MAXIMUS IX CODE Error with new memory?

[AnonymousGuy]

1 hour ago, done12many2 said:

 

The CPU's are made in batches.  You know, a lot cut from one wafer?  

 

Right now L639Fxxx and L639Gxxx are producing some very high clock speeds.  Of course there are CPUs that aren't as good within those batches so luck of the draw still applies, but the highest clocks have come from those batches.  Why don't you go ahead and change that by getting a high overclock.    Good luck man.

Ok mythbusting time.  Those lot numbers on the box are assembly lot numbers, not fabrication lot numbers.  They're basically meaningless for determining overclock potential.  Even material on the same wafer can have 3 sigma variance on the parameters most associated with overclocking.

[done12many2]

31 minutes ago, AnonymousGuy said:

Ok mythbusting time.  Those lot numbers on the box are assembly lot numbers, not fabrication lot numbers.  They're basically meaningless for determining overclock potential.  Even material on the same wafer can have 3 sigma variance on the parameters most associated with overclocking.

 

Generally speaking yes, but the highest clocking chips are coming from two batch numbers so I don't care if batch numbers are just something that some guy writes on the side of a box.  If he keeps writing the same number on the side of boxes that luckily contain the fastest chips, I'm buying one.

 

I'm not sure if you've been down the road with a Kaby Lake chip, but if not, feel free to check out the following threads on oc.net.  I think you'll see the same two batch numbers popping up for the guys hitting the highest clocks at the lowest voltages.  

 

Hell, @Lays and I have the same batch number and we both ended up with 5.1 chips that became 5.2 and higher chips after delidding.

 

The validated 5.2 chips coming out of Silicon Lottery also happen to be from one of these two batches.  

 

While I appreciate your "myth busting," and agree that this has generally been the trend in the past, it isn't exactly shaping up to be the same thing this go around.  

 

As far as "3 sigma variance."  I covered that when I said that not all chips from those mentioned batches are perfect and that it still remained luck of the draw.  The highest chips coming from those two batches remains. 

 

http://www.overclock.net/t/1620203/kaby-lake-5ghz-milestone/0_100

 

http://www.overclock.net/t/1621347/kaby-lake-overclocking-guide-with-statistics/0_100

[AnonymousGuy]

16 minutes ago, done12many2 said:

 

Generally speaking yes, but the highest clocking chips are coming from two batch numbers so I don't care if batch numbers are just something that some guy writes on the side of a box.  If he keeps writing the same number on the side of boxes that luckily contain the fastest chips, I'm buying one.

 

Just for the sake of it, I checked the OP's lot number.  It was ~600,000 die from 8 lots fabbed on a date range from Sept. 26 to October 14.  That's a wide spread of a lot of material.  It's kinda like doing a rain dance every day and then when it finally rains concluding that the dance must have finally worked

 

PM or post your serial number and I can get back to you if I find it was truly an exceptional chip on the parametric curve.

[done12many2]

14 hours ago, AnonymousGuy said:

PM or post your serial number and I can get back to you if I find it was truly an exceptional chip on the parametric curve.

 

I've got 7 of them for you.

 

I can tell you that 3 of them were garbage with regards to overclocking as they were either terrible in clock speed, cache or IMC.

 

It'll be really cool to see if I could have just avoided binning all of these chips by knowing a guy like you who has a method of determining exceptional chips based on serial numbers.  That would save a TON of time and damn sure get you added to my Christmas card mail out.    Thanks.

 

7700K SN:

 

M64N83W702366

M6YY510600388

M6PG615302683

M6939PB403105

M6E42F9701847

M6RW633101119

M6332XX603929

[AnonymousGuy]

4 hours ago, done12many2 said:

 

I've got 7 of them for you.

 

I can tell you that 3 of them were garbage with regards to overclocking as they were either terrible in clock speed, cache or IMC.

 

It'll be really cool to see if I could have just avoided binning all of these chips by knowing a guy like you who has a method of determining exceptional chips based on serial numbers.  That would save a TON of time and damn sure get you added to my Christmas card mail out.    Thanks.

 

7700K SN:

 

M64N83W702366

M6YY510600388

M6PG615302683

M6939PB403105

M6E42F9701847

M6RW633101119

M6332XX603929

Was M6E42F9701847 the best one?

   

[done12many2]

3 minutes ago, AnonymousGuy said:

Was M6E42F9701847 the best one?

   

 

Okay, how in the f*** did you do that?  I expected you to get close, but that's remarkable.

 

If it's something you need to PM me about, feel free to do so.

[AnonymousGuy]

4 minutes ago, done12many2 said:

 

Okay, how in the f*** did you do that?  I expected you to get close, but that's remarkable.

Out of the 7 it's the fastest silicon vs. leakage.My guesses for the 3 shit clocking ones are: -929, -366, and -119.  If I get time I'll make some plots showing where the 7 are vs. the 100k other units in the lots they were apart of.

[done12many2]

2 minutes ago, AnonymousGuy said:

Out of the 7 it's the fastest silicon vs. leakage.My guesses for the 3 shit clocking ones are: -929, -366, and -119.

 

929 - Decent before delid - great after.  

 

I just bought it off a guy on OC.net who tested it prior to selling it to me.  It did 5 GHz at a 1.34 Vcore  (1hr RealBench) and has a great cache.  I popped the top on in thinking that it had potential and it had the worst TIM coverage from the factory that I've ever seen.  Just smaller than a pea sized circular area on the center of die wasn't even making contact with the IHS.  Delid dropped voltage along with temps to the tune of 30c less.  I only got 20c to 25c out of two other delidded 7700k CPUs, so that TIM was screwing stuff up pretty bad on this particular chip.   It's actually running fantastic now and will be heading to one of our LTT members here today.

 

366 - Definitely a higher voltage chip, but scaled okay - got rid of it.

 

119 - No good across the board.  Installed it and pulled it about 3 minutes later.  One of those chips that you just know ain't going anywhere form the get go.

 

Very impressive man!  I'll PM you to find out more about this.

 

 

[TenzoNakami]

22 hours ago, done12many2 said:

 

I've got 7 of them for you.

 

-snip-

Now this is something.

7 i7 7700K. Lots of 7.

 

18 hours ago, AnonymousGuy said:

Out of the 7 it's the fastest silicon vs. leakage.My guesses for the 3 shit clocking ones are: -929, -366, and -119.  If I get time I'll make some plots showing where the 7 are vs. the 100k other units in the lots they were apart of.

 

17 hours ago, done12many2 said:

-snip-

 

Very impressive man!  I'll PM you to find out more about this.

 

 

Now this is unbelievable. What about mine? Any good? I am also looking forward for the PM!

[AnonymousGuy]

5 hours ago, TenzoNakami said:

Now this is something.

7 i7 7700K. Lots of 7.

 

 

Now this is unbelievable. What about mine? Any good? I am also looking forward for the PM!

I'll give it the ole college try!

[AnonymousGuy]

5 hours ago, TenzoNakami said:

Now this is something.

7 i7 7700K. Lots of 7.

 

 

Now this is unbelievable. What about mine? Any good? I am also looking forward for the PM!

Probably a 5-5.1Ghz chip with average thermals.  done12many2's -683 chip is very similar.  

 

I really wish Intel would just sell report cards on chips or do what EVGA does with different tiers of binning (for ASIC quality on Classified GPUs) and changing prices.

[TenzoNakami]

On 1/28/2017 at 1:37 AM, AnonymousGuy said:

Probably a 5-5.1Ghz chip with average thermals.  done12many2's -683 chip is very similar.  

 

I really wish Intel would just sell report cards on chips or do what EVGA does with different tiers of binning (for ASIC quality on Classified GPUs) and changing prices.

How do you do that? (Finding out the quality and such?)

 

And yes, the farthest I tried was 5.2GHz already reaching around 80 degree celcius with my current watercooling setup. With the room ambient sits around 32, it has ~50 degree delta temp? 

 

I wish I could do the same thing, since I could choose which chip to get (Yes, I have access to their storage!)

If I were you I'd be looking up for one of the best die out there 

[AnonymousGuy]

2 hours ago, TenzoNakami said:

How do you do that? (Finding out the quality and such?)

 

And yes, the farthest I tried was 5.2GHz already reaching around 80 degree celcius with my current watercooling setup. With the room ambient sits around 32, it has ~50 degree delta temp? 

 

I wish I could do the same thing, since I could choose which chip to get (Yes, I have access to their storage!)

If I were you I'd be looking up for one of the best die out there 

PM me a bunch of serial numbers and I can tell you the best one out of them.  There's about a 1% chance you'll get a god chip that can hit 5.3+Ghz, with a limited sample size seems to be 10-20% chance of getting a 5.2Ghz chip.  Obviously it all depends on being able to cool the damn thing.  All 7700K's are at least "average" Kaby Lake parts supporting 4.9-5Ghz.  The below average ones probably end up being lower in the SKU stack.

 

I did it for my 5960X but HSW-E had a weird overclocking thing where if it got above 85C it would just not be happy at all, so even though it could theoretically clock high trying to keep 8 cores cool just became the limiting factor (even with chilled water).