Jump to content
HWBOT Community Forums

TaPaKaH

Members
  • Posts

    3657
  • Joined

  • Last visited

  • Days Won

    29

Posts posted by TaPaKaH

  1. what are TPP and UT points ?
    UP - user global and hardware points (same as in rev3)

    TPP - ("new" points) team power global/hardware points which are awarded based on competition between teams (best result from each team counts) and not members

    you can view how many TPP you've earned for your team if you select "Rank by -> team" instead of "Rank by -> user" (default)

    Example: http://hwbot.org/hardware/processor/core_2_e8600_3.33ghz/#rankBy=teams (#1 CPUz on E8600 gives 19.9 TPP + 49.8 UP, so its contribution for NBCL team is 24.88 points)

     

    and that's the flaw with rev4 ... ~28K single GPU vantage is more worthy for the team compared to 6.8GHz on E8600

  2. like I said on MSN a million times, I don't like the global/hardware TPP distribution

     

    I don't think that any hardware category (including the 50-pointers) gives as much hardware TPP as hardware UP, so all grinding now brings even less points to the team than in rev3

    but on the other hand, global TPP is roughly ~3x higher than global UP ... meaning that teams with marketing pimps onboard have a nice handicap

  3. bugreport - if want to see someone's submissions in their profile - some benchmarks are mentioned multiple times (and those are not really cpu benchmarks)

    pic - http://h7.abload.de/img/aaa0uc0.png

     

    also (may be because of that) the 2D hardware points are calculated incorrectly ... an actual sum of all my 2D submissions is 7357.6 points (calculated manually) but hardware masters page says it's 7407

  4. Don't expect much out of those P4 - I had similar air clocking ones, under LN2 they ranged from 6.8 to 7.7GHz.

    C360s are looking interesting, I never had good clocking air samples from that model.

    5634B is a "too late" batch for a C1, all 5634B C1 i ever had sucked, however Gautam managed to get one to 7.6

    347 you may have a typo there, most likely it's an L740... Most malay from 2007 and newer are very consistent, 5000-5200 on air, 7400-7600 LN2. Almost not interesting to bin :)

    btw: with pcb you mean A0109 for exemple ?
    yes, but I'd like both lines of the PCB texts
  5. Christian, you may have been a bit overkill on the Vcore, most chips don't scale past 1.5V on air.

    But anyway, 5400-5600 is very nice on air, expect those chips to be very nice on LN2 unless you hit coldbugs or FSB walls :)

    5200 air isn't bad too ;)

    Do you mind sharing your test results? (batch/pcb/clocks)

    I'm not a specialist ofc, but pretesting on air doesn't really mean near nothing expecting very high freq under LN² :(

    Maybe a 5,2 will react differently with a higher range of vCore...

    If it can't do 5000/1.45V suicide on air, then there's NO WAY it's an 8G chip ... air pretesting helps to get rid of the crap first so that you don't end up testing 30+ chips on LN2 with half stopping under 7GHz.
  6. 6GHz+ 3D benching is still there. The reason why you don't see so many 6GHz+ Pi scores is the well known problem of 990X/new 980X crapping out well below the wPrime/CPUtest-stable clocks. I know 7GHz chips that can't even run 1M at 6.2

    As for wPrime, I think we already got that maxed in the last 980X wPrime contest :)

×
×
  • Create New...