Jump to content
HWBOT Community Forums

ZFeSS

Members
  • Posts

    214
  • Joined

  • Last visited

  • Days Won

    2

Posts posted by ZFeSS

  1. When I try to search fastest SPi 32M LGA771 result(http://url.hwbot.org/1i4iU4i), it shows me that GENiEBEN is fastest, but he is only 5th in QX9775 category. When I narrow my search(adding QX9775 in cpu model), that shows me, that there are only 2 submissions in SPi 32m with QX9775. If you look at QX9775 you will se that it's 8 result there. Other can be found with LGA775 and QX9775 filter, which absolutely wrong. If reason of this bug is in wrong define of socket for QX9775 before, and after changes old submissions still got LGA775 and new - 771, it may happen in other categories(LGA2011 moved to LGA1356, s603 moved to s604, etc).

  2. I smell a GPU-Z bug...

    NV42 definitely clocks better. I've tested some 6800XT and GS cards and can say that NV41 clocks 70-80MHz worse at same voltages. Bug really exists. At all of this cards GPU-Z shows NV41.

     

    All cards are created unequal, by the same definition. If some manufacturer have deliberately decided to make their card slower (=worse), why can't we just accept the fact that the card is not too competitive against other cards of the same model rather than generating a specific category for it?

    Using this logic if non-reference cards get additional memory and packed with DDR2 instead of DDR3 they also should be in the same category, cause "some manufacturer have deliberately decided to make their card slower (=worse)", and at the other side 4850 GDDR5 should be placed with GDDR3 ones, it will be just "lightning" ;)

    Cards with different memory sizes usually packed with different rated ICs. Reference Ti 4200 64MB got 4ns mem, and reference 128 - 5ns, for this reason they even got different default clocks. And there no difference between them in terms of speed at the same clocks in possible benchmarks - 64mb is enough for 03.

×
×
  • Create New...