With all of the excitement surrounding last week's launch of the Thoroughbred core, who wants to talk about 0.18-micron CPUs anymore? Unfortunately for those who follow the server market, they are generally not the first to receive the brand new CPUs. Instead AMD and Intel like to focus their new architectures and technologies on the mobile and desktop fronts first and eventually transition them to the server world when it makes sense.

The reason for this is simple; in the server world the stakes are much higher and thus the processors must be rigorously tested (or validated) under much more strenuous conditions. The processors must also be validated in multiprocessor configurations, taking up even more time. It's for this reason that there's a noticeable lag between the release of a desktop Athlon XP processor and its identically clocked MP counterpart.

In the past AMD committed to a 1-month differential between the launch of an XP part and its MP counterpart, but as you can already guess it has been more than a month since the release of the Athlon XP 2100+. For whatever reason it has taken three months for the Athlon MP 2100+ to become available, and you can guess that the performance boost offered by the 66MHz clock speed increase isn't really noticeable. To top things off, this is still a 0.18-micron part and thus ends up being AMD's hottest running server CPU that's currently available.

Although there's not much to get excited about the new MP release, we decided to take this opportunity to update our database server comparison benchmarks with the fastest CPUs from both camps.

The Chips

The Athlon MP 2100+ looks identical to the Athlon XP 2100+ we reviewed back in March. Our samples did not have the new green-dyed processor substrate but eventually all Socket-A CPUs will be colored green.

Click to Enlarge

The only physical difference between the Athlon MP 2100+ and the Athlon XP 2100+ is that the L1 bridges are not laser cut which makes unlocking the chip much easier. As you can tell from our Price Guides, the Athlon MP processors do sell for almost a $100 premium over their XP counterparts. You're paying extra for the added validation and testing that goes into every Athlon MP processor that ships; whether that is worth the $100 premium is another discussion entirely.

From the Intel camp we have the Xeon running at 2.4GHz. The 2.4GHz Xeon is a 0.13-micron part that is unfortunately restricted to the 400MHz FSB since there are no 533MHz FSB platforms that are officially available for the Xeon just yet. The Xeon core is Northwood derived, with a 512KB on-die L2 cache that helps tremendously in transactional server situations. For more information on the Xeon architecture you can read our original review here.

Click to Enlarge

Intel has also introduced the Xeon MP processor with support for greater than 2-way MP configurations. The Xeon MP processors only have a 256KB on-die L2 cache but they also have an on-die L3 cache. The on-die L3 cache is available in 512KB and 1MB sizes; these processors weren't available to us at the time of testing so we weren't able to include them in the review. We will try to do a piece on them in the future however as the effects of the on-die L3 cache should be interesting, especially in the world of transactional database servers.

The Test

We've already looked at the single processor performance of the Athlon XP 2100+ as a desktop/workstation solution in our review of the processor. Since the Athlon MP 2100+ performs identically there was no reason to revisit all of those benchmarks. We limited this comparison only to our database server tests that we've been using for over a year now.

Here's a description of the three test databases we ran trace scripts on:

As we've mentioned before, there are three databases that power AnandTech: the Web DB, the Ad DB and the Forums DB; we used all three databases to test these server CPUs.

The Web DB is where all of our content is stored; everything from news and reviews to our own internal article rankings are stored in this database. By far the majority of the transactions on this database are selects (reads). Remember that the web site only really offers one way interaction, the readers come to the site and read articles which are contained in this database. The articles are selected from the database and fed to one of the 4 web servers for assembly into a page for your browser. Internally, some update queries are also run, but they were not recorded in the test trace we ran. This database is the smallest out of the three; the DB was only 300MB when we ran the test.

The Ad DB is very similar to the web database in that quite a few selects are running. The select queries are used to pull the ads from the database for display in the user's browser. There are also a number of stored procedures that run along with the selects, but to keep things as simple as possible (at least for this comparison), we omitted them from the test trace. The Ad DB is noticeably larger than the web database, at a large 2.1GBs at the time of publication.

The final database is the Forums DB, which is by far the most transaction intensive database in the AnandTech Network. While the vast majority of the requests to the DB are in the form of selects (users reading categories and threads), there are significantly more inserts and updates (posting, thread/post counts, etc…) than in either of the other DBs. This database is also our largest, weighing in at just under 3GB during the testing and close to 8GB today (we used an older version of the DB from a few months ago).

In the past, when we used database server testing, it was done using a single trace run on the AnandTech Forums. While we're using two additional databases, the test methodology remains the same. We recorded a trace of transactions on each one of these databases for a set period of time. These were live recordings while the website and forums were being accessed just like they would on any normal day. The traces were then played back at full speed (as fast as the server test bed could replay them) and their playback times recorded. We divided the number of transactions replayed by the playback time and reported all scores in numbers of database transactions per second: the higher the better.

For our test bed, we setup a simple RAID 0 array (which would never fly in a real-world situation, no redundancy spells certain death) of two Seagate Cheetah X15 drives. While this gives us less I/O power than is in our current Forums DB made of 6 Quantum Atlas 10K II drives, it was enough for this limited test. In the future, we'll most likely move to a 4 - 8 drive test array as the benchmarks become more strenuous.

The test beds were configured with 2GB of DDR266 SDRAM using two 1GB sticks. Since the E7500 chipset requires memory to be installed in pairs we had to use multiple DIMMs. Also remember that the dual-channel memory controller of the E7500 chipset only ran the memory bus at DDR200 speeds.

Windows 2000 Professional SP2/SQL 2000 Test Bed
Hardware Configuration
CPU	AMD Athlon MP 2100+ (1.73GHz) AMD Athlon MP 2000+ (1.67GHz) AMD Athlon MP 1.2GHz Intel Xeon 2.4GHz Intel Xeon 2.2GHz Intel Xeon 1.7GHz
Motherboard	Tyan Thunder MPX - AMD 760MPX Chipset Tyan Thunder i7500 - Intel E7500 Chipset
RAM	2 x 1024MB DDR266 CAS2.5 DIMMs
Sound	None
Hard Drive	80GB Maxtor D740X - Boot Drive 2 - 18GB Seagate Cheetah X15 - RAID drives
Video Cards (Drivers)	NVIDIA GeForce4 Ti 4600 (28.32)

AnandTech Website Database Performance

When we last used this test the dual 2.2GHz Xeon was on par with the dual Athlon MP 2000+, with Hyper Threading giving it a slight increase in the number of transactions per second that the server could handle.

Final Words - The Future of the Athlon MP?

As we've mentioned before, there's not much to get you excited when talking about the Athlon MP 2100+; it's a 66MHz clock speed increase that we've had on the desktop side for three months now. The performance boost that you get in dual processor environments over the "old" 2000+ is around 4% in some of the best case scenarios, which were illustrated by our database server tests. It's not the Athlon MP 2100+ that we should be focusing on however, it is AMD's strategy going into 2003 that requires our attention.

You've probably heard by now that AMD's Barton core will be released according to AMD's current roadmap. The Barton core will not employ Silicon on Insulator technology; it will simply use the current 0.13-micron process that was introduced with Thoroughbred. The only modification going into Barton will be the use of a 512KB L2 cache which is desperately needed by the current Athlon. Because of the exclusive L2 cache architecture employed by the K7 core and the very large L1 cache, Barton will have more data cache available than the Xeon, despite the fact that both processors will have a 512KB L2 cache.

By the end of this year Barton will be introduced as an Athlon XP processor and in the following months you will see the core transition to the MP line as well. The Barton MP core will be the entry-level workstation and server CPU of choice as it will be much cheaper to produce than the Opterons that will be available.

This brings us to discussion about the Opteron; you've undoubtedly heard talk about ClawHammer (the desktop Athlon based on the Hammer architecture) being delayed until Q1-2003. AMD's official stance (as of today) on this is that the ClawHammer release schedule will be occurring in accordance with their publicly available roadmap. If you take a look at that roadmap you'll see that ClawHammer is listed as being introduced anywhere between the second half of this year and the beginning of next year. We wouldn't rule out a late 2002 release for Hammer but the chances of a early 2003 release are seemingly much more likely now.

According to AMD there are a number of factors that will govern the release time period of the processor; among those mentioned were - the condition of the CPU and yields, chipset and motherboard readiness, OS support and the availability of compelling 64-bit applications to showcase the technology. A combination of all of those factors will determine when ClawHammer hits and also when Opteron is released on the server side.

Until then, the Athlon MP is clearly still quite competitive in the server market but Intel will be able to claim an even more significant performance lead with the introduction of a 533MHz FSB Xeon part. The large-cache Xeon MP processors are also running without competition from AMD at this point and they will continue to do so until sometime in the first half of next year when we expect to see Opterons with 1MB L2 caches. In fact, until 2003 we won't see any serious action from AMD in the server market with the exception of Thoroughbred MP. It seems to be taking all of the kings men to lift that Hammer...