Kevin Houston recently asked which Intel Xeon 5500-series processor was most popular for running VMWare. My short answer: Probably the 2.26GHz Xeon E5520. Long answer follows...
Intel makes more than a dozen different speeds of the Xeon 5500 series processor. Why so many? One reason is that Intel divides them up into three different groups, based on their features. Within each group, there's 3 or 4 different models, each with a different CPU clock frequency.
The Advanced group of processors include full-speed links & features; the Standard have the features but slower links, and the Basic group lacks Hyper-Threading & Turbo Boost:
|Category||QPI throughput||Memory speed||Hyper-Threading||Max Turbo Boost|
|Advanced||6.40 GT/s||1333 MT/s||Yes||+400 MHz|
|Standard||5.86 GT/s||1066 MT/s||Yes||+266 MHz|
|Basic||4.80 GT/s||800 MT/s||No||None|
So which ones are popular for virtualization? I found a good proxy for estimating this: searching for how often each model is mentioned in the VMWare community forums. There, four models rise to the top: The E5504, the E5520, the X5550, and the X5570. The E5520 gets the winner's ribbon for most-mentioned.
Why these four? I believe 3 of the 4 are popular because they're the best price/performance within each of Intel's groups. The last get attention because it's the fastest generally-available Xeon 5500 processor.
To show why the price/performance stands out for the E5504 (a "Basic" processor), E5520 (a "Standard"), and X5550 (an "Advanced"), compare Intel's "1ku" list price (basically their bulk-rate prices) to a processor-oriented benchmark result. I've picked the SPECcpu2006 integer rate benchmark for two reasons: One, it's an established benchmark with lots of published results, and two, the benchmark it's so processor-focused that you usually get the same results no matter what server -- or server vendor -- you use.
I created a Yahoo Pipe that scrapes benchmark results from SPEC.org and processor list prices from Intel, then plots them using a Google chart. Just for fun, I've listed the latest 2-processor AMD Opteron results too. Here's a link to the pipe itself (it's actually a series of pipes.) Although the Yahoo pipe is dynamic, on this blog I'm just going to post a static picture of the chart that it produces:
Each processor model is plotted by list price (vertical axis) and cpu2006_rate (horizontal axis). The further right on the chart, the more performance; the further up, the more expensive.
You'll notice that the results fall into three vertical bands, with the dual-core E5502 as an outlier. Within each band, the SPECint results don't change much, but the price does:
That means within these bands, you might as well pick the cheapest processor, because you don't get much extra horsepower by moving to a pricier CPU. What are the cheapest within each band? It's those 3 models that get all the discussion on the VMWare forums.
As I said, the fourth popular model is the X5570, which is the fastest "normal" Xeon 5500 series processor. (The two W-model processors are actually faster, but since they're geared toward workstations and not servers, they're not as commonly available.) So what makes the E5520 the apparent king of the popularity contest? Probably because it's the cheapest Xeon 5500 processor that's got Hyper-Threading.
So how does this VMWare Communities "proxy" compare to what processors HP ships on its ProLiant servers? Well, it's pretty close. Across all ProLiant servers, six models stand out: the four above, plus the E5540 (a "Standard") and the L5520 (a special Low-Power "Standard").
If you just consider ProLiant server blades, the E5504 drops off that short-list, while the L5520 gets a lot of use, and the X5570 gets heavier representation than my "VMWare communities proxy" would suggest. This makes sense to me, since folks deploying blades tend to be more interested in power-reducing and maximum-performance features.
It doesn’t happen often, but here is one of those situations in life where that dismal science, economics, can be useful and fun.
First a little definition ….
Price transparency is defined as a situation in which both buyer and seller know what products, services or capital assets are available and at what price.
Now, price transparency is a way of life in the business of standards-based, x86 servers. This even includes blade servers.
Go to any of the major vendors’ web sites such as Dell.com, IBM.com, HP.com or even resellers such as CDW.com and you can freely look up at least what each vendor’s list price is. You can bet that we vendors do this all the time to ensure we are ‘competitive’ with each other. And so do all of our customers. In fact, our customers know our list price and our competitors’ before any of us set foot in a customer’s place. That helps keep us vendors on our toes to deliver better products while keeping prices lower for customers. The x86 market is a living example of price transparency.
Until you get to the newest member of the x86 server community …
Cisco claims on their web site that their recently announced Unified Computing System is “Reducing total cost of ownership at the platform, site, and organizational levels”. The glaring omission is “minus the cost”. One would presume this includes a competitively priced set of compute, storage, enclosure, interconnect, management tool, software licenses and support components. But aside from Cisco’s word for it, this cannot be verified.
This is because Cisco, unlike the other server vendors, does not publish their UCS list price on their web site. Nor do their resellers seem to. This makes it difficult for customers (or competitors) to independently validate features to prices in a standards-based industry.
Given Cisco’s traditionally high margins on network plumbing gear (65% vs. the 20% margins of x86 servers), vendors, analysts and customers could be forgiven if they were suspicious of high prices for UCS. In fact one could see some of Cisco’s UCS prices needing to be three times as high as industry averages to meet their business model.
So come on, how about a call for the free market and industry standards. Are we all about the same price? Is Cisco really cheaper?
Michael P. Kendall