SandyBridge 2D madness - let's tweak for the best 1m, 32m, pifast and wprime!

[chew*]

I usually just go in and ask if I can look at the batches... they always say yes lol

They let you just pick through over 120 chips?

The poor guy was bringing them out in cases.

[Kabauterman]

stick to F5x until there is and copy it to backup bios as well with ALT+F12

What is the sense in publishing a Final BIOS what doesn't works better then the last BETA?

[chew*]

This is the best I could pull of so far for sub 6's, 5509 might be doable, note I put more time into this than the 2600K results.

2600K has about a 1 sec advantage at the same clocks so 2500K is def less effecient.

[Linuxfan]

This is the best I could pull of so far for sub 6's, 5509 might be doable, note I put more time into this than the 2600K results.

2600K has about a 1 sec advantage at the same clocks so 2500K is def less effecient.

2200mhz 7-7-5.

[Perica_Barii]

great score, nice cpu and ram vdimm?

[bassplayer]

They let you just pick through over 120 chips?

The poor guy was bringing them out in cases.

Lol. Yeah, there's a glass case full of chips and they just stand next to me while I sort through the boxes!

[TASOS]

UD5 and UD7 are surely nice boards, but too expensive for the mass and "el cheapo suckers" like me Let me know about UD3R, if you say it is ready to rock, I'll get one within' the same second

I believe a UD4 will suit you just fine.

It's a decent board (for it's price) and quick enough with the current beta bioses.

[SuicidePhoenix]

great score chew, could you share your vccsa and vtt voltages?

[chew*]

great score chew, could you share your vccsa and vtt voltages?

tbh i don't think they will matter this chip is a fruit loop compared to the other chips i have used.

They all like 1.14 this hates anything over 1.08.

Other chips don't need system agent, this one needed like 1.05 to even post at anything over 5502 mhz......it just reboot loops to post screen with default system agent.

[Bullet92]

@chew: what's the "secret" of binning cpus by 4 digit number on the chip?

[chew*]

@chew: what's the "secret" of binning cpus by 4 digit number on the chip?

gathering every single good chip info there is put it in a compiler with all 4 digit numbers, have it eliminate all other numbers not in those ranges as bad chips...... so for instance chips from 1XXX to 2599 suck but chips from 26xx to 27xx seem decent........

Then eliminate all batches that we have not seen anything stellar from.

When I saw this chip it like jumped out at me based on this data.

We got a 2600K to, did not see anything really interesting that jumped out at me so we experimented with low 4 digit numbers as the next best alternative and played lotto with an unknown batch.

Out of like 60 chips lowest I could find was 01XX.

I think going by a low 4 digit isn't really usefull with SB, chip does like 5.4, average chip as far as 2600K goes.

I might add that was from a batch with no data at all, L050933, also the chip is extremely voltage hungry.

[SoF]

I believe a UD4 will suit you just fine.

It's a decent board (for it's price) and quick enough with the current beta bioses.

UD4 is nice but back in the days I used to have P35-UD3P and it rocked the hell on earth - friends got UD4 and it rocked the same but I couldn't see why spending more just for some better looking mobo-cooling. I wanna see how the real cheap boards (max ~130€) do - I have P8P67 Pro and MSI C43 here so I need to get an equal gigabyte one. Then if god thinks I deserve some free time, I wanna write a nice comparision review of them - don't try to save dinos work, he'll need to give me some feedback on that UD3(R)

[dinos22]

nice going chewmeiester

[chew*]

nice going chewmeiester

3rd place global just for you dino's

http://www.hwbot.org/community/submission/2126143_

[dinos22]

showoff

[mAlkAv!An]

What vDIMM do you guys need to run your elpida hyper at 1067MHz+ CL7-7-6 for 32M?

[Afrokalle]

Nice benching Chew

Especially that maxxmem

Will try UD4 in next days again and looking forward to push my Hyper's too....

[chew*]

Nice benching Chew

Especially that maxxmem

Will try UD4 in next days again and looking forward to push my Hyper's too....

Thx.

Looking forward to your results.

[chew*]

How about some 01 effeciency on 4870?

God I love that car high.

[Dumo]

Awesome '01 run there chew*

[|ron]

Subscribing to this topic!
Here's what we did at 32M:

6mins 19.219secs
Note that we didn't use Waza, next time we surely will.
Efficiency: 2018.
I think that 108.6bclk 32M stable is something good, too bad the cpu won't go past 5,3ghz

Also, the max bclk achieved is 109.2MHz, not a bad result from our P8P67 Pro!
Hwbot

[Alex-Ro]

Those ram-timings are sick

[|ron]

Yes, nothing to do to tighten them... in that last 32M screenshot, the ram sticks were@LN2, so we were able to push them a little.
Definitely, not the best sticks for sandy bridge.
With some Hypers and Waza, I think we could arrive in the 6min 10secs area
Anyway, first we must change the cpu, 5.3ghz is definitely insufficient

[SuicidePhoenix]

Here a little run, i did the last days.
Is quite okay i think, with better OS and tighter subs there is a bit more to improve.

chew your 1115mhz with 7-7-5 is damn nice, how much vdimm did you use for that?

[chew*]

I doubt what i use matters much man, diff sets of ram need diff volts.....same with cpu's.

My set gets pissed off over 1.76v......use what works for your rig.

I did this run a few days ago when i was working on the 2500K.......then i dropped back down to 1080's worked on it some more and beat it no clue what i did with shot though, swap that out for a 2600K same settings and your probably looking at 6:31.5's

As you can see loops 19-24 was even't a good run, sure it will go faster