FhGFS on ARM

Image

FhGFS is short for Fraunhofer Parallel File System (or Fraunhofer FS) and is developed at the Fraunhofer Institute for Industrial Mathematics (ITWM) in Kaiserslautern, Germany. It can be downloaded and used free of charge from the project’s website http://www.fhgfs.com.Fraunhofer FS

FhGFS is a parallel file system, developed and optimized for high-performance computing and is implemented using a distributed metadata architecture for scalability.

Environment

During the multi-stream test, a constant amount of eight clients was used and the number of storage servers scaled from two to eight.

Striping in FhGFS can be configured on a per-directory and per-file basis. Each directory has a specific stripe pattern configuration, which will be derived to new subdirectories and applied to any file created inside a directory. There are currently two parameters that can be configured for the standard RAID-0 stripe pattern: the desired number of storage targets for each file and the chunk size (or block size) for each file stripe.

Striping for this test was configured across the number of storage servers.

Benchmark Specification – Multi-stream throughput

In this benchmark the total throughput of sequential read and write requests with multiple streams were measured.

FhGFS MultiStream Throughput

Intel® Atom® vs Calxeda® ARM®

Intel have released their latest version of their low power CPU the Intel Atom.   The Atom was designed to take on ARMs dominance in the mobile and low power computing market.  So here a Boston Labs we have compared its performance against our very own ARM based Viridis Server.

Comparison of the Intel Atom C2750 and Calxeda ARM A9

ARM A9 ARM A15 Intel Atom C2750
Cores 4 4 8
Threads 4 4 8
Clock Speed 1.1 1.5 2.4
Instruction Set 32 32 64
Cache 4mb 4mb 4mb

Test Systems

The Intel Atom has been tested using the Supermicro 5018A-TN4.   The System has a single Intel Atom. The System was running Centos 6.4 with the latest updates.

http://www.supermicro.com/products/system/1U/5018/SYS-5018A-TN4.cfm

The Calxeda System was a single System on Chip that is part of a 48 Node Viridis System.  The node was running Ubuntu 1304 with the latest updates.

Benchmarks

A range of benchmarks have been tested to compare the overall performance of the two systems: stream, LMbench and Coremark.

Coremark is used to compare the performance of a single core of a CPU.  This is an important comparison here due to the differences in clock speeds and core count. Stream measures the memory bandwidth for the CPUs  accross four operations: Copy, Scale, Add and Triadd.

LMbench compares the CPUs latency and bandwidths across operations using integers, floats and networking.   The results give an indication of the strengths and weaknesses of a CPU, and can be used to suggest the best option.

Results

armaton_coremark

The Atom core achieved just over twice the number of iterations as the Calxeda Core when running coremark.. This is to be expected as the Atom has a Clock speed 2.18 time faster.  When comparing the core mark results it can be seen that this is reflected in the results with a factor of 2.21 difference.

armatom_stream

The stream benchmark results are showing the the graph above.  As expected there has been a noticeable increase in performance between the two generations of ARM CPUs.

The ARM CPUs have half the number of cores as the Atom.  When taking this along with the slower clock speed we would expect the Atom to produce the best performance, but the results do suggest that the ARM CPU is likely to match the Atoms performance with the next generation.

The LMBench Benchmark produces a large amount of data on the CPUs, too much to show here.  But these are some of the more interesting points.

The Graph below shows the time taken to perform mathematical operations on floats and doubles.  These are fundamental operations in computing and are used heavily particularly in scientific computing.  The performance of these operations will give an indication of the overall performance of a code using the CPU.  The Add and Multiply operations give the performance we would expect with the Atom performing around four times as fast.

The divide operation requires a large amount of time in comparison to the other operations which is why high performance codes often try to minimise their use.  The Atom appears to only perform divisions around twice as fast the the ARM A9 CPU.

The Arm A15 has improved its performance significantly in comparison to the A9.  The next generation will move to a 64 bit architecture which should allow it to compete with the Atom.

Atom-Arm_float

 

 

x86 on ARM; Benchmark Results for Eltechs ExaGear Server

Eltechs ExaGear Server targets datacenters and cloud providers and enables them to further decrease TCO by running Intel software on power-efficient ARM-based servers. It is reliable, easy-to-use and fast.

But how fast is it? In order to find out we have performed a series of tests. The tests take at look at the overhead introduced running x86 applications through our translator technology and take a close look at the performance implications for CPU intensive workloads, IO and network intensives workloads.

Benchmark Description

For this benchmarking exercise, Eltechs ExaGear on Boston’s ARM-based Viridis servers, we have used GeoBenchmark freely available on http://geocomputing.narod.ru/benchmark.html. This benchmark evaluates a systems capability to perform data processing, and provides a good benchmark to stress both CPU in IO system capability. The benchmark was built for the ARM architecture and Intel 32-bit. Results of ARM-based (or native) tests were compared against Intel 32-bit tests started under Eltechs ExaGear Server.

For both sets of tests same Boston Viridis server was used.

Input and Output

These tests simulate processing modules with heavy disk input & output.

During those tests Eltechs ExaGear Server demonstrated exceptional performance of more than 90% out of native in average.

Multi-CPU

These tests are designed to estimate pure SMP performance and performance of SMP computers on the “memory access sensitive” algorithms. They show how good applications scale in multiprocessor and multithreaded environment.

Results clearly show that Eltechs ExaGear server is super scalable and does not impact parallelism of multi-threaded applications.

 

Conclusions

Eltechs ExaGear Server demonstrated excellent performance, in particular under heavy IO tests with nearly zero impact on performance while running Intel applications on ARM-based servers. This proves Exagear to be an excellent choice for running storage applications, disk and networking intensive tasks in translation mode on ARM servers today.

CPU intensive performance was around 50%. In future Eltechs expected to go as high as 80% out of native performance in average.

Scalability tests clearly showed that Eltechs ExaGear Server is highly scalable. It does not affect parallelism of applications and can be transparently used for respective software.

Taking in to consideration the ease of use, transparency for end users, and the immediate results – the Eltechs Exagear Server provides a real solution for the problem of migrating old legacy applications to ARM in the datacentre.

 

 

Difficulty porting certain apps to ARM? Just run x86 binaries on ARM with Eltechs Exagear

BIOS IT, a leading manufacturer of high performance, low-powered server, storage and workstation solutions, today announced a new partnership with Boston and Eltechs to introduce Eltechs ExaGear Server the world’s first cloud solution which provides the technology to run x86 binaries and applications on the power-efficient ARM®-based Viridis server without having to port or recompile.

Eltechs ExaGear Server provides a fully transparent way of allowing users access to the low power benefits of ARM-based solutions, while being able to provide customers with services that are run on conventional Intel-based software. Moreover, Eltechs’ solution helps ISVs run their Intel-based programmes on the Viridis without porting or modifying the initial software; avoiding any heavy investment in ARM-based hardware and decreasing time-to-market.

Eltechs ExaGear Server virtualisation engine is a “middleware” software solution, positioned at the gap between the x86 application and the ARM­based server (see figure below). Within the Eltechs ExaGear Server the user sees no difference between running native ARM and x86 applications – Eltechs ExaGear Server virtualisation engine intercepts x86 applications from the very start, converts them in run­time into ARM­compatible code using binary translation technology, and executes them. The entire process is easy, and transparent to the end­user.

So how does this work? Take a look at the video below to see the product in action:

Power cut in central London today – Quick, get down the shop and get more AA batteries!

Power Cut!

We were in with some customers today around Liverpool Street in central London when there was a power cut (link). Initially we just thought the building was out of action but shortly afterwards we realised all the traffics lights and services in the area were in trouble too. We were all rendered useless without email and web access as the visitors wifi access point was one of the first victims of this power cut (Worrying how dependant we have become!). Desperate times, desperate measures, so we decided to relocate to the nearest watering hole (which was packed for that time of day!)

The events of the day got us talking about UPS protection and how long a server, data-centre and even are area could survive on UPS protection. One of the customers who joined us runs a company data-centre and he commented that theres about 40 minutes worth of UPS protection which protects their critical infrastructure….

The power cut began around 3pm and power was restored around 8pm. Our poor customers data-centre began it clean shutdown 4hrs20mins before power was restored! Around 3:45pm, I suggested they should have been running our ARM Viridis servers in their data-centre. The local sys admin grunted and sighed. Restoring the company infrastructure from scratch was a job that was going to take some time.

What if they had been using our ARM Viridis Server?…

I didn’t have numbers for the UPS but lets assume something along the lines of a 3000VA UPS (link). This would last for 40 minutes at about 1500w. Lets also assume the servers were running at 100% load (for ease of calculation rather than typical use case), we would have 8.5w for a Viridis server (link) or 390w for a E5-2680 server (link, ref Table1)

Servers Type # of Servers/Cores Power Consumption Lifetime on UPS
Intel E5-2680 4 servers / 64 cores 1560w 38.4mins
Viridis Cortex A9 180 servers / 720 cores 1530w 39.2mins
Viridis Cortex A9 16 servers / 64 cores 136w 441mins / 7.3hrs

So an equivalent number of ARM cores would have survived on the company UPS! Shame it took a real power cut to realise this..

Boston ARM Viridis Server Extends Its HPC Capability with Allinea Development Tools

Allinea

2013-07-01,  Warwick, UK – Developers porting and debugging HPC/server applications to the ARM architecture received a major boost with the announcement that development tools from the parallel software tools experts, Allinea Software, are now available on the Boston ARM Server platform.

Boston’s ARM-based server has been designed from the ground up to deliver an incredibly power-efficient and highly scalable platform.  The Boston Viridis servers are based on the low power Calxeda EnergyCore ECX-1000 SoC, with quad-core ARM A9 Cortex CPUs and fabric interconnect. With each 2U chassis packing up to 48 servers, customers can populate over 1000 servers per rack, all interconnected via a high bandwidth low latency interconnect. The Boston Viridis server offers a unique low power, high-density platform ideal for specific HPC workloads.

The Allinea DDT debugger and Allinea MAP performance profiler are extremely scalable parallel development tools that enable more efficient use of resources, whilst reducing the complexity and risk of software development. Developers will resolve software defects quickly using Allinea DDT, and tackle any performance bottlenecks discovered during the scale out to multiple servers with Allinea MAP.

“We are delighted to have Allinea Software join the growing ranks of thought-leaders supporting the Calxeda platform,” said Karl Freund, VP of Marketing, Calxeda, “Clearly, having the leader in scale-out development tools available on EnergyCore will be a big help for anyone porting and testing solutions for ARM-based low power servers.”

Allinea Software’s tools were selected due to their well-known ease of use and popularity in High Performance Computing (HPC) and provide the most capable integrated debugging and profiling tools on the market. Allinea DDT and Allinea MAP are used to resolve bugs and improve application performance for multi-core and multi-process applications, and are the de-facto standard for users of the world’s largest supercomputers.

“This represents a vital enabling step for our customers”, said David Power, Head HPC at Boston Limited, “The availability of tools for this platform makes porting, debugging and profiling a more efficient process. The tools are very well known and respected throughout the HPC community and having them ported to our platform will mean users not only have a familiar development environment, but also access to one of the most scalable debuggers and profilers in the industry. It’s yet another migration from the x86 barrier addressed!”

“Energy efficiency is one of the major challenges facing us all today – and developers are keen to discover the benefits that different processor architectures can bring to the applications that matter to them,” explained David Lecomber, COO of Allinea Software. “By supporting this ARM platform, we are providing new flexible approaches to software developers who are seeking to develop software more effectively by using the right tools.”

 

This blog post has been taken from our good friends over at Allinea. Ref: http://www.allinea.com/news/bid/95028/Boston-ARM-Viridis-server-extends-its-HPC-capability-with-Allinea-development-tools

 

 

 

Interworx and Boston at HostingCon 2013

At HostingCon last week, Boston/Calxeda and Interworx demonstrated the Interworx control panel and clustering technology on a 24 server Boston Viridis system. Ever since the spike in interest from World Hosting Days back in March in Germany (where we had an apache web server farm running), we’ve been working hard to enable hosting providers to create Virids-based offerings. A key part of that is the hosting control panel, and we’re pleased to be working with InterWorx on creating the first control panel compatible with our Viridis ARM servers.

The InterWorx Control Panel is a modern web hosting and linux server management system that provides tools for system admins to command their servers and for end users to oversee the operations of their website. Using the RedSleeve distribution of Linux and one of Boston Viridis 24-server systems, the InterWorx team is now able to launch their software in a 24-node ARM cluster.

Check out the Video from Brett Wiewiora of InterWorx who demonstrates the first control panel compatible with ARM servers on a Calxeda-based Viridis server at Hosting Con 2013.

Boston, Calxeda and Fedora Project Deploy Servers on Path to Make ARM a Primary Architecture

Boston Viridis Servers With Calxeda EnergyCore® Accelerate Application Development and Lower Barrier for Porting and Supporting ARM Software Ecosystem for the Fedora Project

St Albans, UK, May 16, 2013 – Boston, a leader in the low power server market, today announced the deployment of Boston Ltd. Viridis servers embedded with Calxeda EnergyCore® chips for the Fedora Project. The availability of the Calxeda-based server cluster is a critical step in helping the Fedora community enable ARM into a primary architecture. This installation, the first enterprise/server-class ARM servers deployed for the ARM port of Fedora, is targeted at the project’s software build infrastructure and accelerates the ability to target the ARM architecture, something the community sees as an emerging industry trend for datacenters in this era of ultra-efficient servers.

“It is nice to see the impact of ARM-based production servers on the Fedora Project,” said Robyn Bergeron, Fedora Project leader, Red Hat. ”The new hardware, and help from Calxeda, Red Hat and other Linaro Enterprise Group members, will facilitate ongoing development of Fedora for the ARM architecture.”

Earlier this year, the Fedora Project, a Red Hat-sponsored and community-supported open source collaboration, announced the availability of Fedora 18, the latest version of its free, open source operating system distribution. Among new ARM architecture-specific features added to Fedora 18 is support for industry standards such as Pre-boot Execution Environment (PXE-boot) technology, a method that is frequently used in datacenter automation to simplify operating system installation on servers. That feature was heavily leveraged by the Fedora Project while deploying the cluster of four high-density Calxeda-powered Viridis systems (each with 24 servers inside).

For the deployment, the Fedora Project migrated off of old infrastructure to the Calxeda-based server cluster which enabled physical consolidation of multiple generations of developer boards into a uniform server environment for the ongoing build and validation activities of the project. This deployment of ARM servers is the first time the Fedora community has been able to use the standard deployment tools like PXE and Kickstart on an ARM system. Since deployment into production, the performance and reliability of the new hardware continues to deliver impressive performance, accelerating application development, and lowering the barrier for porting and supporting the ARM software ecosystem.

“We are thrilled to have been chosen as the launching pad for accelerating ARM into a primary Fedora architecture. The Fedora Project team’s experience – from install to deployment to production – is a testament to compatibility of Linux code on Calxeda: it just works,” said Karl Freund, Vice President, Marketing, Calxeda. “That is what datacenters will expect and demand from ARM platforms, and we plan to deliver.”

Boston triumphant at Green I.T. Awards!

We are delighted to announce that Boston’s ARM®-based Viridis server has been awarded ‘One To Watch Product’ at the fourth annual Green IT Awards! The ceremony took place last night at the Grand Connaught Hotel in London in front 150 professionals from a range of organisations involved in the green IT and energy efficiency arenas.

The Boston Viridis, which started life in November 2011 as a concept for a revolutionary low power server designed to rival the traditional x86 server, was officially launched in June 2012. Upon it’s unveiling at ISC 2012 in Germany it became the world’s first ultra-low power server to harness the CPU technology of ARM Holdings – whose processors are synonymous with consumer electronics, mobile devices and tablets.

Now nearly a year on from its launch the innovative server has been awarded the Green IT Magazine ‘One To Watch Product’, beating competition from Cannon Technologies Ltd and Nimbus Data Systems.

Congratulations to everyone involved for this fantastic achievement!