Jump to content
HWBOT Community Forums

Bilko

Crew
  • Posts

    283
  • Joined

  • Last visited

  • Days Won

    3

Posts posted by Bilko

  1. 51 minutes ago, WhiteWulfe said:

    So's XTU and Catzilla... And Vantage...  And if you ask the right person, Firestrike is just an epeen "look at the massive hole in my wallet" bench. :P

    (note: previous sentence was kind of being silly, and not really serious)

     

    I asked about it since people complain that the newcomers aren't sticking around, and one complaint that seems to come up even amongst the "regulars" is there isn't much variety.

    If a VRMark was to be included I would make it Blue Room

    Orange Room is DX11 and launches in Windowed mode (Changing it means invalid run) which means that Crossfire wouldn't be enabled on older cards while Blue Room being DX12 works just fine with everything and while for benching mGPU isn't a big focus I'd still like the option :)
     

    • Like 1
    • Thanks 1
  2. 20 hours ago, unityofsaints said:

    Now that it's been over 48hrs since the background change it's worrying me that some buggy "first half" submissions are still up in the air:

    1. There are two REX BCLK subs for France (sub 1, sub 2), both using 775 chips. This is on obvious bug, as every team would be running 6 REXes (or maybe just 6 different chips in the same REX :) if that was allowed.
    2. My  406 MHz AM3 BCLK sub not counting because of the CPU socket vs. motherboard socket mixup in in that stage. It's a tough one because the rules actually still state the CPU socket counts but @Leeghoofd made the call earlier in this thread that it should be motherboard socket - can we have clarity once and for all on this matter?
    3. The LGA1151v1 / LGA1151v2 split not getting implemented yet has led to various scores (mostly in the BCLK and CB15 stages) not getting counted, e.g. AG-TAS's 484 MHz sub conflicting with Chili-Man's 525 MHz.

    I'm sure other teams are affected by these bugs too, I'm just using mostly Australia ones as examples as it is easier to keep track of them and it's hard to search for this stuff if you don't know what was benched.

    There are less than two weeks until xmas so the amount of time left to rebench is limited. I know the mod team is busy so I don't expect a fix soon but some kind of definitive answer on whether (and how) these issues will be fixed would be much appreciated ?

     

    132-00.png

    • Like 1
    • Haha 1
  3. 2 minutes ago, cbjaust said:

    For the Reference Frequency Stage it seems to be based on the Socket the CPU is intended for. Here unityofsaints' score using an AM3 Phenom II X4 970 on the C4E (an AM3 board):

    http://hwbot.org/submission/4004746_unityofsaints_reference_frequency_crosshair_iv_extreme_405.85_mhz

    is apparently being trumped by macsbeach98's score using an AM3 Sempron 145 on the AM2+ 790FXB-M2RSH:

    http://hwbot.org/submission/3974543_macsbeach98_reference_frequency_lanparty_dk_790fxb_m2rsh_429.97_mhz

    I think the Reference Frequency stage should be based on the Socket of the Mainboard rather than the CPU being used as seems to be the case. For CPU benchmarks it should be as it is - i.e. based on the CPU socket.

    Pretty sure Unitys submission will be re added to the comp and you're correct, Ref clock goes off Motherboard socket and CPU benches go off CPU socket used.

    It took a while but we finally have an answer for this......I think

    • Like 1
×
×
  • Create New...