Warning signs processor is a fraud?

GHoosdumGHoosdum Icrontian
edited June 2004 in Hardware
I just bought an XP3200+ for my girlfriend's PC from a site that I will not name right now. I'm not sure it's legit. My reason is that the processor sold for about 40 bucks less than Newegg was selling it. There were other vendors offering the XP3200+ in the same price range as where I bought it, but only one vendor (the place I bought it from) had decent resellerratings.

I looked at the processor, and there's no obvious signs. The bridges look like they've been untampered with, and the label on the proc says it's a 3200. The edges of the label aren't quite as straight as on the 2500 I'm comparing it to, though.

Are there any other ways I can check to make sure this processor is legit?

Comments

  • EyesOnlyEyesOnly Sweden New
    edited June 2004
    Install it.
  • edited June 2004
    Like EO said, install it and see what it boots up at, also check in bios what vcore it is booting with. If it boots with anything but around 1.65v vcore with bios set to use default vcore, then I would say it's highly likely that it is a rebadged 2500.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited June 2004
    GHoosdum wrote:
    I just bought an XP3200+ for my girlfriend's PC from a site that I will not name right now. I'm not sure it's legit. My reason is that the processor sold for about 40 bucks less than Newegg was selling it. There were other vendors offering the XP3200+ in the same price range as where I bought it, but only one vendor (the place I bought it from) had decent resellerratings.

    I looked at the processor, and there's no obvious signs. The bridges look like they've been untampered with, and the label on the proc says it's a 3200. The edges of the label aren't quite as straight as on the 2500 I'm comparing it to, though.

    Are there any other ways I can check to make sure this processor is legit?

    Well, look at the batch codes, make sure they parse and match to a 3200+ if you can, maybe a local dealer has a batch ref, or AMD in its spec docs has parsing info. If batches do not parse or do not match to what CPU is claimed to be, someone has a processor sold to you that either is not up to snuff and was resold on gray market to them, or has rebadged the thing and not been VERY careful. Lots of rebadgers are not careful with fine details.

    CPUs that also underperform overall as a batch in field can get resold at discount, by batch without particular CPU being tested. CPU not run can be sold as new, legally. But might be not up to snuff as to label. I would also ask AMD to trace this CPU, if it looks like and performs not like thta which it is labelled, report as a possible bad CPU or as a possible counterfit badged CPU. The results should come back farily well with this apporach, as AMD does not like rebadgers or dark gray CPUs. Purchasers are supposed to RMA known bads directly, but it is cheaper for them to resell\dump them in many cases. Price alone makes me think it is worth looking into, unless the seller happens to be a Platinum Partner of AMD, in which case they might be passing on extreme volume discounts in part.

    This CPU might be marginally defective, or misbadged at AMD, that has happened when a missort occurs, also. So mfr direct label validation done promptly or ASAP at problem detect time is best route to get satisfaction, with name of seller to you revealed. AMD may query the seller themselves, they will with a proven misbadge that is reported from more than one person and the RMAs in this case can be done so AMD has misbadging or dark gray market reselling evidence to hand.

    I have seen CPU mfrs ship tested CPUs that perform to label spec of misbadged CPU in exchange for something like this, with very minimal questions asked, so they can trace and have evidence at hand. Advice, talkt o AMD support or email them with message flagged as urgent priority and explain you have a limited resolution time clock ticking to justify urgency flag. Give them everything printed on CPU, line and character perfect, I have seen copyright dates wrong or not changed on misbadges also.
  • GHoosdumGHoosdum Icrontian
    edited June 2004
    Here's the full OPN code:
    AXDA3200DKV4E Y867212231042
    AQXDA0351XPFW (m)(C)1999AMD

    Here is a photo of two processors I have. The '3200' is on the right, there's a 2500 on the left. Do 3200s have different bridges cut than 2500s? I can't see a difference on the bridges between the 2 chips.
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited June 2004
    DKV4E checks out to be a Barton.

    Is the label assuredly integrated into the package of the chip? It's not a sticker, right?

    Check the underside. Make sure there are no pins connected via pencil, metallic trace pen, or wire.
  • KwitkoKwitko Sheriff of Banning (Retired) By the thing near the stuff Icrontian
    edited June 2004
    That label on the right doesn't look right. Can you take a closeup of both labels?
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited June 2004
    It looks like a sticker was placed over the real label.
  • GHoosdumGHoosdum Icrontian
    edited June 2004
    Here's a close shot of both the labels. Unfortunately, my camera gets blurry when it's close, so this is the best I can get.

    The label on the right isn't a sticker (I've tried to peel it to no avail) and it seems to be on the package properly, but it seems to have less even edges and just seems *wrong* compared to the 2500+. However, there are no pins connected whatsoever.
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited June 2004
    Well, it's a good first sign.

    The stickers from the factory are placed onto the OPGA with a ligh-pressure stamping press, and then the CPUs are sealed with that glossy, clear overcoat.

    So if it's embedded in the package, that's good.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited June 2004
    I'll tell you something-- there are different kinds of labels used by different fabs sometimes. My 2500+ came from Malaysia, is a different label, but is genuine per AMD (yes, I CHECKED WITH THEM). NO AMD Logo on top label itself on my 2500+, just centered text. Take your CPU of 3200+ kind, see what if anything is written on bottom, ok???

    Only other thing I can say to do is try it, make sure mobo does not have to be flashed to go to that CPU for voltage and\or detect or set limit reasons, and if it does not perform to spec with about 40 hours of burn in, talk to AMD and seller about a "defective on arrival CPU."

    To be honest, that "3200+" die looks USED or tested in burn in, could be wrong though.
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited June 2004
    The dies range in color from orange to purple. Ain't no thang.
  • GHoosdumGHoosdum Icrontian
    edited June 2004
    Installing it now. I'll post results later. Thanks guys!
  • GHoosdumGHoosdum Icrontian
    edited June 2004
    OK, so far, so good, I suppose. I put it in and it clocked up to 200MHz FSB no problem. The board booted default at 100MHz for some reason, I think it must be a default setting after the CCMOS jumper is tripped.

    It's running at the 3200+ speed at 1.664 VCore dead stable on the Antec 380W TP(when left at 'default' in BIOS, which reads it as 1.65) so it seems like a legit 3200+ so far, or a rebadged 2500+ that can really do 3200+ at default speeds.

    Temp is a little high: 42C at full load, but then again, I put the case fans and CPU fan on the "Fan Only" connectors on the PSU, so it's also very nearly dead silent.
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited June 2004
    Temp is a little high: 42C at full load
    That's not high at all for full load.
  • GHoosdumGHoosdum Icrontian
    edited June 2004
    For the processor, no. For what I'm used to... ;)

    The same PC with a 2500+ OC'd to 2.2GHz, 53CFM fan running full blast with 6 case fans (it was screaming!) ran 39C full load proc temp. I'm willing to give up that 3C for dead silence, I guess.
Sign In or Register to comment.