Feature Intel's Core i9-9900KF May Overclock Better Than 9900K

[abryant]

Welcome to the Core i9-9900KF. Yes, this is not a typo. The eight-core 16-thread Core i9-9900KF is the newest mainstream processor from Team Intel. Read more here.

ALLEN 'SPLAVE' GOLIBERSUCH

 

[Math Geek]

one day i want a company to send me a ton of samples just cause i'm awesome.

interesting result. out of curiosity are you able to disable/remove the igp power pins from one of the 9900k 's and see if that has the same effect? i'd be curious what that means for the normal chip and what "disabling" the igp in BIOS actually does.

 

[alecuba16]

one day i want a company to send me a ton of samples just cause i'm awesome.

interesting result. out of curiosity are you able to disable/remove the igp power pins from one of the 9900k 's and see if that has the same effect? i'd be curious what that means for the normal chip and what "disabling" the igp in BIOS actually does.

Indeed, he could just put a little bit of loctite on the power pins, this way is easily reversible.

 

[splave]

Indeed, he could just put a little bit of loctite on the power pins, this way is easily reversible.

I'll see what I can do

 

[alecuba16]

I have a quick question, are the KF samples consecutively numbered? (maybe they came from the same wafer?)

Because I'm thinking that maybe the 5 samples of KF were made from a very good wafer. This information could be useful for the article.

 

[splave]

They were a mixed batch but good thinking.

I isolated the igpu pins and unfortunately it results in the motherboard seeing it as a conflict and not booting. One of my contacts predicted this behavior and I can confirm. I appreciate the cool idea though!

 

[alecuba16]

They were a mixed batch but good thinking.

I isolated the igpu pins and unfortunately it results in the motherboard seeing it as a conflict and not booting. One of my contacts predicted this behavior and I can confirm. I appreciate the cool idea though!

Interesting result, thanks for at trying out.

So maybe the motherboard is still powering on/checking the IGPU, and perhaps the differences between K and KF are because the KF is a new revision? I'm speculating right now.

 

[Math Geek]

would likely have to have at least new microcode somewhere that tells it to not expect anything from those igp pins. or however it would work.

but was happy you looked at it. i guess as time goes we'll see more and more benchmarks from the cpu's and can get a better idea what average performance may look like. but then again the small gain seen here is still real world little to get overly excited about unless you're going for pure benchmark numbers.

but the geek in me still gets giddy when i see this type of stuff

 

[Co BIY]

My best guess is that this decision to sell these as KF parts is pretty recent and they are all from recent late in the game production runs that generally have most of the "bugs" worked out and are well optimized. I suspect your larger pool of "K" parts data stretch back into the earlier less optimized runs. I wonder if you could account for date of production in some way if the apparent quality difference would disappear.

 

[littleleo]

I guess a lot of failures were in the iGPU in the Intel CPUs fab, so they decided to sell them rather than chuck them. Which is fine, but charging full price as the ones that have iGPUs isn't fine, in fact another 4 let word comes to mind.

 

[Notorious^]

I wonder if this is the chip MSI just released a BIOS microcode update for on 3/27 for the MEG Z390 ACE board.

 

[Math Geek]

charging full price as the ones that have iGPUs isn't fine, in fact another 4 let word comes to mind.

think this is pretty much the general thought on the issue around here. but in the end, they'll still sell out so not like Intel has any reason to really care.....

 

[littleleo]

think this is pretty much the general thought on the issue around here. but in the end, they'll still sell out so not like Intel has any reason to really care.....

I think these are some of the CPUs that failed to make the full grade, perhapos the iGPU was bad so they disabled it and sell them.

 

[TJ Hooker]

I think these are some of the CPUs that failed to make the full grade, perhapos the iGPU was bad so they disabled it and sell them.

That's normally what one would think when parts of the die are disabled, but it seems hard to believe that Intel would be having yield issues after ~4 years and 3 iterations worth of refinement of their 14 nm process...
Maybe it's because of their shortages, such that even if the defect rate is relatively good they're trying to salvage every last die they can?

 

[littleleo]

That's normally what one would think when parts of the die are disabled, but it seems hard to believe that Intel would be having yield issues after ~4 years and 3 iterations worth of refinement of their 14 nm process...
Maybe it's because of their shortages, such that even if the defect rate is relatively good they're trying to salvage every last die they can?

Well there is a huge Intel CPU shortage so it isn't that hard to believe.