A few weeks ago, I built myself a brand new machine. I wanted something 64bit and low powered that could act as a VMware or other virtualization server if necessary. I also wanted it relatively cheap.
So between Newegg, Geeks.com, ZipZoomFly and Amazon, I did just that. My primary focus was a "Green PC". I didn't want to sacrifice power (cpu operations) for power (electricity). I also wanted it to be relatively cheap ($500-600). I identified the two of the biggest power suckers in just about any mainstream pc:
- the CPU
- the PSU (in efficency)
I started with an AMD 64 x2 BE-2350. This CPU was on sale for less than 100 bucks and advertised 45w of power usage ... which is fantastic! Many high powered CPUs these days use roughly 100w of electricity, easily.
My other major piece was the PSU. Most power supplies are about 60-75% efficient, meaning much of the juice you're pumping into your PC is being used to heat your room. Aantec's "EarthWatts" PSU's have a rating of 85% or higher. Not only do you save that electricity in the form of heat, but also running the fans to cool down the system.
Once I had those two major parts, I continued. I got a deal on 2gb of DDR2 800 ram, which operates at 1.9v ... which is low by todays standards. The motherboard I chose was nothing special, an ECS A770M-A, the main reason I chose it because it supports PCI-E x16 and up to 32gb of memory.
The video card was also not the latest and greatest. I'm a big fan of Nvidia, and I went with their 6600 series card. (My last card was a 5500). This card supports dual monitors, has 512m of ddr video memory, and probably runs cooler (and with less power) than a top of the line 8800 or 9000 series. Also, Geeks.com had it for about $50!
Next, my old machine had three IDE hard drives and two DVD drives, one a burner and one just a reader. Can you say power consumption? I consolidated the 320gb of drivespace I had with my old machine with one 500 gb SATA drive. I also went with one DVD+/-RW drive instead of two.
Not counting the dual LCD monitors, my new setup uses about *half* the electricity of my old one. I have one of those Kill-a-Watt voltage/wattage monitors. And I took the readings before my last tip.
I installed Slamd64, a 64-bit version of Linux, but could have very well used Ubuntu Server or Workstation. One thing Ubuntu does "out of the box" was enable CPU frequency scaling. I, too, wanted to take advantage of this.
I made sure to enable the loading of AMD's powernow-k8 module (in /etc/rc.d/rc.modules) as well as the "ondemand" cpufreq governor. (Slamd64 has the "userspace" governor enabled by default compiled into the kernel). The main difference between these two, the ondemand gives you power when you need it, while the userspace lets you choose what specific power levels you want.
Finally, these aliases in root's ~/.bash_profile topped it off:
alias hpwr='echo userspace > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor && echo 2100000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed'
alias lpwr='echo ondemand > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor'
My available frequencies were 1000mhz/cpu core on the low end and 2100 on the high end. While my machine idles (I leave it on all the time) it doesn't soak a lot of power unless I access it from remote and tell it to. If I'm working on something important, I can temporarily type "hpwr" and perform my task, then back to "lpwr" when I'm done!
I'll have to take idle measurements with the Kill-a-Watt again to see how low it really goes. Or, maybe I'll just wait to see if I get a nice surprise in my next power bill!
|Non-Linux FOSS: libnotify, OS X Style||Jun 18, 2013|
|Containers—Not Virtual Machines—Are the Future Cloud||Jun 17, 2013|
|Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer||Jun 12, 2013|
|Weechat, Irssi's Little Brother||Jun 11, 2013|
|One Tail Just Isn't Enough||Jun 07, 2013|
|Introduction to MapReduce with Hadoop on Linux||Jun 05, 2013|
- Containers—Not Virtual Machines—Are the Future Cloud
- Non-Linux FOSS: libnotify, OS X Style
- Linux Systems Administrator
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Validate an E-Mail Address with PHP, the Right Way
- Technical Support Rep
- Senior Perl Developer
- UX Designer
- Introduction to MapReduce with Hadoop on Linux
Free Webinar: Hadoop
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?