March 9, 201115 yr Hi Folks, I've been running a Sempron 140 in my unRaid server since I built it a bit over a year ago. It's been running quite well for me, and no complaints. It's a 45 watt chip, and it served files like nobody's business . Fast forward to today: I need to reduce my electrical bill! Having my main pc as well as my server running 24 hours a day is just not helping. Solution: I moved the apps I had constantly running onto my unRaid server, and voila! Instant 100 - 150 watt reduction in usage, as my PC no longer needed to be on 24/7. Of course, now that I was using my little Sempron for more then just basic file serving, I was concerned it might not have enough oomph (who am I kidding? I didn't notice any slowdowns, performance-wise, but we have to think ahead, right?? ). After a bit of research on these boards (Thanks for the advice, folks!), I ended up deciding on a low-power (electrical-wise) Athlon II: the X4 600e. The specs claim this more advanced chip, with 4x the number of cores that my Sempron 140 had, uses the same amount of juice (45 watts).... It's true! I couldn't believe it! At idle, and when all drives spun up and serving files, the new cpu uses pretty much the same amount of power as the old one: These numbers are measured at the wall using a kill-a-watt meter (and the 30 watts that my router, switch, modem, and UPS use has already been removed from the numbers) : Sempron Boot: 150 watts Idle: 60 - 65 watts Spun Up: 100 -110 watts Parity: 130 watts Athlon 600e Boot: 160 watts Idle: 60 - 65 watts Spun Up: 100 -110 watts Parity: 150 watts Note: These are all approximate numbers. Give or take 5-10 watts... As you can see, the Athlon II does use more juice under load, but when idle or serving up content, power usage is pretty much the same. So I just installed the new chip, and I haven't tested differences in more processor intensive activities like unraring etc., but one thing I noticed immediately (which was unexpected), was that the unMenu interface is significantly more responsive.... Not sure I understand why that would be, as I can't imagine it is a terribly processor-heavy task, but for whatever reason, switching from menu to menu, and rendering the new content is noticeably quicker... I hope this post gives some ideas to people considering upgrading AMD processors Some data: My unRaid server: 8 2TB WD Green drives 1 1.5 TB WD Green drive 2 1.0 TB 7200 RPM Seagate drives 750 watt Corsair Power Supply 4 gigs ram Gigabyte board (see sig) 4 120mm case fans 1 80mm case fan Athlon II 600e synthetic benchmark http://www.cpubenchmark.net/cpu_lookup.php?cpu=AMD+Athlon+II+X4+600e According to Passmark CPUMark, the performance seems to be fairly close to the i3 540... Next step will be some underclocking/undervolting to see if I can gain some more power savings while keeping up a higher level of performance then the Sempron... Thanks for reading! Cheers, DB.
March 9, 201115 yr ...according to your meterings provided (150W during boot), your 750W PSU seems well over-sized in order to serve in an efficient range. It should run at least in the range of 20-25% of the nominal output most of its time - most likely that is when unRAID is idle). But according to the rules of thumb for your drives connected (9x green with 2A each and 2x non-green with 3A each) your readings upon boot should say something way over 220W, I'd say. Does your kill-a-watt have a peak metering function? ...well, of course if your plans are to throw in another 10 drives more, your PSU is fine
March 9, 201115 yr Author ...according to your meterings provided (150W during boot), your 750W PSU seems well over-sized in order to serve in an efficient range. It should run at least in the range of 20-25% of the nominal output most of its time - most likely that is when unRAID is idle). But according to the rules of thumb for your drives connected (9x green with 2A each and 2x non-green with 3A each) your readings upon boot should say something way over 220W, I'd say. Does your kill-a-watt have a peak metering function? ...well, of course if your plans are to throw in another 10 drives more, your PSU is fine You're right about the PSU being over-sized . The plan has always been to max out my pro license though, and I figured I may as well get the PSU to support it from the get-go . Expansion stalled for awhile, as my case ran out of drive slots (well, there is one more, but the EARS drive I bought recently to go in there did not pass pre-clear, so I returned it). That said, NCIX just delivered my new Norco case yesterday! Hurrah! As soon as the last bits arrive (still missing the SAS to SAS cables), I will start moving everything over to the new case and add another few drives (once they go on sale, of course). re: Max boot: Not sure what to say. Pretty sure I've never seen it go that high, but my data recording procedure is fairly lax . You have me curious now. I'll shut it down and reboot a few times tonight and see if the numbers vary greatly.... I'm pretty sure the kill-a-watt has a peak metering function. Though I imagine it won't tell me the date when it happened, so I won't know if it was pre or post the new cpu...
March 9, 201115 yr I'd like to see some kind of comparison between the two processors personally. Id like to build another machine sometime, but I guess I'm rather happy with the 140 for now. LOL Found one for the Sempron 140 http://www.cpubenchmark.net/cpu_lookup.php?cpu=AMD+Sempron+140 Sempron 140 - 769 - $30+ AMD Athlon II X4 600e - 2601 - $119.92
March 9, 201115 yr Author I'll see if I can run some benchmarks on both chips over the weekend to get some more data. Anyone have a good suggestion for something quick and easy to test with? My benchmarking experience outside of Windows is pretty weak . re price: I got mine from NCIX, pricematched to BestDirect for $89: http://www.bestdirect.ca/products/222596/AMD/AD600EHDGIBOX/
March 9, 201115 yr I'm pretty sure the kill-a-watt has a peak metering function. Though I imagine it won't tell me the date when it happened, so I won't know if it was pre or post the new cpu... ...I have a metering unit that provides two memory slots for min/max values that can be reset individually from the accumulated readings....don't know about the kill-a-watt though.
March 9, 201115 yr Author To be honest, it's not actually a "Kill-a-Watt" branded meter. I'm using the word like I would use "Hoover" for vacuum or "Kleenex" for ummm.... facial wipes(?) . The one I have is made by "Blue Planet" or something like that. It appears to do the same stuff as the actual "Kill-A-Watt" brand though... Edit: Crap. Just remembered that I used it to see how much juice my space heater pulled on the different settings... If there is a max recorded in there, it'll be reflecting that value.... Guess I'll reset it and see what it does over the next few days...
March 10, 201115 yr Just curious but what apps did you move to your unraid server? Also, i thought the sempron was a 45w chip and the athalon II was 65w?
March 10, 201115 yr Author Just curious but what apps did you move to your unraid server? Also, i thought the sempron was a 45w chip and the athalon II was 65w? I moved Sabnzb, Sickbeard, couchpotato and transmission over . The athlon II seem to range from 20w all the way up to 95w. The u versions are rated at 20 and 25, the e versions are 45 watts, and the rest are 65 - 95watts http://www.amd.com/us/products/desktop/processors/athlon-ii-x2/Pages/AMD-athlon-ii-x2-processor-model-numbers-feature-comparison.aspx
March 11, 201115 yr Author But according to the rules of thumb for your drives connected (9x green with 2A each and 2x non-green with 3A each) your readings upon boot should say something way over 220W, I'd say. Well, I retested again tonight with a couple of shutdown/reboots, and it looks like your estimate was more accurate then my less-than-scientific initial testing! I reset the watt meter and it captured a maximum spike of 242watts ( - 30watts for router etc), so it looks like, for an instant, the server pulled 212 watts! I stand corrected! That said, all the other numbers have stayed consistent, and I'm pleased to report, that for the majority of the time (idle or serving media), the 600e uses no more power then the Sempron 140!
March 11, 201115 yr Well, I retested again tonight with a couple of shutdown/reboots, and it looks like your estimate was more accurate then my less-than-scientific initial testing! I reset the watt meter and it captured a maximum spike of 242watts ( - 30watts for router etc), so it looks like, for an instant, the server pulled 212 watts! I stand corrected! That said, all the other numbers have stayed consistent, and I'm pleased to report, that for the majority of the time (idle or serving media), the 600e uses no more power then the Sempron 140! Thanks for taking the time and for keeping us posted.
March 15, 201115 yr That said, all the other numbers have stayed consistent, and I'm pleased to report, that for the majority of the time (idle or serving media), the 600e uses no more power then the Sempron 140! Do you have any plans on selling the Sempron 140?
March 15, 201115 yr Author Not sure, to be honest. I just upgraded to a norco 4224 as well and have a centurion 590 and 2 4in3s. Haven't decided if I want to build another UnRaid box or sell the parts....
March 15, 201115 yr Not sure, to be honest. I just upgraded to a norco 4224 as well and have a centurion 590 and 2 4in3s. Haven't decided if I want to build another UnRaid box or sell the parts.... Damm, I would've bought that CM590 and the Sempron if I hadn't just sent payment for another case a few days ago to another member here.
March 15, 201115 yr I have a 605e running HandBrakeCLI roughly 40-60 hours/week in addition to unRaid. It is rated at the same 45W as the 600e and has been wonderful.
March 22, 201115 yr Well since building a unRaid box is my long term goal to fix my WHS issues, I figured I would go ahead and create an account to add to this thread. Please note this is an Athlon II which means it supports Cool n Quiet. Cool n Quiet will throttle back the processor to save even more juice if you have a mobo that supports it. I currently have a Phenom II X6 1090T in my main machine at home. It is currently over clocked to 4 GHz on all 6 processors, but it will throttle down to around 888 mHz when it is not under load. (I think it was 888, but I will check this afternoon.) I also found out this processor uses less volts when over clocked than stock with turbo enabled. I will put the meter on it after work today to give a full throttle vs. idle vs. with cool n quite enabled. I will also get you stock voltages vs OC’ed voltages. I am quite confident the 600e will work the same as my 1090T when it comes to Voltages, Cool n Quite, and Turbo. I also currently plan on using a Athlon X4 610e on a HTPC with a mobo that supports Cool n Quite to cut back on the power usage. Edit: I don't understand the testing results. AMD Phenom II X6 1090T 3.2GHz (125W) Test #1 - (OC’ed to 4 GHz core @ 1.328 volts - Cool n Quiet Enabled) --- Kill A Watt style tester --- - 204 W – Max boot up - 140 W – Idle at desktop (Clocked down to 1000 MHz @ 1.328 V) - 274 W – Max at desktop using Handbrake (4000 MHz @ 1.328 V) --- CPUID Monitor --- - 60 W min CPU only at idle. - 141.6 W max CPU only. Test #2 - (OC’ed to 4 GHz core @ 1.328 volts - Cool n Quiet Disabled) --- Kill A Watt style tester --- - 204 W – Max boot up - 143 W – Idle at desktop (4000 MHz @ 1.392 V) - 274 W – Max at desktop using Handbrake (4000 MHz @ 1.392 V) --- CPUID Monitor --- - 141.6 W min CPU only at idle. - 141.6 W max CPU only. There is a 81.6 W difference with Cool n Quite enabled shown with CPUID Hardware Monitor, but the Kill A Watt tester doesn’t show any difference in total watts used.
March 22, 201115 yr Well since building a unRaid box is my long term goal to fix my WHS issues, I figured I would go ahead and create an account to add to this thread. Please note this is an Athlon II which means it supports Cool n Quiet. Cool n Quiet will throttle back the processor to save even more juice if you have a mobo that supports it. I currently have a Phenom II X6 1090T in my main machine at home. It is currently over clocked to 4 GHz on all 6 processors, but it will throttle down to around 888 mHz when it is not under load. (I think it was 888, but I will check this afternoon.) I also found out this processor uses less volts when over clocked than stock with turbo enabled. I will put the meter on it after work today to give a full throttle vs. idle vs. with cool n quite enabled. I will also get you stock voltages vs OC’ed voltages. I am quite confident the 600e will work the same as my 1090T when it comes to Voltages, Cool n Quite, and Turbo. I also currently plan on using a Athlon X4 610e on a HTPC with a mobo that supports Cool n Quite to cut back on the power usage. Do you know if the Sempron line has the same cool and quiet to downclock the CPU to save power?
March 22, 201115 yr Yes... (Looks like there is a little more to it with the older Athlon IIs that just enabling it in the bios.) http://www.amd.com/us/products/technologies/cool-n-quiet/Pages/cool-n-quiet.aspx Scroll down to: Cool'n'Quiet Technology Installation Guide for AMD Athlon processor-based systems (Look at the pdf link. Wonder if this is necessary for Linux / Unix.) http://www.cpu-world.com/CPUs/K10/AMD-Athlon%20II%20X4%20600e%20-%20AD600EHDK42GI%20(AD600EHDGIBOX).html
March 22, 201115 yr Looks like there is a Linux driver. I don't know if this will work with unRaid though, and I would do some testing before putting it into your production unRaid box. Search Page: http://support.amd.com/us/psearch/Pages/psearch.aspx?type=2.1&product=2.1.2&contentType=Tech%20Download%20Processor
Archived
This topic is now archived and is closed to further replies.