Statistical evidence Intel most likely cherrypicks cpus for reviewers

[leadeater]

22 minutes ago, DildorTheDecent said:

>Implying Intel doesn't have in house overclockers.

 

Come on man, of course they have an in-house overclocking unit.

 

How else would they have made this?: https://www.intel.com/content/www/us/en/gaming/how-to-overclock.html

 

Also @Lays if you're still around. Worked at Intel as an overclocking technician.

I'm not saying they don't, I'm saying they are not testing tens thousands of CPUs to find the best ones to send to thousands of reviewers because globally there is a lot.

 

Like I said originally binning for product SKU and binning for OC is very different, the K SKU is literally the golden dies that people are talking about.

[JoostinOnline]

2 minutes ago, leadeater said:

I'm not saying they don't, I'm saying they are not testing tens thousands of CPUs to find the best ones to send to thousands of reviewers because globally there is a lot.

 

Like I said originally binning for product SKU and binning for OC is very different, the K SKU is literally the golden dies that people are talking about.

I think it's fair to assume that they test if a K model can hit a reasonable overclock before sending it out. Basically an extra binning step. I absolutely would, and I wouldn't count it as wrong either. It's like testing to make sure that a product actually works before sending it out. Customers get DOA products all the time, but I've never heard of a reviewer getting one. I don't know why people are picking on Intel either. I'd be shocked if every other company didn't do this as well. 

 

I recently bought two stock 980's. One of them hit a decent overclock, but the other wouldn't budge a bit before you got artifacts. Obviously crappy models exist, even though they're rare. Does anyone think a company would risk sending a bottom of the barrel product to a reviewer who's going to test overclocking ability? It would be embarrassing. 

[leadeater]

3 minutes ago, JoostinOnline said:

I think it's fair to assume that they test if a K model can hit a reasonable overclock before sending it out. Basically an extra binning step. I absolutely would, and I wouldn't count it as wrong either. It's like testing to make sure that a product actually works before sending it out. Customers get DOA products all the time, but I've never heard of a reviewer getting one. I don't know why people are picking on Intel either. I'd be shocked if every other company didn't do this as well. 

 

I recently bought two stock 980's. One of them hit a decent overclock, but the other wouldn't budge a bit before you got artifacts. Obviously crappy models exist, even though they're rare. Does anyone think a company would risk sending a bottom of the barrel product to a reviewer who's going to test overclocking ability? It would be embarrassing. 

And that is far more likely, checking it's not a bad one, than checking thousands of CPUs to find the halo ones. Filtering out the bad ones from reviewers is much simpler to do and requires a lot less time, like you said I'd do it too.

[porina]

I don't want to give the guy a click. Can someone summarise the data sources and alleged conclusions thereof? I only just saw this thread, after seeing him getting dragged into other "discussion" if you can call it that on twitter... is he using hwbot?

[Sniperp2704]

[epsilon84]

Alright guys hows it going?

 

Let's put it this way - his channel basically exists to go against the grain, and his viewers love it, just look at the comments section, it's basically everyone telling him how great he is for 'exposing Intel'.

 

This is his third ConLake video I believe?

The latest one seems to be grasping at straws though, and points to a lack of knowledge and/or experience with overclocking, or as others mentioned, he is just playing dumb and being deliberately inflammatory to get views. 

 

As someone who overclocks, I'm aware of the many variables at play - motherboard selection, VRM quality, LLC settings, and also how strenuously do you test for stability? Do reviewers test stability the same way as Silicon Lottery? I highly doubt it. 

 

Finally, what are the odds he will go to the same effort and see if AMD sends out cherry picked samples too? I would say almost zero, because he would lose practically all his subs if he does

 

This isn't journalism, this is agenda based content. Know the difference.

[Lays]

On 4/1/2018 at 3:12 PM, leadeater said:

I'm not saying they don't, I'm saying they are not testing tens thousands of CPUs to find the best ones to send to thousands of reviewers because globally there is a lot.

 

Like I said originally binning for product SKU and binning for OC is very different, the K SKU is literally the golden dies that people are talking about.

That's actually essentially what I did while I worked there.  Not ten's of thousands, but definitely thousands.

 

Reviewers aren't sent the best chips we have, those are usually reserved for OC events to break world records etc. Reviewers were receiving like B quality chips whilst A quality were saved in house.