Despite its age, however, S939 is still quite capable. It loses out to the latest that Intel (especially) and AMD have to offer when it comes to just raw computing power, but for my own personal purposes, it's still more than sufficient.
Think of it this way, perhaps: S939 is a Lamborghini Countach; in its heyday it wore the performance crown and forced its biggest rival to come up with something really special to best it. There is no doubt S939's performance advantage, and the subsequent inroads AMD made into Intel's dominant market share, woke the sleeping giant and angered it enough to want to retake the losses it had incurred during the Netburst era.
As surely as the Countach was supplanted, so has S939; it has been three long years (and counting) since it, and AMD, had a legitimate claim to being the most powerful and fastest computing platform available to the regular consumer, as well as the choice of performance enthusiasts. Intel's Core/Core 2 series, as well as the subsequent Nehalem generation of CPUs, have firmly put S939's halcyon days in the rearview mirror.
On the AMD side, S939 was first replaced with AM2. AM2 was basically just a refresh of S939's K8 roots. It was equivalent to the Lamborghini Diablo: Touted as a major advance, in reality it initially was merely a little bit more powerful, just a tad bit faster, than its immediate predecessor. As with the Diablo replacing the elder sister, the Countach, AM2 replacing S939 left me decidedly cold. I felt no desire to love the Diablo, and the same was true with AM2.
I guess you just can't ever forget your first love.
Compared to its immediate predecessor, AM2's differences boiled down to two salient features: 1) It now could use DDR2 (which had supplanted DDR in the marketplace); and 2) it slightly elevated the performance potential. AMD's move to terminate S939 was calculated to keep in step with developments in the memory sector; with their integrated memory controllers (IMC), AMD's chips had to be redesigned to be able to use DDR2, since 1st generation DDR was consigned to marketplace obsolescence. In most other respects, though, under the IHS, AM2 was still K8, with all its inherent strengths and limitations.
Indubitably, K8's biggest weakness was that it was impractical to cram four logical cores within one processor package. In other words, AM2 (and S939) were limited to a maximum of two cores per CPU due to power consumption and TDP concerns primarily. In a world where Intel was not only destroying AMD's best with just two cores, but with the prospect of having a quad-core processor coming to market imminently, AM2's status as an upgrade step up from S939 was looking less and less logical. Despite a gradual escalation in clock speeds, AM2 was simply not enough of a speed increase for me to justify starting over and buying new equipment (just to say I had the latest AMD kit).
Through constant development, though, AM2 gave rise to AM2+, signaling AMD's willingness to move on from its venerable K8 design genetics onto something far better. AM2+ opened the door to AMD's first quad-core CPUs, the Phenom X4, as well as the consumer market sector's first tri-core CPUs, the Phenom X3. There were also still AM2+ AthlonX2 dual-cores available for AM2+, but these were basically the ultimate refinements of the K8 microarchitecture.
The Phenom series, though, was quite different.
Although the first Phenom series for AM2+ was not all that great, with its fairly limited overclocking ability, it could be argued that it still was a significant release for AMD. If nothing else, it was the very first quad-core from AMD. Moreover, as often happens, once weaknesses and limitations are identified, they can be attacked and rectified.
And thus we have come to AM3, and the Phenom II series.
As an AMD fan, I was intrigued when I first read tales of superb overclocking coming from the Phenom II generation of CPUs. In a single stroke, it seemed as if the second generation of Phenom processors overcame the biggest reason why the original Phenoms did not earn the interest and admiration, much less the sales dollars, of performance computing enthusiasts. Tales of insane clock speeds (upwards of 6GHz), even if from cherry-picked chips with hyper-exotic cooling equipment, were enough to grab the collective attention of the enthusiast crowd. The fact that Phenom II quads could finally beat Intel's Core 2 quads (albeit NOT the yet-to-be-released Nehalem chips at the time) was also very exciting to AMD enthusiasts.
AM2/AM2+ definitely did NOT inspire any desire for me to upgrade from my S939 systems. But news of AM3 making waves certainly did make me consider upgrading sooner than my self-imposed target date of May 2010 (by which time my first, and still OCed and functional, original system will be four years old).
Four years is an eternity in the tech industry; I know that many OCNers would scream in abject horror that I've not yet moved on at all from S939. However, while I am an enthusiast, I have to be really selective when it comes to adopting new hardware. If there's one absolute law in this hobby, it's that it costs a lot of money to stay on top of the performance tree.
To be quite frank, I did consider going with a Core i7 for my next platform upgrade. Unquestionably it is the king of the performance mountain right now; not even AM3 can keep up with it, especially when we're talking about overclocked performance. But some factors dissuaded me from seriously pursuing this option.
Chief is money (sorry to disappoint you, but it wasn't AMD fanboyism ). Ultimate performance will always be the most expensive, and I don't have a lot of disposable income. Even in terms of value as I perceive it, Core i7 makes little sense to me.
A second factor is my actual needs for all that excess power. As I'd written earlier, even given its age S939 is still plenty adequate for my personal needs. I don't benchmark (totally useless activity, in my opinion); my systems can play any game I own (no, I don't own nor play Crysis, so don't go there); I don't do a ton of media creation/processing, but when I do my system doesn't slow to a crawl and beg for mercy. Perhaps my needs aren't typical of the average performance computing enthusiast (one of my favorite oxymoronic juxtapositions), but my needs are satisfied.
A third factor is a strong curiosity to learn about what AMD currently has to offer. AM2 OCing isn't much different from S939 (or any other K8) OCing, but AM3 Phenom II OCing looks decidedly different. As OCN's AMD Editor, I feel a responsibility to know as much about AMD OCing as possible. I'm not comfortable about doling out advice on something I don't personally have experience in; I don't feel good about merely regurgitating what someone else said or wrote without KNOWING for myself whether or not the information/opinion is correct or factual or accurate.
So yes, I'm going to be turning over a new leaf. I am amassing the core components of an AM3 system. I already have the CPU and the motherboard, as well as the HDDs, the PSU, Windows 7 pre-ordered, and the video card. I have a case, but it needs some modifications by thlnk3r before I put everything in. I also have all the watercooling parts I intend to use on this build (it will also be my first watercooled PC project). Basically, I need to amass a budget for some DDR3 RAM.
I'm so excited about this. Learning something new is so stimulating. Just as I was filled with enthusiasm when the Diablo was pushed aside by the Murcielago, so I am filled with fervor now that AM3 and Phenom II have arrived and are in my hands and are here to stay.
This is also my 100th OCN Blog Post. Thank you so much to all who have read, and hopefully still are reading, my words here. Many of you have engaged me and others through my blog and shared many provocative thoughts. I dedicate my first 100 blog posts to my regular readers; you know who you are.
I know the journey to the next 100 posts will be just as fun.