Jump to content
HWBOT Community Forums

TaPaKaH

Members
  • Posts

    3674
  • Joined

  • Last visited

  • Days Won

    31

Everything posted by TaPaKaH

  1. Back in the early 1366 days (Q4 2008) people used to bench D9GTR on this platform. Samsung 1Gbit D (then called HCF0) was seen as an inferior option. Then (in Q2 2009) Hypers came along and the game changed. Out of D9GTR, I would say that Corsair 1800C7 Dominators were seen as the best. Sure, CSX Diablo were Hyped, but the binning was not centralised, meaning that their local resellers could assign specs to kits that were nowhere near passing.
  2. Sort of. You can set it, GeiL even made kits rated for CL1.5, but it's not noticeably faster than CL2.
  3. This is one of my rejects. I went through over 1600 pcs E6300/E6400 in total, 17 pcs could do this or higher FSB (583-584) on air. So it's among the top 1% in this respect. Still, I reckon these cascade runs are limited by core
  4. 1125 tRCD 7 will be very hard to do on ambient, even for GTX2. 1125 7-8-6 should be doable.
  5. P5Q3 vanilla sold, some prices reduced
  6. Is 5-8-5 any good for performance? Loving that Rampage I Apex :D
  7. +1 to TerraRaptor. AquaMark3 on 7600GT doesn't benefit from CPU clocks above 5GHz even on SandyBridge, so it's extremely unlikely for a 5.3GHz Coffeelake to outweigh 350-400MHz GPU clockspeed deficit by nearly 30K in score.
  8. some E8600's and 478 collection sold
  9. push
  10. What can this chip do on air for R15 and 32M?
  11. Isn't replacing of DLL files technically considered "illegal practice" as it alters the way in which the benchmark works?
  12. added P5Q3 vanilla
  13. Having a 760-770MHz stability with Hypers too, I guess? :)
  14. I'm amazed that you can get any form of FSB reliability above 600 with Wolfdale on single stage. It seems that no matter what I do, the behavior is only stable for me when the CPU is on LN2.
  15. Try POSTing 2:3 either at 570-580 or at 649+ FSB.
  16. No chance to make 2:3(266) or 5:8(333) to work? On some boards these (especially 5:8) seem much easier than other dividers.
  17. added four E8600s
  18. On most kits you can narrow it down by specs, manufacturing date and OC behavior.
  19. Unless you like the looks, there is nothing special about OCZ Blade. The heatsinks are not very efficient and don't allow modules to be stacked (if that's how you keep them). Moreover, OCZ build quality is by far the worst of all memory brands I know, including obscure Chinese stuff. It seems that they were using lower JEDEC IC bins to save money and had absolutely no idea how to assign specs, which resulted in their weird lineus (such as many 1333C9 models that are NOT rated at 1.5V and still fail specs). Add non-existent support / warranty due to bankruptcy, and you'll quickly realise that any sort of OCZ memory is a waste of money.
  20. Far from. There are multiple generations and versions of PSC DDR3. The ones best known and sought after are ones whose density is 1Gbit and part number begins in XD.. , XE... and maybe also XF. These are usually referred to as PSC X (as opposed to PSC whose part number begins in T).
  21. 2-5kg package to Greece costs 25 Euros with PostNL.
  22. Added a bunch of 965, P45 (all 640+ 32M) and Z77 boards.
  23. added a decent W3680 and a 3.55GHz 32M capable 939 chip
  24. added an ASUS Commando
×
×
  • Create New...