.

Most enthusiasts are comfortable with the restrictions and limitations to the processor in order to get the speed. Anyone willing to take such steps would be well served by the XP-M processors. Those looking to just put together a system without overclocking would be best to stick with the standard Athlon XP desktop models.

Manufacturer's Site