Back to Article

  • extide - Monday, January 30, 2012 - link

    That's a nice lookin heatsink, what cpu does that come on? Reply
  • Kristian Vättö - Monday, January 30, 2012 - link

    It's a separately sold heatsink, unfortunately. Part number is E88216-001 and seems to go for around $40. Reply
  • Metaluna - Monday, January 30, 2012 - link

    Also, it looks like a bolt-through mount (i.e. no pushpins). All the companies like Scythe who have produced giant tower heatsinks with plastic pushpins take note. Reply
  • IntelUser2000 - Monday, January 30, 2012 - link

    My friend that got a i7 970 got it with the CPU purchase. Reply
  • Kristian Vättö - Monday, January 30, 2012 - link

    That specific model in the picture is only sold separately as it's for LGA 1156/1155. Your friend most likely has E75476-001, which is the stock cooler for LGA 1366 CPUs released after i7-980X (it was the first one to get it). They look almost identical though, only the motherboard connectors are a bit different :-) Reply
  • Rick83 - Tuesday, January 31, 2012 - link

    Having installed a bunch of Scythe coolers which use pushpins, I've never had a problem with them, and installation has usually been quite simple.
    Don't see the reason for the complaints - unless you live a in a region prone to earthquakes.
  • Taft12 - Tuesday, January 31, 2012 - link

    Ever installed a plastic pushpin cooler multiple times?

    (It's a trick question, the plastic will break, stretch or twist before you get the chance)
  • extide - Thursday, February 02, 2012 - link

    Those plastic pushpins arent that hard to deal with, if you know what you are doing you can easily install and uninstall one several times. However holding a large and heavy heatsink I would prefer metal, but honestly I dont get why everyone has such a hard time with the plastic push pin mounts. Reply
  • Piano Man - Monday, January 30, 2012 - link

    It'll be interesting to see if this 2550K is a special bin that will allow for higher OCs over the 2500k. Reply
  • Mr Perfect - Tuesday, January 31, 2012 - link

    Meh. I'm betting that Kristian is right, this is merely the We-Have-2500Ks-With-Dead-Graphics bin. Reply
  • Olbi - Monday, January 30, 2012 - link

    this Celerons are very expensive. It costs for ours paid 3.55 times more, where avarage man get for hand only 1200 zł , which in dollars is ~399$. I hate my GOV for this shit, what they do now. Reply
  • DanNeely - Monday, January 30, 2012 - link

    They're laptop celerons; a category that's always been significantly more pricey than its desktop equivalents. Reply
  • SlyNine - Wednesday, February 01, 2012 - link

    I've downloaded some decrypting software to try and decrypt your message. As soon as, or if, my computer can decrypt your message I'll have a response. Reply
  • r3loaded - Monday, January 30, 2012 - link

    "The 100MHz increase in frequency isn't worth the loss of the IGP" - it certainly is if you're packing a discrete graphics card. And if you're interested in an unlocked multiplier, chances are high that you will have a discrete card anyway. If these chips are binned higher too, enthusiasts will certainly flock to them. Reply
  • JarredWalton - Monday, January 30, 2012 - link

    Our point is that you can use Quick Sync and the IGP with Z68 motherboards, so even if you have a discrete GPU you can still benefit from the IGP. Reply
  • A5 - Monday, January 30, 2012 - link

    Do any of the open source encoders support Quick Sync yet? And is the quality still bad? Reply
  • Senti - Monday, January 30, 2012 - link

    What is the real use of Quick Sync? It's pure marketing hype. Any sane encoder would prefer quality of x264 over some fps. Add there Hi10P and most hardware encoders/decoders immediately becomes completely useless. Reply
  • Iketh - Monday, January 30, 2012 - link


    I don't understand why quicksync is regarded so highly here. I push it aside as hype, as you said. I'll never use it. x264 all the way.
  • JonnyDough - Monday, January 30, 2012 - link

    Without an IGP it shouldn't get as hot, allowing for that measley 100mhz bump. I would guess these might make some very nice overclockers... Reply
  • MrSpadge - Monday, January 30, 2012 - link

    Nah, they're not that close to the limit, yet, that a few W from the IGP matter.
    And sure they're good overclockers.. as are the 2500K.
  • MrSpadge - Monday, January 30, 2012 - link

    I wouldn't buy or recommend a Sandy without the IGP. Even with a discrete GPU it's going to be used for a few years and then passed on to someone else, where it will very probably do some surfing & office work. And guess what's the most power efficient, yet completely sufficient GPU for this situation? Reply
  • Midwayman - Monday, January 30, 2012 - link

    The nice thing is that the IGP is entirely sufficient to turn the box into a media server. Plus its nearly silent when running. Great 2nd life for the computer. I have an old c2d pulling that duty now and I know my i7-2600k will be headed that way a few years down the line. Reply
  • jak3676 - Monday, January 30, 2012 - link

    Perfect for those of us with P67 mobo's who would never notice the lack of IGP anyway. If you're using a larger SSD for boot + application drive, I don't see the benifit from the caching on the Z68 - may as well get the 2550k and a P67 and save a few dollars.

    If you're going Z68, you'd probably still want the 2500k though.
  • bupkus - Tuesday, January 31, 2012 - link

    Absolutely agree. I've sold several of my older computers which I specifically selected for on board graphics while still retaining my ATI HD 4850 to this day.
    It's a great way to sell/give older tech to friends that can't or don't need new hardware.
  • Mr Perfect - Tuesday, January 31, 2012 - link

    Good point. I'd love to give away my last tower, but took the video card to the new build. Since it doesn't have an IGP, I would have to buy a card just to make it giftable. >_> That was four months ago...

    IGP is also great when you're trying to diagnose a discreet card. Or, worse yet, when you have to RMA one. You at least have something to fall back on.
  • Shadowmaster625 - Monday, January 30, 2012 - link

    Does anyone even know of a product that uses any of these 8xx or 7xx celerons? It seems like a phantom chip. Reply
  • Kristian Vättö - Monday, January 30, 2012 - link

    There are a few in NewEgg: Reply
  • Natfly - Monday, January 30, 2012 - link

    Do you happen to know what stepping/revision these are? Reply
  • Kristian Vättö - Monday, January 30, 2012 - link

    Same D2 as other desktop Core i5 parts. Reply
  • ckryan - Monday, January 30, 2012 - link

    The 2550K is kinda odd. I didn't want the IGP in my 2500k, but it came in handy on occasion. Doesn't the 2500K's Xeon cousin lack the IGP as well? I don't think the 1155 Xeons have a disabled IGP, it's just not there. I'm not sure what the advantage would be, as I would think the number of K parts with defective IGPs is very miniscule. And if you really don't want the IGP, just get a P67 board. Most P67 boards are more power efficient than Z boards anyway. The IGP only consumes a tiny amount of power when used in a Z system with discrete graphics anyway, and I don't think it's going to keep anyone from shattering any OC records.

    The SB Celeron is no joke. I have one in a low power SSD testing rig that is always on. The G530 is very competent, and the IGP is sufficient to adequately tackle the few acceleration tasks asked of it. Plus, if you use an H67 board instead of the H61, you get the platform advantages in a low power, quiet, and well behaved system. Resist the temptation of the single core Celeron 4xx series. It actually uses more power since it doesn't ever downclock or reduce voltage. I know some single cores for low power systems -- but it doesn't make any sense. Even at above MSRP pricing, those who can use the G530 will find the price worth it.
  • MrSparkle404 - Monday, January 30, 2012 - link

    SNB Xeons, if it ends in a 5, IGP is enabled, 0, disabled. is your friend.

    E3-1235 : 4 cores, HT, IGP

    E3-1230 : 4 cores, HT, no IGP
  • Taft12 - Tuesday, January 31, 2012 - link

    Intel: Industry leader in incongruous product naming conventions! Reply
  • Emulex - Monday, January 30, 2012 - link

    noticed the celeron dual core was $34.99 and $49 for z68 mobo at microcenter. the passmark on the celeron indicated it was as fast as the phenom quad core (older) or a c2d 3.0ghz.

    I think someone has their numbers messed up because the older celeron dual core lga1155 had a msrp of $45 on intel's site.

    celeron = 1mb cache
    pentium = 2mb cache
    core = 3/4/6/8. cache.

    Quite honestly most casual users with a $1/gb SSD and $49 nvidia card would not be able to tell the difference between the cache. and the lowest rank celeron (check the microcenter flyer) destroys everything amd has to offer.

    someone needs to check the prices here man. i just opened up my microcenter flyer tonight and noticed $49 z68 mobo + $34.99 celeron dual core 2.4ghz celeron g530 (LGA1155)

    why would they sell a celeron for $107 when you can buy a G530 celeron for $34 ($39 not on sale)?
  • Kristian Vättö - Tuesday, January 31, 2012 - link

    These Celerons are mobile versions, some even being ULV. Hence the price. Reply
  • maroon1 - Tuesday, January 31, 2012 - link

    Actually the desktop version of Pentium processors has 3MB L3 cache

    And Celeron G5xx has 2MB L3 cache
  • Taft12 - Tuesday, January 31, 2012 - link

    Anyone have performance numbers comparing the Pentium G6xx to Celeron G5xx? Especially if it shows performance at equal clockspeed? (so we can see just what difference 2MB vs 3MB of L2 cache makes... I'm betting very little in most cases) Reply
  • ibrihim - Thursday, February 02, 2012 - link

    "No matter what type of workload there is, Celeron G540 is stably 4% behind Pentium G620, which is most likely determined by the differences in clock frequency. It means that cutting L3 cache memory from 3 Mb to 2 MB (by 1.5 times) is a purely cosmetic adjustment, which has more effect on the subjective perception of the technical specifications rather than the actual performance in applications. So, the slight difference in price between Celeron and Pentium processors has only slight influence on their performance.

    Taken from and I detest pushpins.
  • mathew7 - Tuesday, January 31, 2012 - link

    But where are the AES-NI celerons? Or i3s? I really want to use encryption on a fileserver, but that's all that requires CPU performance. Otherwise the cheapest celeron is enough. Reply
  • Taft12 - Tuesday, January 31, 2012 - link

    Sandy Bridge CPUs have more than enough horsepower to AES encrypt on the fly on a home fileserver without the special instructions. Intel has decided you need to pay up if you want that feature. Reply
  • Visual - Tuesday, January 31, 2012 - link

    When we are all waiting for Ivy Bridge, Intel is still selling us old news.
    I wish AMD were more competitive so we don't get such artificial slowdowns to technology progress.
  • wut - Tuesday, January 31, 2012 - link

    Uh, last time I checked Intel's tick-tock is right on schedule. There is still a new iteration once per year. Reply
  • extide - Thursday, February 02, 2012 - link

    Right, but if AMD was more competitive we WOULD see better value from Intel as well. Less crippled processors and stuff released at higher speeds for less money. Intel has plenty of room to lower their profit margin *if they need to* but right now, they dont, so they make as much money as they can. Business 101. Reply
  • dilligad - Tuesday, January 31, 2012 - link

    if 2550k is gpu-less they should name it 2550p or 2550kp, 2550k is totally misleading. Reply
  • Taft12 - Tuesday, January 31, 2012 - link

    Apparently you haven't been paying attention to Intel product naming for a good number of years now -- it's been misleading since the P4 days. Reply

Log in

Don't have an account? Sign up now