Jump to content
HWBOT Community Forums

ground

Members
  • Posts

    634
  • Joined

  • Last visited

  • Days Won

    10

Posts posted by ground

  1. i7 920 common data:

    - batches 3845A to 3849A are the typically binned chips. Best chips I've seen so far of these batches can do 4.5/1.24V 32M on air or water, 230-240 BCLK air and scale to 5.4 (though thats really rare)
    -batch sorry enough people are buying this batch now so I’m removing all data I can is a more exotic batch that I recently stumbled over; as far as I could tell by limited data nobody besides me seriously binned this range. I've so far found a W3520 and an i7 920 that are both doing 4.5/1.2-1.21V on air for 32M; the W3520 scaled to 5.53 32M 1.45V easy on Cascade and is probably not at its limit. The 920 I tried testing on LN2 but it was, at least for me, far too hungry to be controllable at all. 5.45 1M at 1.45V was the max I managed with about 15L of LN2 in maybe 1 1/2 hours (chip is insanely leaky - heats up the pot by 20° with an F1 pot during 1M). Gonna test that chip on a strong cascade when I get around to it again. So far had about 4 chips in this range and all did 5.2+ and seemed to scale to ~1.55-1.61V depending on the temperature (the lower the less voltage scaling). The W3520 also does 5.3 GHz uncore at 1.55V VTT on Cascade which seems quite nice. 

     

    With 45nm there seem to be a lot of chips that don't seem to scale with voltage at all (even with the good batches), some not even scaling beyond 1.5V on LN2. Have seen a couple chips that did 4.5/1.3V air and could barely do or couldn't even manage 5GHz on LN2. 

    Much later chips can sometimes be quite nice but those you would need to bin a lot of. Of my 12 W3530s (all 3040-3150 batches) I had one chip validate >5.3 and scale with a ton of voltage (saw scaling up to 1.8V during post and 1.75V booting up).

  2. did you try raising PCIe frequency for more BCLK? 221 lines up with the pcie at 100 wall... You can estimate the required PCIE with ~targetBCLK*0.452 (on asus with good chips typically the max postable is 261 and 118, at least with 32nm (45nm can post higher PCIe for some reason). '10 32nm usully manages 225-255 BCLK on ambient, '11 chips can be more capable (max I've seen on ambient was 273 though I'm sure there are better ones out there). Haven't seen chips that were IMC limited at the 10x mem multi for 32nm yet.
  3. 4 minutes ago, Mr.Scott said:

    So......in less than 3 minutes, you had a LN2 bench done, stripped of pot, processor changed, pot remounted, and a second bench done?

    Bullshit

    No confusion here.

    I mean... they could be running with the pot just standing on the CPU (I've done that before binning) and since the runs are at basically at stock it doesn't seem impossible I guess. Just gonna try seeing a way its plausible. I'm 99% sure they'll rerun if enough people complain, its not like either score is particularly impressive.

    Edit: Also joining the complaining party since I wanna see runs that aren't basically stock...

  4. 31 minutes ago, mickulty said:

    BTW as far as the underlying logic, if anything it's weirder that DDR3 and DDR4 RDIMMs aren't listed separately since the CAD bus is done totally differently on them.

    I mean for those there are at least IMCs that can use udimms as well as rdimms without any issue, sometimes even on the same boards (x99 especially). Pretty sure that doesn't apply to any platforms with FBdimms or do I remember it wrong?

    • Thanks 1
  5. Anyone here has experience getting hyper to run with 32nm 1156 chips? We have a couple 650s and 550s here that won't post with hyper at anything beyond 800 MHz if even that while handling PSC at >1050 or so...

    edit: and once again noting that almost nobody else running high QPI which makes it easy :D 

  6. On 7/20/2019 at 6:58 PM, Leeghoofd said:

    CB R15 (3 scores) -LGA1156 (server cpus allowed)

    Quick suggestion gauging interest - would it be more popular if its 2C/4T/8T instead of 4c8t for all 3 subs (which would be unavoidable otherwise)? Not sure how that would go for moderation though... Either way, I'm happy to see 1156 :D 

    • Like 1
×
×
  • Create New...