Jump to content
HWBOT Community Forums

Casanova

Members
 View Profile  See their activity

  • Posts

    280

  • Joined

  • Last visited

  • Days Won

    3

 Content Type 

Posts posted by Casanova

    First LN2 overclock session in history

    in OC Archeology

    Posted

    I remenber that when I did the first ln2 action in France that was because of the Japanese over and the Tom's Hardware ln2 experiment.

    That was early on for the pot market history and tools but definetly not at the origin.

    Shammy and fredyama might know more on this...

     

    Thanks for this info :)

    I found a video from 1984 where japaneses were making experiments with Supercomputers using Ln2, but tbh that's not a "truly" overclock session.

    First LN2 overclock session in history

    in OC Archeology

    Posted · Edited by Casanova

    The 1st known overclocking on 80x86 was made by People on 8086 in 1980, upgrading from a 4.77Mhz to above 6Mhz ... It was requiring to unsolder the Quartz on the motherboard, and replace it.(I was part of this wagon)

    It went from then to the 486DX33Mhz, Where the OC got it up to an incredible 40Mhz ...

    The 486sx was overclocking badly, as it was the 1st Microprocessor with Integrated memory controller.

     

    The 1st overclockable CPU with Frequency multiplier was the 486DX2 66Mhz, that got overclocked to 100Mhz ... wow! (since then, Frequency multiplier is very often the limiter)

     

    Internally, at Intel, people have been Overclocking the transistors since then (DX with Frequency multiplier) with LN2, because it allows us to draw very important curves for life expectancy. (Life expectancy of a transistor is a function of how many commutes it does, so, if you want increase it by 50%, you get your answer 50% earlier, then, you have a good math model)

    The 1st Public demonstration of overclocking with LN2 was done by one of my good Friend, Art Webb, then, working in the demo group of intel, I know he demonstrated a Pentium III, and a Pentium 4 ... I could find internal literature where he was involved into on Deschutes (Pentium II) (internally)

     

     

    Francois

     

    Thanks A LOT Mr. Piednoel, it's really an honor to have someone with such a huge experience and knowledge about overclock history helping us here.

    I know you must be a very busy person at Intel, but if somehow you have some time left, it would be a great help if you could share content (pictures, videos, scientific paper etc...) about "old" extreme overclock with Ln2 (if you have it).

     

    Thanks again :)

    First LN2 overclock session in history

    in OC Archeology

    Posted · Edited by Casanova

    There must have been.

    At least since late 90's

     

    If i remember correct , the Japanese were using LN2 since slot-1 era.

     

    Fredyama probably knows better.

     

    Great, thanks for the info TASOS, i already contacted him, let's hope he can help :)

     

    I was one of the fists, we made pots from brass pipe and soldered a brass plate to it. I still have them, very crude. I made CPU and CPU pots before I made phase change.

     

    The answer is buried on XS. Kyosen or Shamino would be my guess.

     

    Thanks FUGGER, i'll dig in there then :)

    BTW, could you please send some pictures of this pot? Also if you can share more photos/videos of it i'll appreciate it :)

    First LN2 overclock session in history

    in OC Archeology

    Posted

    This one? Im still 7 or 8 years old then :D

     

     

    Thanks a lot for this link :)

     

    Way way way before then. Even South Africa was doing LN2 by then, and we certainly weren't the first :D

     

    Do you think that Ln2 was on since 90's ?

    I'm asking because i remember that in 1995 (i was 14yo), i lived in a different city (in Brazil called Barretos, San Paolo state) my father took me and my brother to visit his friend in a new laboratory were they had tons of dewars with aninal samples to do research.

    That was my first contact with Ln2, i remember the guy trowing Ln2 in our hands and on my brother's head, just for fun.

    Them some other guy arrived and asked for a dewar that they combined to be borrowed.

    After a while i remember my father's friend telling my fatther that that guy was using the Ln2 to do "computer research".

    I confirmed this history with my father last weekend in a nostalgic conversation about the past.

    I keep wondering how many crazy enthusiastic overclocker guys on this world went extreme and we will never hear about them.

     

    If someone else knows a history or photos/videos please share this, it's overclock history, we need to keep as many records as we can :)

    2D Masters

    in Announce event and group sessions

    Posted · Edited by Casanova

    Hello everyone.

    I'm organizing and promoting a (really) small overclock tournament on facebook:

     

    https://m.facebook.com/groups/606748646118363?view=permalink&id=995839243875966&ref=bookmarks

     

    I made some rules, since it's limited to 4.6ghz for intel and 5ghz for Amd.

    Also, i created 2 (two) formulas to balance as many generations as possible, including the number of cores.

    Of course it's far from perfect, but it's working well so far (the tournament has been running for about 10 days).

     

    Since i could not find sponsorship, there is no prize ;( but the spirit is to make the competitors seek for system optimization and tweaks.

     

    [OVERCLOCK TOURNAMENT]

     

    2D MASTERS

     

    Hello, and welcome to the 1st Tournament Overclock 2D Masters!

    This is an open overclocking competition for any processor from any generation.

    The tournament will have separate rankings for Intel and AMD.

    The rules for competition are the following:

     

    1. Frequency limitation: 4.6ghz Intel and 5GHz processors for AMD processors, with a margin of + 10mhz limit of tolerance, ie the maximum frequency for the result to be considered valid are 4610mhz and 5010mhz for Intel to AMD. Results sent often above these values ​​will not be considered for scoring purposes and will be disregarded;

    2. Place the paper official tournament wall;

    3. Turn successfully each of the following benchmarks:

    • HWBOT Prime

    • Geekbench 3

    • Superpi32m

    • Cinebench r11.5

    • Cinebench r15

     

    4. are not allowed any software or human interference that alter the perception of speed / score of the benchmark program (hack the benchmark program), sabotaging it in order to make him believe that lapped faster.

    5. Tweaks and optimizations of OS and hardware via bios or application are allowed and even encouraged, so the competitive extract as much as possible from your system;

    6. After the result of the benchmark, the competitor must open the CPUZ software with mandatory 3 tabs: "CPU '', '' Mainboard '' and '' Memory ''. Optionally the competitor can also open the "SPD" tab, provided the mandatory 3 tabs are open, too;

    7. The competitor should printar the result screen. The print should contain the open benchmark on the screen to their score, the mandatory tabs CPUZ and the official wallpaper of the tournament (no need to display the entire wallpaper, but at least one part to identify the official paper of the tournament) ;

    8. After the print, the competitor must save it and send the saved file in this post official tournament.

    PUNCTUATION

    The score for ranking of the composition will be the sum of the results obtained in the benchmarks, weighted according to the equation:

     

    INTEL: P = ((G * ΣR) / N) - (Gi * SP)

    AMD: P = ((G * ΣR) / N) - (Gi * (SP / 2))

     

    where:

    P: Weighted score;

    G: Multiplier processor generation;

    ΣR: sum of the results obtained in the benchmarks, excluding the SuperPi (in Geekbench 3 is the sum of scores single and multi, divided by 10, in Cinebench r11.5 will be the score multiplied by 200 in Cinebench R15 will be the score multiplied by 2, and press hwbot be divided by the score 4).

    Gi: Factor generation negative weighted to superpi32m.

    N: number of physical processor cores;

    SP: Result obtained in Superpi32m, converted to seconds.

     

    The multiplier factor "G" in the formula will comply with the following criteria

    INTEL:

     

    • CORE 2: 1.5x

    • Nehalem: 1.32x

    • SANDY: 1.182x

    • IVY: 1.161x

    • Haswell: 1.093x

    • BROADWELL: 1.042x

    • Skylake: 1x

     

    AMD:

    • BULLDOZER: 1.202x

    • piledriver: 1.136x

    • Steamroller: 1.068x

    • EXCAVATOR: 1x

     

    NUMBER OF CORES: the benchmark result will be divided by the number of physical processor cores.

     

    https://docs.google.com/spreadsheets/d/1OckXhumke6aFT52Sox8TGuVrk610G4DunESnk0ycKm4/htmlview

     

    AWARDS: Still got no sponsor for this tournament, but we are still evaluating the idea to fabricate custom medals for the top three finishers in each category.

     

    RANKING: The competitive position in the ranking will be given by the sum of the weighted results in the benchmarks.

     

    RANKING INTEL

     

    1st place: 2008.39 points - Alex Silva Tersaiol

    2nd place: 1771.22 points - Edson Souza

    3rd place: 1767.38 points - Maykon Helver

    4th place: 1742.83 points - Jaime Luiz de Souza

    5th place: 1687.87 points - TheLaw Pedro

    6th place: 1429.50 points - Casanova

    7th place: 1312.20 points - Gabriel Ledema Sales

    8th place: 1137.48 points - Antonio Simiao

    9th place:

    10th place:

     

    RANKING AMD

     

    1st place: 547.33 points - Jaime (Joe90br)

    2nd place: 266.18 points - Lucas Fank

    3rd place:

    4th place:

    5th place:

    6th place:

    8th place:

    9th place:

    10th place:

     

    Contestants can compete with any AMD and / or Intel processor, may compete in both categories. Competitors may rise results with different processors in the same category, but will be considered only the processor to get more points.

     

    PERIOD: The tournament begins today, lasting one month, ending at 23h59m of 10/20/2016.

     

    LINKS:

     

    Wallpaper: Attached to this post.

    CPUZ: CPU-Z | Softwares | CPUID

    HWBOT Prime: http://url.hwbot.org/1ahLmxJ

    Geekbench 3: Geekbench 3 - Cross-Platform Processor Benchmark - Geekbench

    Superpi32m: http://url.hwbot.org/1x9EUxE

    Cinebench r11.5: http://url.hwbot.org/1EPq8jT

    Cinebench r15: http://url.hwbot.org/1FrlRDn

     

    Good luck to all competitors! :)

    Adjustment for Global Points - Work in progress

    in HWBOT Development: bugs, features and suggestions

    Posted

    I think I said it in a earlier post but I want to make a separate comment about this... I think 15 GP, 40 HP and 25CP aren't well distributed, 15 GP maybe is too low in comparision with the amount of the other points, I think they should be 20 GP slots at least, maybe 25 GP Slots since lot of people have dual-core and quad-core CPU now, and maybe even a X99 CPU and/or a high end GPU, and could probably fill 20/25 slots.

     

    Also, who in the earth enters 25 competitions in a year? I think less than 50 people of the thousands of users that HWBOT has, 25 CP slots is totally unnecesary from my point of view...

     

    About 40 hardware points, you need some years at HWBOT to achieve 40 good scores that give decent points, I think 30 would be fine so new users can be a little bit more competitive...

     

    Resuming, I would make 20/25 Global Point Slots, 30 Hardware Point Slots, and 10/15 Competition Point slots

     

    I agree.

    More HP is good for overclockers that bench many different hardware, and more slots for GP is needed because there are increasingly benchmarks listed. And considering that OCe have a year reset, 10 is enough.

     

    My suggestion:

     

    20 Global points slots

    35 Hardware points slots

    10 OCe points

×
×
  • Create New...