The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.

    Core2 ALU unit defective. Possible?

    Discussion in 'Hardware Components and Aftermarket Upgrades' started by coolink, Feb 14, 2008.

  1. coolink

    coolink Notebook Geek

    Reputations:
    0
    Messages:
    76
    Likes Received:
    0
    Trophy Points:
    15
    I ran Sandra arithmetic benchmark with my Intel T7100 (1.8Ghz)and the results are a bit strange for the Dryhstone numbers. It gaves me only 13.000MIPS, that is compared to T2000 series. I saw another bench with Sandra and the same processor T7100 giving 16.000MIPS. Is it possible that my processor has a defective ALU unit? How can I know whats going on with this particular performance issue, since my computer is running ok? The CPU has 3ALU units. Is it possible that one ALU unit is damaged?
     
    coolink, Feb 14, 2008
    #1
  2. coolink

    coolink Notebook Geek

    Reputations:
    0
    Messages:
    76
    Likes Received:
    0
    Trophy Points:
    15
    Any ideas? Is it possible to test the internal CPU parts by software?
     
    coolink, Feb 15, 2008
    #2
  3. Jamaicanyouth

    Jamaicanyouth Notebook Evangelist

    Reputations:
    58
    Messages:
    400
    Likes Received:
    0
    Trophy Points:
    30
    Did you do this test on battery?
     
    Jamaicanyouth, Feb 15, 2008
    #3
  4. hydra

    hydra Breaks Laptops

    Reputations:
    285
    Messages:
    2,834
    Likes Received:
    3
    Trophy Points:
    56
    Are you sure CPU was running full speed at the time of test? If you had a damaged ALU you might be getting other errors? 2+2=5 :eek:

    Kidding aside, maybe try Everest and compare to like processors in case Sandra is having a hissy fit.
     
    hydra, Feb 16, 2008
    #4
  5. Jalf

    Jalf Comrade Santa

    Reputations:
    2,883
    Messages:
    3,468
    Likes Received:
    0
    Trophy Points:
    105
    Two things.
    First, waiting more than one hour before bumping your thread probably won't kill you. Try it some time.

    Second, your ALU's get tested a couple of billion times per second already. If they didn't work, I think you'd know it.
    They're only the most important part of the CPU. The ALU (Arithmetic Logical Unit) is the part that's responsible for actually performing computations (integer computations, at least). So all your software would be have *very* weird (and crash within a couple of milliseconds, most likely) because they got the wrong results back from computations. If that wasn't bad enough, they may also be responsible for memory address computations (Athlon 64 has 3 separate ALU's for this, so each ALU only does one or the other. Not sure about other CPU's, but there's no fundamental reason why one ALU shouldn't be able to do both)
    So even if your applications somehow avoided crashing due to incorrect arithmetics, they'd certainly crash from trying to access the wrong memory addresses (because of errors in computing the address)

    Most likely, the reason for the difference you're seeing is something different, such as imprecise benchmarks (After all you have stuff running in the background that may skew the results), slower memory, or even different bios or drivers. Perhaps even one of Intel's CPU microcode updates have affected performance. There's a lot that can affect benchmark results.

    But if the CPU was in any way defective, you'd have seen far more severe symptoms. ;)
     
    Jalf, Feb 16, 2008
    #5