Page 2 of 5 FirstFirst 12345 LastLast
Results 16 to 30 of 64
  1. #16
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529

    Arrow

    Barton 2500+ tutorial for AMD desktop CPU speed estimator
    For system specifications look above.
    I chose this example cos showing differences between specific estimators vs general.

    Step 1:
    We follow the same procedure and the first MSF was 2284MHz.
    2284/12=190.33MHz --> FSB is not integer.
    Because of this processor is similar like my samples it is enough if we find one reference point.
    Measurements were:



    Step 2:
    Let's estimate MSF2.
    Vcore2 was 1.709V multiplier = 12X, MSF1=2284MHz, Vcore1=1.659V.
    We enter data as follows:



    As you can see 4th area has possibilities 45% and predicts well cos 2321 < 2340 < 2345.
    On the other hand 5th area has possibilities 72% > 45% but did not predict well.
    This is normal (see ABXzone forum thread O/c Analysis).
    With simple words possibilities is a direction where things are going to but in some cases exist exceptions.
    An event that has 100% possibility (or probability) will always happen and other one with 0% never happens!
    But for Vcore4=1.810V



    only 4th and 6th regions predicted well.
    This happens cos high extrapolation step (look ABXzone forums).

    Notes
    1) Avoid extrapolation values greater than 1.07.
    For our example Vcore4 is 1.810/1.659=Vcore1 at reference point --> 1.091.
    Normally high extrapolation values are not accurate.
    2) In the future possibility values maybe change cos depending on data.
    Last edited by MrSeanKon; 09-05-2005 at 01:33 PM.

  2. #17
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529
    At this point user should start seeking for MSF2 checking his/her system at upper limit of 6th region choosing a CPU stability program such as S&M or Prime95 for 5-10 minutes (see Mapping method) etc...

    Step 3:
    Things are better now cos we have two reference points.
    Let's run O/c Analyzer exponents to see what is going on:





    AMD desktop CPU speed estimator now:



    Vcore3=1.759Volts and MSF3=2377....
    Xm....
    LOL let's use BAN method (see in previous examples).
    Well the maximum new o/c exponent should be the same of course=0.816.
    We input data but the reference point is always (MSF1,Vcore1) that's why it is very important to seek very carefully in the beggining for first reference point.
    See why in OcForums USA (O/c Analysis tutorial).



    6th region predicts 2370 - 2418MHz....
    LOL BAN RULEZ and [2370, 2418] area --> [2370, 2396]....
    Thus we had to try 12 X 198=2376, 12 X 199=2388 and 12 X 200 = 2.4GHz.
    At 2.4GHz CPU failed quick test reboot 12 X 199 failed again and at 12 X 198.08 passed!
    Only two tries and MSF3=2377MHz!
    Last edited by MrSeanKon; 09-05-2005 at 01:36 PM.

  3. #18
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529
    Step 4:
    We continued similar and found (MSF3, Vcore3) + (MSF4, Vcore4) + (MSF5, Vcore5).
    O/c Analyzer exponents (MSF1 - MSF5):





    O/c Analyzer find freq/FSB:



    It is obviously why I put n=0.638.
    We expect that new n < 0.638.

    AMD desktop CPU speed estimator supports 1 - 3 reference points.
    Why not four?
    What can you do now?
    Internal computations with four or five reference points did not give significant precision in the past.
    OK I cannot say more about this it is hard to present the algorithms.
    General rule
    If you have found four reference points then in AMD desktop CPU speed estimator you should input values for 2nd, 3rd and 4th points not 1st, 2nd and 3rd or 1st, 2nd or 4th.
    Similar if you have found five reference points you should input values for 3rd, 4th and 5th reference points not for 1st, 2nd and 3rd or 2nd, 4th and 5th and so on.
    The answer base on Numerical Analysis theorems but it is beyond of my scope to explain why this or that etc...
    So we BAN LOL 1st point and we input data now:



    Note that possibilities values are greater than before.
    6th region predicts [2370, 2418] but O/c analyzer says that 2495 is the MSF6 --> BAN LOL -->
    2469 < MSF5 < 2495MHz.
    I started 11X227=2497MHz and CPU was unstable after at 11 X 226 CPU passed the quick CPU stability test so MSF5=2483MHz!.
    Just two tries!
    Oh my what a waste time!

    Conclusion
    BAN method RULEZ!
    Last edited by MrSeanKon; 03-16-2006 at 03:02 AM.

  4. #19
    Joined
    Jan 2003
    Location
    "show me state" usa
    Posts
    212

    Re: Overclocking Bible: Post your comments/questions here!

    Quote Originally Posted by MrSeanKon
    This photo shows the voltage between red and yellow.
    Note I put probes in inverse order that's why you see a negative value.



    Here is 6pin molex.
    Orange cable is +3.3Volts.
    So orange-black gives you the answer "How is my +3.3 PSU voltage??"

    Why in the world would you put your probes in the 12v and 5v wires on the molex? That's not "inverse order", both probes are on + wires, inverse would be the +probe in black- and the -probe in red/yellow+. Also, a good multimeter will give 3 digits precision without all that battery stuff, if you have to do all that to get 3 digit precision you really need a better DMM (not that I can think of a good reason to need 3 digit precision on the 5v rail anyway ) Other than the voltage measuring it's a pretty cool deal... looks like a lot harder way to oc though..... but I'm an Intel guy and it's a lot easier to oc
    CPUs: 660 3.6es, 840 3.2es, 950 3.4es

    MBs: (2) Asus P5WD2 Premium

    Memory: OCZ PC5400EB Platinum DDR2, Corsair PC5400UL

    Video: ATI X800pro, ATI X850XT PE, ATI X1900XTX

  5. #20
    Joined
    Jun 2004
    Posts
    3,035

    Re: Overclocking Bible: Post your comments/questions here!

    Quote Originally Posted by boshuter
    Why in the world would you put your probes in the 12v and 5v wires on the molex? That's not "inverse order", both probes are on + wires, inverse would be the +probe in black- and the -probe in red/yellow+. Also, a good multimeter will give 3 digits precision without all that battery stuff, if you have to do all that to get 3 digit precision you really need a better DMM (not that I can think of a good reason to need 3 digit precision on the 5v rail anyway ) Other than the voltage measuring it's a pretty cool deal... looks like a lot harder way to oc though..... but I'm an Intel guy and it's a lot easier to oc
    I think the original poster was trying to say that the negative value is shown on the voltmeter because the "+" probe is going to the 5V and the "-" probe is going to the 12V line so the resulting voltage shown is 5-12 = -7V. Please remember that english isn't his first language.

  6. #21
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529

    Thumbs up Thanks again scrandman!

    You replied for me!
    At least we can communicate...
    Of course red line gives (is OK the verb??) positive voltage not negative...
    boshuter I know more expensive multimeters with 3 digits precision for voltages > 2V but they cost 200 Euros or more!
    Last edited by MrSeanKon; 03-16-2006 at 03:02 AM.

  7. #22
    Joined
    Jan 2003
    Location
    "show me state" usa
    Posts
    212

    Re: Overclocking Bible: Post your comments/questions here!

    Hey.... I'm sorry, I never thought about the language snafu Great guide, you must have put a LOT of work into it. Keep up the good work
    CPUs: 660 3.6es, 840 3.2es, 950 3.4es

    MBs: (2) Asus P5WD2 Premium

    Memory: OCZ PC5400EB Platinum DDR2, Corsair PC5400UL

    Video: ATI X800pro, ATI X850XT PE, ATI X1900XTX

  8. #23
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529

    Thumbs up Thanks boshuter!

    Well I have a brilliant idea!
    You can correct my manuals!
    Think something like English teacher!

    Anyway:
    1) First apologize me sometimes I make mistakes.
    2) Well of course you know PSU Red line is +5Volts (by default).
    But this is voltage = potential difference.
    Earth's potential is by default zero volts.
    Consider Everest mountain.
    When you say Everest mountain is 8848 meters high you mean the top there is 8848 meters above from sea level.
    We determinate sea level = zero meters high = reference point etc...
    Thus when Red probe is connected to a High potential (+12V) and Black to a Low potential (+5V) multimeter shows a positive value.
    If you connect Red probe to -5V and Black to -12V (ATX 24pin) multimeter shows a positive value again although both voltages are negative.
    That's why I connected "inverse" the probes to show this in previous photo.
    3) I have o/ced as you can see Intel and AMD systems (more AMDs) I don't think Intel o/cing is easier each platform has some specials but this program is based on general ideas.
    That's why I persist measure your Vcore, Vdimm etc with multimeters don't trust monitors (see FAQ).
    Last edited by MrSeanKon; 09-20-2005 at 10:56 AM.

  9. #24
    Joined
    Jun 2005
    Location
    Nova Prospekt
    Age
    39
    Posts
    6

    Re: Overclocking Bible: Post your comments/questions here!

    This thread has a lot of info about o/c etc....i guess that some links about DDR modules/Chip info will be usefull here.i hope to help someone to avoid serching a lot.

    here are the links

    http://www.techpowerup.com/memdb/

    http://www.biwa.ne.jp/%7Eyok/TENJIKAN-DDR.htm

    http://forum.donanimhaber.com/m_938129/tm.htm

    http://www.xtremesystems.org/forums/...ad.php?t=50010

  10. #25
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529

    Thumbs down Damn!

    Sorry for your inconvenience our site is under construction and you cannot download O/c Bible.
    I called Administrator and he will correct the problem as soon as possible.
    But I must find a better site hope in the future.
    Last edited by MrSeanKon; 11-08-2006 at 07:40 AM.

  11. #26
    Joined
    Mar 2004
    Location
    mtl,canada
    Age
    41
    Posts
    37,286

    Re: Overclocking Bible: Post your comments/questions here!

    dont worry greco.we arent bichin hard enough
    anywya,gave it to a frend and hes liek calling it the best soft hes seen related to oc ever...
    ull have more dls from his uni colleagues and so on.

    add a couple 120mm fans overhead to drop temps by 10+ celsiusD.S.C-12(2)-disclaimer : whatever u do with your hardware/software is your
    responsibility, which i dont hold if u break anything

    GRAB here OcBible v1.55 and Guidemania v1.21
    scary pooch?pooch#1taking ball from dogpooch pulln on ropeme on board after 2 years
    of no board
    http://gonny.se/ <-cool lady http://www.mediafire.com/?8dmg21ms1vx
    IN MEMORY OF HOW BM WAS TREATED HERE...

  12. #27
    Joined
    Jul 2001
    Location
    Round Rock, Tx
    Posts
    1,787

    Re: Overclocking Bible: Post your comments/questions here!

    They're very helpful over there so don't be shy to go ask questions.
    Self Banned for disgust at lack of moderator impartiality.

  13. #28
    Joined
    May 2005
    Location
    Visit Athens...
    Posts
    529

    Thanks moshpit but I must clear some things

    1) My spare/free time is not enough and in general I don't have a schedule.
    2) But this does not mean that I will stop supporting my program at abroad.
    OK it is hard sometimes to visit 6 supported forums (PcPer, XtremeSystems, ABXzone, OCAU, Aoaforums and Ocforums) so any user should be patient when ask something.
    3) I have spent 3 years for this program until now and of course I will not stop this project but sometimes cos my bussiness and personals maybe away.

    Best regards to all users and Moderators here.
    Last edited by MrSeanKon; 06-08-2006 at 03:43 AM.

  14. #29
    Joined
    Mar 2004
    Location
    mtl,canada
    Age
    41
    Posts
    37,286

    Re: Overclocking Bible: Post your comments/questions here!

    good luck greco,bump in here whenever ull be online
    :snif:

    add a couple 120mm fans overhead to drop temps by 10+ celsiusD.S.C-12(2)-disclaimer : whatever u do with your hardware/software is your
    responsibility, which i dont hold if u break anything

    GRAB here OcBible v1.55 and Guidemania v1.21
    scary pooch?pooch#1taking ball from dogpooch pulln on ropeme on board after 2 years
    of no board
    http://gonny.se/ <-cool lady http://www.mediafire.com/?8dmg21ms1vx
    IN MEMORY OF HOW BM WAS TREATED HERE...

  15. #30
    Joined
    Sep 2003
    Location
    Bedford, England
    Age
    32
    Posts
    10,213

    Re: Overclocking Bible: Post your comments/questions here!


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •