Jump to content
HWBOT Community Forums


  • Posts

  • Joined

  • Last visited

  • Days Won


Everything posted by TaPaKaH

  1. not THAT much harder than you might expect. At least with TP45HP (you can ask anyone else if you don't trust me) going from 5-5-5-15 to tighter mem settings (like 5-4-4 or 4-4-4, if your mems can do that) doesn't hurt FSB by more than 2-3MHz not ANY timings. at 7-6-5-18 the thing would not even POST. I find 7-6-6-15 the best in terms of stability/performance for benching at memspeeds over 1000MHz.1.61Vmch = yes, not mistaken. It was the sweet spot for this specific board as X48 does not like high mch (unlike p35/p45) - 600+ PL8 you can see these days from a Rampage Extreme which is based on a X48 chipset, which has worse FSB/tRD ratio than P35- 790i at 616 ... well, you saw from couple of guys (jody/vince) that it IS possible, didn't you? - 8800GTS at 1240core. There were people who ran 1200+ on dry ice and 1250+ on LN2. Why don't we accuse them in cheating as well? - 4Mb to be precise. Back in early C2D days it was common that "white box" ESes had cache misread by CPUz. This thread with all the name calling is going nowhere, IMHO. If you got any concerns, please post them to hwbot crew and I'll be happy to work them through
  2. We all know who gets the best "handpicked" hardware, so, let's follow the same principal and call them cheaters, shall we? Well, I admit that I have spent a lot of time handpicking my CPU at the warehouse I used to work at, but I don't think I have the best one (boble and andre had 6.8G chips, jowi has 6.53G 32M "without pushing it much" chip) the scores are irrelevant to the memory kit I was using. We could hit roughly same clocks with two other DDR3 sets I currently own. (those are not handpicked in any way) No. I've seen multiple D9GMH sets that could do 600MHz CL4 32M at 2.35V DMM or lower (mine can't do that) Why do you think it's best? I reckon mine's one of the worst Anyone, and I DO MEEN ANYONE who had put mch subzero could hit 650+ benchable.
  3. James, in case if you never had experience with P5E64 WS Evo and RampageExtreme - they are exactly same in terms of OC (not considering "gem" board factor), only that the Evo is less bling and less $ (me and mrlobber don't have multi kilo$ budget extra to spend on "bling") The board was not handpicked in any way, it was originally bought from a German e-shop that doesn't give a damn about their customers. I DO NOT GET HANDPICKED HARDWARE DIRECT FROM MANUFACTURERS. You need to learn a lesson (same as I did a while back): don't draw conclusions about hardware you've never owned As for 660MHz CL4 with 2.61V, I have to admit, I have seen better mems
  4. James, who is pathetic in this case is actually you. If the handpicked hardware sent you by manufacturers (Corsair, ASUS, Giga) can't reach certain frequency you claim it impossible, right? As for the Vmem on the PiFast score, it was 2.61 actually (same as on other results) (it's actually read 2.608V by BIOS monitoring tool, I compared that tool with DMM reading on that board with lower Vmem and it's +/- 0.01v) And why are you claiming that fsb650 is impossible on an ASUS X48 board if ... *cough, cough* ... you've hit similar FSB speeds yourself? EDIT: oh, I get it, ASUS didn't send you a handpicked P5E64WS Evo yet, my bad
  5. Fred, may be let's have a public vote, shall we? or, let's make two team rankings - all points score - best scores per hardware score that way we can see who's the biggest and who's the most diverse team
  6. Taking pictures of actual hardware doesn't solve the problem fully, as it only proves THE POSSIBILITY, not THE FACT that the results were done using different hardware. I think we will be getting such hardware sharing reports for on and on until hardware and software developers don't find a way to ID a cards' S/N in a screenshot. Until that happens - question: shared or not shared hardware will rely only on participants' honesty... (not pointing at anyone's scores by saying this ... simply showing that rules are imperfect ATM) Massman, I suggested that division ages ago, but the prob is that, again, we don't know who to specify as a "professional" overclocker as most people (including me) have no sponsorships ... they borrow hardware from their workplaces / friends ... how do you consider them: pro/non-pro ?
  7. who, do you think, is a better overclocker? - guy who runs highend hardware close to its limits scoring global WRs (quality) - guy who runs thousands of CPUs near stock frequency (quantity) since those two are completely different categories - there are two separate ranking at hwbot for either kinds of benchers
  8. xt0m, at least that's how it works in the Baltics
  9. before recieving the CPU you need to sign a warranty letter saying that if CPU dies - you will compensate it's price to Intel
  10. Not all Intel offices are so friendly. Some may require a review or any publicity to back up sending of CPU.
  11. you should better check Danny's thread at XS where he tests similar kit of GSKill. Having 600MHz spec I don't think 625 1M will cause any difficulties
  12. George o/c, OPB's case is already being worked through by hwcrew, very soon we'll see the final verdict. OT (answering your question) I've been trying to play with mchbar a bit, but it gains close to nothing in 1M.
  13. George, sorry for confusion. Will be done ASAP As for coolaler - you know how hwbot works 1.Found suspicious score 2.Don't start flame in public 3.Report the score with your points why it's not valid 4.Let hwcrew decide on it's validity... Me and other people already expressed my opinion - that we see it as valid
  14. To clear things up: 7.469s 1M / 8:03 32M (QX9650) were done on a mrlobber's P5K3 board. We benched together, the owner of CPU was he, so he owns the result. no conflict with hwbot rules. Since he didn't want to kill the CPU trying high Vcore for the WR - I've bought it from him (can show online bank transfer printouts, if necessary) and we tried it with BlitzExtreme board (I've wrote rig. specs at OCX) and did 7.406s 1M which was WR. The owner of CPU was me, if hwbot has anything against it - we will remove one 1M score as that may look like "local hardware sharing" even though I paid full price for the CPU. The QFS logo on the hwbot version of the screen is put due to support I recieved from Foxconn programme in face of 2x1Gb HyperX DDR3 kit which I used for the record. If it is confusing for some, I am ready to upload "normal" screen with no logo. I think that everyone knows, that BlackOps doesn't have "pre45nm" BIOS to boost 1M scores, do you agree? George, let's not start name calling in public again, but I believe that coolaler's record is valid. - As we know, he cooperates with Intel and he INDEED! has access to handpicked pre-released hardware. - He showed CPUz verification / various FM links of CPU running in those MHZ area, so there was no problem for him to run 1M at that clock. - If you think that getting 6.3GHz+ on a E8xxx is impossible, you'd better check on latest results from TeamJapan where like 5 people were getting old-gen E8500s to such clocks.
  15. not speaking seriously: come to think of it - those who pay $500 for pro version get better results ... because they have the chance to tweak their OS better by removing networking components and submit later
  16. again , we're not discussing VISTA over here - everyone has it's own opinion on it! looking at the thinking of those who've voted "FOR" hwbointing that bench I've noticed more of "I can afford it -> I vote for" rather than "Paying $$$ for being able to bench is OK -> I vote for"
  17. we're not talking about hardware requirements here , it just the question of whether or not should the community accept the principal of "non-free" benchmark...
  18. Let's then make a new "benchmark" category at hwbot : everyone goes to their bank ask the lady to print their bank balance take a pic submit and compare Benchmarks that need you to pay money are simply a "NO!!!!!!" because they put everyone in unfair conditions. If it continues - then they'll prompt you to throw 25c coin if you want to continue benching (just like in an arcade game machine)
  19. but we need similar conditions to all the cards and some just won't run that high settings
  20. Suicide runs (no stability tests) 'd be pointless as most of the modern cards have separate 2D and 3D clocks - so you can set 2000/2000 for 3D mode on whatever card you own - run GPUz in 2D mode - GPUz 'll read 2000/2000 - you have #1 spot stupid if we're talking about "benchable" rankings - best way to combine GPU/MEM clocks without separating the categories is a bench that is 100% GFX dependant Sadly, 01 nature isn't such anymore - with 9800GX2 cards you have a heavy CPU bottleneck there
  • Create New...