Is there really any gain in getting a L part? Guessing Idle power should be about the same it would just limit the max draw? Couldnt find any good comparison on idle power.
Well, they won't get hot.
Lol not surprised there since they are lower power parts, but idle should be about the same correct? Upgrading my NAS server since I am putting 2 10gig nics on it and dont think my poor little 4130t can handle the interfaces + the expanding pool. I have the room for the cooling, but not 100% on if the L tax is gonna be worth it or if the idle savings will be about the same power draw.
I should know that .... but I don't. ... derp! :D
I have a system running on an E3-1240L v3 and that thing is cooled by a Noctua NH-C14S. Without a fan! Actually the whole system is running on passive cooling. So there are use cases for those chips. ;)
"Letter Prefix:
X = Performance
E = Mainstream (rack mount)
L = Power Optimized
W = Workstation for up to 2 processors per motherboard.
The letters also refer to the power consumption figures of the CPUs in question. Note that the numbers associated with the letters are for the "Nehalem"-based Xeon family and older Xeons that use the letter designations have different TDP values associated with the letters.
L = Thermal design power (expected power draw with a 100% CPU load running normal software) of 60 watts or less. These are the "low voltage" chips.
E = TDP of 80 watts. Intel refers to the Xeon 5500/5600s with this letter prefix as the "basic" line and as a result, the idle power draw on these chips is frequently worse than any of the other series of chips, despite the other chips possibly having a higher full-load power draw.
X = TDP of 95 watts. These are the "premium" mainstream chips as their TDP is not so excessive that they are difficult to cool in a rack server, yet their idle power consumption is far lower than the E-series chips.
W = TDP of 130 watts and higher. These are the fastest, hottest Xeons made and they are designed to be used in workstations and pedestal servers with better cooling than an average 1U or 2U rack server."
found this on another forum but would love to see this actually tested. Also doesnt say that idle between E and L is different just that the X are better.
@wendell do you have an Xeon L vs Regular you could do a video on, or have any knowledge on this? I think it would be an interesting comparison. Also a video on 10gig cpu requirements would be interesting.
Nehalem is ages old. That is socket 1366 stuff.
naming is probably pretty similar in terms of naming convention, but yeah couldn't find anything solid on this easily gonna keep digging but figured someone here might have my answer.
all that will probably be different is the maximum multiplier that the CPU can clock to. Most if not all modern Intel CPUs have a base clock of 100MHz. As far as I know, the only thing that would help power consumption at idle would be how the processors are binned, and Xeons are already probably as well binned as you can get. They're the same silicon as the desktop parts, same architecture and everything, but they're cherry-picked for 24/7 operation with constant workloads.
That's what I figured but couldn't find any solid information on it.
There's probably no solid information on it simply because the difference in power consumption at 100MHz (Idle) would be so negligible between one or the other that other factors like VRM componentry, power supply condition, age, etc would have a more noticeable effect on the overall power consumption of the systems they're in than the processors themselves. I'd make the guess that any difference at those speeds would be considered to be within margin of error. To accurately measure the idle power consumption of two different CPU SKUs in the same architecture at 100MHz would require specialized equipment and quite a lot of time with a lot of variables to work with.
you mean a kill-a -watt and the two cpus?
I'm using a Xeon E3-1265L v3 (4c 45w TDP) in my FreeNAS box. It idles at about the same power consumption as my old i3-4130T (2c 35w TDP) which I transplanted into my pfsense box. It peaks up above 200w under heavy load but that doesn't happen often even when decoding X265 videos, if that helps any.
Nope, I mean a (theoretical) testing board that had the necessary onboard components to closely and accurately measure just how much power is going through the CPU at idle. have it clock to 100MHz and measure the current and voltage. You couldn't measure the difference accurately with just any old Kill-A-Watt because honestly, just moving the power source from 120V to 208V would allow the PSU to operate more efficiently, which would throw off your results. There are a lot more variables in a machine that affect power consumption of a machine at idle much more, I would say an order of magnitude more, than the idle power consumption of two CPUs of the same architecture rated at different clockspeeds.