Intel Scales Up Cores and Memory with New Westmere EX CPUs

By Michael Feldman

April 6, 2011

This week Intel launched its new Westmere EX lineup, the latest Xeons aimed at large-memory, multi-socketed servers. The new chips come in 6-, 8- and 10-core flavors and will be sold under the name Xeon E7. According to Intel, these latest CPUs deliver 40 percent greater performance than the previous generation Nehalem EX (Xeon 7500 and 6500) processors while maintaining the same power draw.

Compared to the 45nm-based Nehalem EX line, the E7 silicon is on 32nm process technology, which allowed them to add a couple of more cores and an additional 6 MB of L3 cache to the top-end chip. Despite that, Intel only grew the transistor count modestly, from 2.3 billion to 2.6 billion. The thrust was to make the cores smarter and more efficient at their job, not to rely on the brute force of Moore’s Law.

The E7s are 42 percent quicker than their Nehalem ancestors, at least at integer throughput (using the SPECint_rate_base2006 benchmark). One might wonder how Intel accomplished this since they only increased the core count and L3 cache by 25 percent apiece. Apparently 11 percent of the performance increase is the result of optimizations in the latest Intel Compiler XE2011. The rest of the performance bump can probably be attributed to the faster clock for the E7. (Intel pitted a 2.4 GHz E7-4870 against a 2.26 GHz Nehalem X7560 in their benchmark tests.) Floating point throughput (SPECfp_rate_base2006) increased at a more modest 32 percent.

Using the OpenMP benchmark (SPEC OMP2001) for shared memory throughput, the E7-4870 only delivered an 18 percent boost compared to the Nehalem X7560. On some real-life memory-intensive HPC workloads, however, performance was on par with the integer and FP results. For example, Intel reported that throughput improved 21 to 37 percent when exercising the E7s on a number of EDA analysis tools. It remains to be seen how other big memory HPC codes fare on the new hardware.

Besides the core count bump, the other notable E7 feature is its support for larger memory capacity. For a four-socket server, the E7 will scale up to 2 GB of RAM and 102 GB/second of bandwidth, which is twice as good as Nehalem EX. Intel accomplished this by adding support for 32 GB DIMM chips. (The E7 still relies on the same 16 DIMM slots per socket.) These 32 GB DIMMs tend to be rather expensive, though, and so far the server OEMs are only offering E7 systems with 16 GB DIMMs. But 1 TB in a four-socket box is quite useful in its own right, and will be able to handle some rather large in-memory databases.

Perhaps more importantly, the E7 chips can be paired with low voltage memory modules (LV DIMMs) to help curb energy consumption, especially on terascale-sized DRAM configurations. Intel has also added integrated memory buffers to further reduce power draw.

Unlike the Nehalem EX line, the E7 family is divided into three different processor series according chip socket support. The E7-2800 series is geared for two-socket systems, while the E7-4800 series is designed for machines with four CPUs. The quad-socket setup is probably the sweet spot for the E7 family given that four CPUs in one server is apt to be less expensive than 2 dual-socket boxes; plus you have twice the memory headroom. The E7-8800 series is for eight socket machines. These CPUs priced at a premium, but if you’re looking for an x86 SMP machine with up to 80 cores (160 threads) and multiple terabytes of memory, this is the CPU for you.

At launch, 19 server makers announced E7-based platforms, including the usual suspects like IBM, HP, Dell, Cisco, and Oracle. The principle destination for these chips will be “mission-critical” enterprise servers, the segment Intel first pursued in a major with its Nehalem EX line. To chase that application space, Intel has incorporated a number of new security and RAS features which, according to them, puts their latest x86 offering on par with RISC CPUs and even their own Itanium chip. Mission-critical enterprise computing is estimated to be worth about $18 billion per year — about twice that of the HPC server market.

But a number of vendors — SGI, Cray, Supermicro, and AMAX, thus far — are also using the E7s to build scaled-up HPC machinery. SGI for example, has latched onto the E7s to refresh their Altix UV shared memory products. The low-end Altix UV 10 and mid-range Altix UV 100 both benefitted from the extra cores and memory capacity.

For example, the UV 100 now scales to 960 cores and 12 TB of shared memory in just two racks. The top-of-the-line Altix UV 1000 can also use the new E7 CPUs, but for architectural reasons and OS limitations still tops out at 2,048 cores and 16 TB of memory. However, you can still take advantage of the more performant 8-core and 10-core E7s, so a UV 1000 can squeeze out more FLOPS per watt than before, and can scale past 20 teraflops of peak performance.

Cray’s CX1000-S is also being offered with E7 chips. Although, Cray didn’t announce specific configurations, as in the Altix UV, the higher performing E7s would make this SMP box faster and/or more power efficient.

Finally, both Supermicro and AMAX have come up with four-socket and eight-socket E7-based servers (these might actually be the same hardware). The top-end offerings delivers up to 80 cores and 2 TB of memory in an 5U form factor, while the four-socket servers provide half that scalability, but in a 1U, 2U, or 4U package. The 8-way offerings can be outfitted with up to four NVIDIA GPUs if you want to pair the E7 parts with some extra vector acceleration. Although these Supermicro and AMAX systems are geared for HPC, at least the non-GPU versions are also being positioned for big memory enterprise workloads.

As you can see from the chart above, these high-end CPUs are priced accordingly. The top-end 130-watt E7-8870 is over $4,600 in quantities of a thousand. More mid-range E7s will run half that, and even the 10-core chip for dual-socket systems runs over $2,500. Intel apparently believes that they are worth the premium, and given that these chip are being paired with lots of expensive DRAM and software, the CPU itself is probably the one of the best-valued components in these high-end shared memory servers.

Regardless, the E7 parts will be less expensive than RISC processors, the Itanium, or any proprietary CPU. At the other end of the price spectrum, Intel will have to contend with AMD, which is planning to launch its Bulldozer-class “Interlagos” CPU in Q3. Those chips come in 12-core and 16-core versions and can populate four-socket servers. So for users with SMP workloads that are chewing on terabytes of data, the x86 architecture is looking a bit more tempting.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

DoE Awards 24 ASCR Leadership Computing Challenge (ALCC) Projects

June 28, 2017

On Monday, the U.S. Department of Energy’s (DOE’s) ASCR Leadership Computing Challenge (ALCC) program awarded 24 projects a total of 2.1 billion core-hours at the Argonne Leadership Computing Facility (ALCF). The o Read more…

By HPCwire Staff

STEM-Trekker Badisa Mosesane Attends CERN Summer Student Program

June 27, 2017

Badisa Mosesane, an undergraduate scholar who studies computer science at the University of Botswana in Gaborone, recently joined other students from developing nations around the world in Geneva, Switzerland to particip Read more…

By Elizabeth Leake, STEM-Trek

The EU Human Brain Project Reboots but Supercomputing Still Needed

June 26, 2017

The often contentious, EU-funded Human Brain Project whose initial aim was fixed firmly on full-brain simulation is now in the midst of a reboot targeting a more modest goal – development of informatics tools and data/ Read more…

By John Russell

DOE Launches Chicago Quantum Exchange

June 26, 2017

While many of us were preoccupied with ISC 2017 last week, the launch of the Chicago Quantum Exchange went largely unnoticed. So what is such a thing? It is a Department of Energy sponsored collaboration between the Univ Read more…

By John Russell

HPE Extreme Performance Solutions

Creating a Roadmap for HPC Innovation at ISC 2017

In an era where technological advancements are driving innovation to every sector, and powering major economic and scientific breakthroughs, high performance computing (HPC) is crucial to tackle the challenges of today and tomorrow. Read more…

UMass Dartmouth Reports on HPC Day 2017 Activities

June 26, 2017

UMass Dartmouth's Center for Scientific Computing & Visualization Research (CSCVR) organized and hosted the third annual "HPC Day 2017" on May 25th. This annual event showcases on-going scientific research in Massach Read more…

By Gaurav Khanna

How ‘Knights Mill’ Gets Its Deep Learning Flops

June 22, 2017

Intel, the subject of much speculation regarding the delayed, rewritten or potentially canceled “Aurora” contract (the Argonne Lab part of the CORAL “pre-exascale” award), parsed out additional information ab Read more…

By Tiffany Trader

Tsinghua Crowned Eight-Time Student Cluster Champions at ISC

June 22, 2017

Always a hard-fought competition, the Student Cluster Competition awards were announced Wednesday, June 21, at the ISC High Performance Conference 2017. Amid whoops and hollers from the crowd, Thomas Sterling presented t Read more…

By Kim McMahon

GPUs, Power9, Figure Prominently in IBM’s Bet on Weather Forecasting

June 22, 2017

IBM jumped into the weather forecasting business roughly a year and a half ago by purchasing The Weather Company. This week at ISC 2017, Big Blue rolled out plans to push deeper into climate science and develop more gran Read more…

By John Russell

DoE Awards 24 ASCR Leadership Computing Challenge (ALCC) Projects

June 28, 2017

On Monday, the U.S. Department of Energy’s (DOE’s) ASCR Leadership Computing Challenge (ALCC) program awarded 24 projects a total of 2.1 billion core-hour Read more…

By HPCwire Staff

DOE Launches Chicago Quantum Exchange

June 26, 2017

While many of us were preoccupied with ISC 2017 last week, the launch of the Chicago Quantum Exchange went largely unnoticed. So what is such a thing? It is a D Read more…

By John Russell

How ‘Knights Mill’ Gets Its Deep Learning Flops

June 22, 2017

Intel, the subject of much speculation regarding the delayed, rewritten or potentially canceled “Aurora” contract (the Argonne Lab part of the CORAL “ Read more…

By Tiffany Trader

Tsinghua Crowned Eight-Time Student Cluster Champions at ISC

June 22, 2017

Always a hard-fought competition, the Student Cluster Competition awards were announced Wednesday, June 21, at the ISC High Performance Conference 2017. Amid wh Read more…

By Kim McMahon

GPUs, Power9, Figure Prominently in IBM’s Bet on Weather Forecasting

June 22, 2017

IBM jumped into the weather forecasting business roughly a year and a half ago by purchasing The Weather Company. This week at ISC 2017, Big Blue rolled out pla Read more…

By John Russell

Intersect 360 at ISC: HPC Industry at $44B by 2021

June 22, 2017

The care, feeding and sustained growth of the HPC industry increasingly is in the hands of the commercial market sector – in particular, it’s the hyperscale Read more…

By Doug Black

At ISC – Goh on Go: Humans Can’t Scale, the Data-Centric Learning Machine Can

June 22, 2017

I've seen the future this week at ISC, it’s on display in prototype or Powerpoint form, and it’s going to dumbfound you. The future is an AI neural network Read more…

By Doug Black

Cray Brings AI and HPC Together on Flagship Supers

June 20, 2017

Cray took one more step toward the convergence of big data and high performance computing (HPC) today when it announced that it’s adding a full suite of big d Read more…

By Alex Woodie

Quantum Bits: D-Wave and VW; Google Quantum Lab; IBM Expands Access

March 21, 2017

For a technology that’s usually characterized as far off and in a distant galaxy, quantum computing has been steadily picking up steam. Just how close real-wo Read more…

By John Russell

Trump Budget Targets NIH, DOE, and EPA; No Mention of NSF

March 16, 2017

President Trump’s proposed U.S. fiscal 2018 budget issued today sharply cuts science spending while bolstering military spending as he promised during the cam Read more…

By John Russell

HPC Compiler Company PathScale Seeks Life Raft

March 23, 2017

HPCwire has learned that HPC compiler company PathScale has fallen on difficult times and is asking the community for help or actively seeking a buyer for its a Read more…

By Tiffany Trader

Google Pulls Back the Covers on Its First Machine Learning Chip

April 6, 2017

This week Google released a report detailing the design and performance characteristics of the Tensor Processing Unit (TPU), its custom ASIC for the inference Read more…

By Tiffany Trader

CPU-based Visualization Positions for Exascale Supercomputing

March 16, 2017

In this contributed perspective piece, Intel’s Jim Jeffers makes the case that CPU-based visualization is now widely adopted and as such is no longer a contrarian view, but is rather an exascale requirement. Read more…

By Jim Jeffers, Principal Engineer and Engineering Leader, Intel

Nvidia Responds to Google TPU Benchmarking

April 10, 2017

Nvidia highlights strengths of its newest GPU silicon in response to Google's report on the performance and energy advantages of its custom tensor processor. Read more…

By Tiffany Trader

Nvidia’s Mammoth Volta GPU Aims High for AI, HPC

May 10, 2017

At Nvidia's GPU Technology Conference (GTC17) in San Jose, Calif., this morning, CEO Jensen Huang announced the company's much-anticipated Volta architecture a Read more…

By Tiffany Trader

Facebook Open Sources Caffe2; Nvidia, Intel Rush to Optimize

April 18, 2017

From its F8 developer conference in San Jose, Calif., today, Facebook announced Caffe2, a new open-source, cross-platform framework for deep learning. Caffe2 is the successor to Caffe, the deep learning framework developed by Berkeley AI Research and community contributors. Read more…

By Tiffany Trader

Leading Solution Providers

MIT Mathematician Spins Up 220,000-Core Google Compute Cluster

April 21, 2017

On Thursday, Google announced that MIT math professor and computational number theorist Andrew V. Sutherland had set a record for the largest Google Compute Engine (GCE) job. Sutherland ran the massive mathematics workload on 220,000 GCE cores using preemptible virtual machine instances. Read more…

By Tiffany Trader

Google Debuts TPU v2 and will Add to Google Cloud

May 25, 2017

Not long after stirring attention in the deep learning/AI community by revealing the details of its Tensor Processing Unit (TPU), Google last week announced the Read more…

By John Russell

Russian Researchers Claim First Quantum-Safe Blockchain

May 25, 2017

The Russian Quantum Center today announced it has overcome the threat of quantum cryptography by creating the first quantum-safe blockchain, securing cryptocurrencies like Bitcoin, along with classified government communications and other sensitive digital transfers. Read more…

By Doug Black

US Supercomputing Leaders Tackle the China Question

March 15, 2017

Joint DOE-NSA report responds to the increased global pressures impacting the competitiveness of U.S. supercomputing. Read more…

By Tiffany Trader

Groq This: New AI Chips to Give GPUs a Run for Deep Learning Money

April 24, 2017

CPUs and GPUs, move over. Thanks to recent revelations surrounding Google’s new Tensor Processing Unit (TPU), the computing world appears to be on the cusp of Read more…

By Alex Woodie

DOE Supercomputer Achieves Record 45-Qubit Quantum Simulation

April 13, 2017

In order to simulate larger and larger quantum systems and usher in an age of “quantum supremacy,” researchers are stretching the limits of today’s most advanced supercomputers. Read more…

By Tiffany Trader

Messina Update: The US Path to Exascale in 16 Slides

April 26, 2017

Paul Messina, director of the U.S. Exascale Computing Project, provided a wide-ranging review of ECP’s evolving plans last week at the HPC User Forum. Read more…

By John Russell

Six Exascale PathForward Vendors Selected; DoE Providing $258M

June 15, 2017

The much-anticipated PathForward awards for hardware R&D in support of the Exascale Computing Project were announced today with six vendors selected – AMD Read more…

By John Russell

  • arrow
  • Click Here for More Headlines
  • arrow
Share This