Finally, a Xeon MP with four cores and modern blueprint

Intel bids adieu to aging NetBurst microarchitecture


Intel has at last brought its high-end Xeon processor out of the stone age, replacing its geriatric microarchitecture and doubling the number of cores to four.

The chip maker on Wednesday officially christened the Xeon 7300 series, a lineup of six quad-core processors that are built on the Core microarchitecture. Careful readers will recall that Intel unveiled Core 15 months ago after CEO Paul Otellini declared NetBurst, the previous blueprint, "long in the tooth".

With its impressive performance and modest power requirements, Core has proven key in helping Intel staunch the loss of sales of lower-end server and consumer chips to AMD. Now Intel is hoping it will spark new demand among enterprise-class customers by expanding the capabilities of an x86 processor.

The 7300 processor, dubbed Tigerton, comes with a new chipset known as Clarksboro. Together, they offer up to four times the memory capacity, double the performance and more than three times the performance per watt compared with systems running Tulsa, Intel's previous top-of-the-line high-end Xeon.

Previously, multi-processor Xeons, which are designed for servers with four or more CPUs, had only two cores and were built on the aging NetBurst design.

At a press conference in San Francisco today, Intel VP Tom Kilroy, who oversees the company's enterprise group, touted benchmarks that showed Tigerton outgunning Opteron, with 2.47 times the performance of the SPECjbb2005, 1.92 times on SPECint_rate_base2006 and 1.55 time on the TPC-C.

According to Intel partners on hand to discuss the new chip, the killer app for all this increased horsepower is virtualization. Organizations don't have enough room to store all the servers they need, and even if they did, they wouldn't be able to afford the bills for powering and cooling them.

Virtualization is largely viewed as the panacea to all this data center apocalypse. Just load the software and use it to create multiple virtual servers running a host of different operating systems. Migrate the apps running on the patchwork of smaller, inefficient machines, and you've saved yourself a bundle.

But all that consolidation requires a server that's up to the task, and that's where Caneland, the name of the new Intel platform, comes in. It has the ability to accommodate 256 GB of memory, comes with a multiple high-speed interconnects to share memory among processors and supports shared L2 cache of up to 8 MB per processor.

It also comes with Intel's promise that customers who invest in Caneland will be able to use a VMware technology called VMotion to easily migrate enterprise apps from old platforms to newer ones. Up to now, VMotion has only worked when customers were migrating to a system in the same processor platform. Now, Intel is pledging that VMotion, which is also used for load balancing, won't be limited to a single platform.

The 7300 series ranges in speeds from 1.6GHz for an 80-watt version to 2.93GHz for one running at 130 watts. A 50-watt version runs at 1.86GHz. Prices range from $856 to $2,301 for quantities of 1,000.

We found most of Intel's announcement relatively sober and straightforward. But there's one thing that still puzzles us: Why would Intel take so long to jettison an architecture its own execs admit is no longer competitive, particularly in a market segment where performance matters the most?

The answer wasn't nearly as easy to decipher as we had hoped. Mainly it boiled down to all the work Intel has to do with its OEM partners to make sure the chip meets the rigors of enterprise-class customers. And then there was last year's launch of Tulsa, which despite being based on an antiquated design, meant that Intel had to wait a decent length of time - more than a year in this case - for another major refresh.

Finally, though, our patience won out.

"There's no question that in some cases, because we didn't have this, we weren't able to win some customers' business," Kilroy confessed. "Core allowed Intel to win back market share. To have had it across the board would have been nice." ®


Other stories you might like

  • VMware claims ‘bare-metal’ performance from virtualized Nvidia GPUs
    Is... is that why Broadcom wants to buy it?

    The future of high-performance computing will be virtualized, VMware's Uday Kurkure has told The Register.

    Kurkure, the lead engineer for VMware's performance engineering team, has spent the past five years working on ways to virtualize machine-learning workloads running on accelerators. Earlier this month his team reported "near or better than bare-metal performance" for Bidirectional Encoder Representations from Transformers (BERT) and Mask R-CNN — two popular machine-learning workloads — running on virtualized GPUs (vGPU) connected using Nvidia's NVLink interconnect.

    NVLink enables compute and memory resources to be shared across up to four GPUs over a high-bandwidth mesh fabric operating at 6.25GB/s per lane compared to PCIe 4.0's 2.5GB/s. The interconnect enabled Kurkure's team to pool 160GB of GPU memory from the Dell PowerEdge system's four 40GB Nvidia A100 SXM GPUs.

    Continue reading
  • Nvidia promises annual datacenter product updates across CPU, GPU, and DPU
    Arm one year, x86 the next, and always faster than a certain chip shop that still can't ship even one standalone GPU

    Computex Nvidia's push deeper into enterprise computing will see its practice of introducing a new GPU architecture every two years brought to its CPUs and data processing units (DPUs, aka SmartNICs).

    Speaking on the company's pre-recorded keynote released to coincide with the Computex exhibition in Taiwan this week, senior vice president for hardware engineering Brian Kelleher spoke of the company's "reputation for unmatched execution on silicon." That's language that needs to be considered in the context of Intel, an Nvidia rival, again delaying a planned entry to the discrete GPU market.

    "We will extend our execution excellence and give each of our chip architectures a two-year rhythm," Kelleher added.

    Continue reading
  • Now Amazon puts 'creepy' AI cameras in UK delivery vans
    Big Bezos is watching you

    Amazon is reportedly installing AI-powered cameras in delivery vans to keep tabs on its drivers in the UK.

    The technology was first deployed, with numerous errors that reportedly denied drivers' bonuses after malfunctions, in the US. Last year, the internet giant produced a corporate video detailing how the cameras monitor drivers' driving behavior for safety reasons. The same system is now apparently being rolled out to vehicles in the UK. 

    Multiple camera lenses are placed under the front mirror. One is directed at the person behind the wheel, one is facing the road, and two are located on either side to provide a wider view. The cameras are monitored by software built by Netradyne, a computer-vision startup focused on driver safety. This code uses machine-learning algorithms to figure out what's going on in and around the vehicle.

    Continue reading

Biting the hand that feeds IT © 1998–2022