Of course, that's a rhetorical question. Sadly, I know I'm not alone when it comes to the pile of work that you come back to after taking a vacation. You know. The one where, it's so big that it's depressing and you're not sure where to start? That pile also accentuates my love/hate relationship with RSS. I do love it and quite frankly see it as potentially being the solution to a lot of other problems out there (like spam). But man o' man. Go away for five days and what sort of RSS reader do you come back to (in addition to a forklift-sized load of e-mail)? I feel like sending a note to all the people whose feeds I monitor to see if we can establish an official blogger vacation week. That way, we all come back to empty RSS folders.
So, after digging through a forklift-load of e-mail, I turned my attention to an overly swollen RSS reader (I'm using the RSS subscription/reading capabilities in Mozilla's Thunderbird 2.0 (Release Candidate 1 is out), I thought I'd rant on a few historical items as a way of jumping back into the Niagara Falls (my favorite metaphor for the blogosphere).
Speaking of Niagara (the codename used by Sun for its latest generation of power-saving silicon), Sun may have some challenges moving Niagara gear off the shelves if any of its prospective customers give Don MacAskill's blog a read. For Sun though, there's still a silver lining. MacAskill, the chief exec at SmugMug, wrote so glowingly of Sun's Opteron-based X2200 servers that I feel like I need to buy one just for looks:
It’s been two months since we divorced Rackable and married Sun as our new server & storage vendor and lots of people have been asking how it’s going. So while the 'marriage' is still early the server side of things is going really really well. We’re still starry-eyed in love.
MacAskill goes onto to extol the virtues of deploying the X2200 as one of the driving forces in the SmugMug infrastructure but, in saying "our experience with Sun’s storage hardware isn’t nearly so rosy (in fact, it’s downright bad)," drops a hint that someone with a "C" in their title at Sun should be calling him right about now. MacAskill probably knows a little bit about what he's talking about given that eBay and HotMail were originally hosted on a network that he built. The rest of his incredibly balanced review of the X2200 (full of recommendations to Sun) makes us wish he was reviewing gear for ZDNet when he's not checking for defective RAM.
Elsewhere in server news, IBM's server customers will be able to have their cake and eat it too now that Big Blue is offering low power servers in a choice of flavors: AMD or Intel. Wrote CNET News.com's king of the server beat Stephen Shankland:
Big Blue has put Intel's new 1.6GHz and 1.86GHz 50-watt quad-core Xeon 5300 "Clovertown" processor in two rack-mounted System x servers, the 3.5-inch thick x3650 and 1.75-inch thick x3550, and in the HS21 blade server. In addition, its two-processor LS21 and four-processor LS41 blade servers now are available with new dual-core AMD Opteron chips, a 2.6GHz HE (high-efficiency) model that consumes 65 watts and a 1.8GHz EE (extreme-efficiency) model that consumes 40 watts, said Scott Tease, marketing manager for IBM's BladeCenter products.
In deference to AMD and Intel, both of whom now talk green (now that the green theme is the in thing to talk about), a chip's power consumption definitely matters. But Shankland rightfully doesn't venture much past chips specs here given the way a power consumption advantage could easily be ameliorated by some other power hungry component that's in those systems with those chips. Robert Mullins coverage of the same "low-carb" news (the idea being less energy consumption equates to less carbon emissions) cuts to that chase pretty quickly:
IBM also designed the new BladeCenter servers without internal disk drives, which use 10 to 12 watts. Instead, IBM is using a 4G-byte modular flash drive that it claims uses 95 percent less power than a spinning disk drive. The flash drive can be used as a Linux operating system boot drive and as a storage device to complement shared storage on the IT network."
Once you have the equivalent of an EnergyStar sticker on the outside of the system (unfortunately, we don't), there still remains the question of what sort of field that system can plow, and how long it will take. Much the same way chip wattage is a tough indicator of a system's overall power consumption, chip frequency (the "Ghz") is a terrible indicator of overall system performance. Again, other components come into play. Bottom line: When it comes to benchmarking systems for their performance or power consumption in the context of cost, there's a huge grey area revolving around system componentry that makes the story of a chip more difficult to tell, and easier to abuse (if your a chipmaker). That's why the industry has the current benchmarking mess on its hands that it does, sometimes forcing me to read the riot act to ZDNet's own bloggers about objectivity when writing about benchmarks and chips.
It's a riot act I won't be reading to The Tech Report where Scott Wasson, in his 64-bit benchmark comparison of the current generation of Intel and AMD destkop-targeted chips, does a masterful job of not only showing both Intel and AMD how to line one company's offerings up against another, it also exposes the soft-white underbelly when it comes to determining what he rightfully calls "value:"
.....These things are never entirely simple, though, so we should roll out some caveats. One of the big ones involves those FX-series processors. You'll need two of them in order to populate a Quad FX motherboard, so they're priced (and listed above) in pairs. However, there's currently only one Quad FX motherboard available, and it costs about $350, which throws the value equation out of whack.
The value equation sometimes goes off-kilter the other way when AMD employs guerrilla price-war tactics like selling the Athlon 64 X2 4600+ for $125.99 on Newegg, well below the slower 4400+. AMD has several of these "Crazy Hector" deals going at Newegg right now, and none of them seem to involve the Athlon 64 models we've identified as direct competitors to specific Core 2 Duo models. That's probably an intentional facet of AMD's strategy. This practice throws a wrench in our nice, neat comparsion, but there's little we can do other than tell you about it.....
OK. So no comparison is perfect. But what's better? Disclosing that difficulty up front as The Tech Report does? Or manipulating it? Oh, by the way, Wasson concludes that Intel has maintained "a performance edge at most price points, despite standing pat through AMD's aggressive pricing moves and new model introductions." Advantage (in the 64-bit desktop performance market): Intel.
Back to servers.
Not content to let IBM get away with being the Baskin-Robbins of servers (including theoretically lower energy servers), Dell has apparently tossed a little more Opteron into its mix as well. eWeek's Scott Ferguson reports:
Staring April 10, the Round Rock, Texas, PC maker will add several additional Opteron models to a new server offering called the PowerEdge 2970. In addition, Dell plans to add Opteron processors to a second new system, called the "Energy Smart" 2970 server, later this year.
"Energy Smart?" Sounds dangerously close to the aforementioned "Energy Star" (it's only a matter of time). Dell is also apparently tee'd up to dally with AMD's forthcoming quad core Opterons. Wrote Ferguson:
Besides dual-core processors, Dell executives said the two systems have been designed to allow customers to upgrade to AMD's quad-core Opteron processor—"Barcelona"— when the chip maker unveils these models later in 2007.
With some of the top system makers clearly jazzed about AMD's offerings, perhaps the gloom and doom surround the chipmaker (see ExtremeTech's How can AMD be fixed?) is a bit premature. Sure, AMD is having some short-term difficulties that may very well have been exacerbated by Intel's felonious manipulation of benchmark data (sadly, after drawing our attention to Intel's misdeeds, AMD did virtually the same thing). But as long as system manufacturers continue to build product around a chipmaker's offerings, I'm not so sure that a short-term hiccup is a harbinger of longer-term things to come. Behind closed doors, AMD has been pretty cocky about what it has in its pipeline. Not that I'm just going to roll over and take AMD's word for it. But the last time I blew AMD's cockiness off, I ended up eating crow. So, despite any need for additional foul in my diet, please excuse my reluctance.
One thing I believe to be true (in turns of How AMD can be fixed?). Building Digital Rights Management (DRM) into its chips (see Adrian Kingsley-Hughes' How AMD hopes to turn things around) probably isn't the answer. Not only does move by AMD appears to comes at a time when the entertainment cartel is showing signs of backing away from DRM schemes (eg: EMI's release of DRM-free music on the iTunes Music Store), but also at a time when the complexity of DRM-laden technologies like CableCard are failing in practice (see how CNET's Rich Brown struck out in his attempt to get a supposedly CableCard-compatible PC to tune into a "CableCard signal").
OK. So you've had enough about servers, AMD, Intel, and benchmarking for one post. About the only thing you need to figure out now is what to run on them. Windows? Linux? Or what about OpenSolaris? The Register's David Norfolk recently pondered this very question.