I9 9900k

[SoldierRBT]

Here are the lowest voltage (after vdroop) pass on cinebench with the Asus Maximus XI Apex (LLC7). Memory was stock clocks (3600MHz CL16)

5.0GHz 1.074v



5.3GHz 1.217v

 

[Distracted]

That's some low vcore!

Have you tried Cinebench r20? There's a thread in the cpu section that is recording results and there are several 9900k on there for comparison.

https://forums.overclockers.co.uk/threads/the-official-ocuk-cinebench-r20-benchmark-thread.18849380/

 

[woppy101]

Here are the lowest voltage (after vdroop) pass on cinebench with the Asus Maximus XI Apex (LLC7). Memory was stock clocks (3600MHz CL16)

5.0GHz 1.074v



5.3GHz 1.217v

Wow that vcore, when did you get that chip?

 

[moorhen2]

Those voltages are not the load voltages, we need to see what they are really. 5.3ghz at 1.217 seems very low to be honest.

 

[Silent_Scone]

The package draw is rather low, so I'd imagine the voltage is, too. Nice chip.

 

[Robert896r1]

@Silent_Scone When we were discussing RAM a couple of weeks ago, you'd mentioned that for 24/7 usage, I should be stabilizing 1T timings. Can you please elaborate on that a bit? RAM info is hard to find so I'm trying to learn what the impact of 1T vs 2T is. If there's good reading you can point, that'd be appreciated also.

 

[SoldierRBT]

Wow that vcore, when did you get that chip?

I got the chip from Amazon US a couples of months ago. It's a pretty good chip but I've seen better than mine.

Those voltages are not the load voltages, we need to see what they are really. 5.3ghz at 1.217 seems very low to be honest.

Those are the voltage under load. For the 5.0GHz pass it was 1.10 Vcore in BIOS LLC7 and for the 5.3GHz I believe it was 1.255 Vcore and 1.217v under load with LLC7. It also passed 1 hour of Realbench 5.1GHz at 1.154v under load with LLC6. I'll upload a photo later today.

Edit:

 

[tyler_jrb]

There won’t be many better if those voltages are stable trust me.

Mine does 5ghz stable on prime small fft’s at 1.15v, .025v lower and it crashes.

I’ve used realbench/aida 64 in the past but I tend not too now, have been stable on realbench, crashed on prime. Subsequently crashed in rendering.

But that is a 1/1000 chip ,very nice either way .

 

[moorhen2]

I got the chip from Amazon US a couples of months ago. It's a pretty good chip but I've seen better than mine.



Those are the voltage under load. For the 5.0GHz pass it was 1.10 Vcore in BIOS LLC7 and for the 5.3GHz I believe it was 1.255 Vcore and 1.217v under load with LLC7. It also passed 1 hour of Realbench 5.1GHz at 1.154v under load with LLC6. I'll upload a photo later today.

Edit:

But CPUZ is showing 1.154v under load, so it cant be 1.10v in the bios with LLC 7.

 

[SoldierRBT]

But CPUZ is showing 1.154v under load, so it cant be 1.10v in the bios with LLC 7.

Completely different tests and clockspeeds. The 1.154v under load is Realbench 1 hour 5.10GHz all cores and the 1.074v under load is cinebench 5.0GHz. Here is another photo of same cinebench 5.0GHz test with HWiNFO64

 

[Danny75]

@SoldierRBT

The Vcore you are looking at isn't right. The same reading (slightly different location for my board) shows incorrect as well (marked in red in below pic). The correct Vcore readings are marked in green:



You are using 1.169v for the 5.0GHz run, or very close to what the VID is.

 

[SoldierRBT]

@SoldierRBT

The Vcore you are looking at isn't right. The same reading (slightly different location for my board) shows incorrect as well (marked in red in below pic). The correct Vcore readings are marked in green:



You are using 1.169v for the 5.0GHz run, or very close to what the VID is.

You're wrong. VID and VCORE values are voltages at which a processor is designed to operate at within in it's frequency range. The processor communicates VID to the motherboard's VRM, which in turn delivers that correct voltage (Vcore) to the processor at the given frequency/moment. VID voltage varies with frequency. The VID is programmed into the specific processor. You can have two of the same processors with different VID and the lower the VID is an indication of the quality of the processor and generally the lower the VID means better/more stable overclocking. VID is the voltage that is called for by the processor and Vcore is the voltage that is applied.

In other words, VID is voltage level requested by the CPU to the motherboard's VRM to supply it and VCore is the Actual voltage delivered/supplied by the motherboard to the CPU.

I don't know about your readings but your HWinNFO is outdated. The newest version is 6.03-3710

 

[Danny75]

You're wrong. VID and VCORE values are voltages at which a processor is designed to operate at within in it's frequency range. The processor communicates VID to the motherboard's VRM, which in turn delivers that correct voltage (Vcore) to the processor at the given frequency/moment. VID voltage varies with frequency. The VID is programmed into the specific processor. You can have two of the same processors with different VID and the lower the VID is an indication of the quality of the processor and generally the lower the VID means better/more stable overclocking. VID is the voltage that is called for by the processor and Vcore is the voltage that is applied.

In other words, VID is voltage level requested by the CPU to the motherboard's VRM to supply it and VCore is the Actual voltage delivered/supplied by the motherboard to the CPU.

I don't know about your readings but your HWinNFO is outdated. The newest version is 6.03-3710

I knew that Vcore operates within VID range already so I agree with what you posted. But on full load the Vcore will be equal to, or very close to the VID. Which was my point (not that Vcore is always same as VID). And since my HWiNFO64 shows a wrong "Vcore" reading, the obvious assumption is that it's wrong on yours too. The HWiNFO64 version is new enough that it doesn't alert me yet to update. But I will get the same version you are using and check. Meantime, why don't you run the same test again, with CPU-Z open, and grab a screenshot of the Vcore it's using while running the test?

 

[Danny75]

Just checked. Version 6.03 is a beta version released 20th March. 6.02 is the official. That's why it hadn't bugged me to update, I have beta update alerts disabled.

 

[Danny75]

Beta still showing a wrong Vcore and a right Vcore:

 

[SoldierRBT]

I knew that Vcore operates within VID range already so I agree with what you posted. But on full load the Vcore will be equal to, or very close to the VID. Which was my point (not that Vcore is always same as VID). And since my HWiNFO64 shows a wrong "Vcore" reading, the obvious assumption is that it's wrong on yours too. The HWiNFO64 version is new enough that it doesn't alert me yet to update. But I will get the same version you are using and check. Meantime, why don't you run the same test again, with CPU-Z open, and grab a screenshot of the Vcore it's using while running the test?

Looks fine to me.

 

[Danny75]

Looks fine to me.

It might end up looking fine to me as well. Thing is the bench isn't running. Grab a screenshot of CPU-Z over Cinebench while it's running please.

 

[Distracted]

Vcore displays correctly in hwinfo for asus boards.

I know some other boards use vrout and that is considered the most accurate vcore reading, the asus equivalent as of the maximus 11 series is vcore and is just as accurate.
For those boards that use vrout they have a weird bug that is displaying dummy values as well as accurate readings, which is probably what you are seeing on your hwinfo Danny.

Tldr. Asus maximus 11 series boards display vcore correctly in hwinfo.

 

[Danny75]

Vcore displays correctly in hwinfo for asus boards.

I know some other boards use vrout and that is considered the most accurate vcore reading, the asus equivalent as of the maximus 11 series is vcore and is just as accurate.
For those boards that use vrout they have a weird bug that is displaying dummy values as well as accurate readings, which is probably what you are seeing on your hwinfo Danny.

Tldr. Asus maximus 11 series boards display vcore correctly in hwinfo.

Ah ok, cheers Distracted. And sorry SoldierRBT, that wrong reading on mine threw me off.

One other thing you guys might know that I don't - why isn't Vcore hitting VID level on full load? Is it because of the LLC keeping a bit of the Vcore for itself?

 

[Distracted]

Some updated stats on the silicon lottery for the 9900k.

Spoiler: Stats 17/11/18

Some stats taken from silicon lottery. Keep in mind their sample size is small and may not represent your own chances when buying a 9900k.

I think this is useful info for those overclocking. No info yet on why those particular voltages were chosen or any temperature data.


As of 11/17/18, 100% of tested 9900Ks were able to hit 4.8GHz or greater.

• CPU Multiplier: 48
• BCLK: 100.0
• CPU Vcore: 1.275V
• AVX Offset: 2

As of 11/17/18, the top 82% of tested 9900Ks were able to hit 4.9GHz or greater.

• CPU Multiplier: 49
• BCLK: 100.0
• CPU Vcore: 1.287V
• AVX Offset: 2

As of 11/17/18, the top 46% of tested 9900Ks were able to hit 5.0GHz or greater.

• CPU Multiplier: 50
• BCLK: 100.0
• CPU Vcore: 1.300V
• AVX Offset: 2

As of 11/17/18, the top 14% of tested 9900Ks were able to hit 5.1GHz or greater.

• CPU Multiplier: 51
• BCLK: 100.0
• CPU Vcore: 1.312V
• AVX Offset: 2

There are no 5.2GHz chips being offered and the settings and frequencies above are with non delidded CPUs.

With regards to delidding they say that "Peak core temperatures under a heavy overclocked load typically decrease anywhere from 4°C to 8°C for Intel 9th generation CPUs."

Takeaways from this, 1 in 5 9900ks tops out at around 4.8GHz.

More than 50% of 9900ks wont reach 5GHz(!)

I imagine those numbers will change with a larger sample, and if the voltages were changed then more would hit higher frequencies, but presumably heat is a problem otherwise the vcore they are using would be set higher.

The biggest shock to me is the 5GHz percentage. The numbers lend credence to those reviewers that complained about temps and poor overclocks.

Spoiler: Stats 07/12/18

Some updated figures regarding the silicon lottery.

Keep in mind their sample size and that it may not represent your own chances when buying a 9900k.

I think this is useful info for those overclocking. No info yet on why those particular voltages were chosen or any temperature data.

As of 12/07/18, 100% (no change) of tested 9900Ks were able to hit 4.8GHz or greater.

• CPU Multiplier: 48
• BCLK: 100.0
• CPU Vcore: 1.275V
• AVX Offset: 2

As of 12/07/18, the top 85% (+3%) of tested 9900Ks were able to hit 4.9GHz or greater.

• CPU Multiplier: 49
• BCLK: 100.0
• CPU Vcore: 1.287V
• AVX Offset: 2

As of 12/07/18, the top 41% (-5%) of tested 9900Ks were able to hit 5.0GHz or greater.

• CPU Multiplier: 50
• BCLK: 100.0
• CPU Vcore: 1.300V
• AVX Offset: 2

As of 12/07/18, the top 11% (-3%) of tested 9900Ks were able to hit 5.1GHz or greater.

• CPU Multiplier: 51
• BCLK: 100.0
• CPU Vcore: 1.312V
• AVX Offset: 2

Silicon lottery now have a larger sample size and as such the percentage of silicon meeting their requirements for each bin has changed.

As of 07/12/18 100% of 9900ks can hit 4.8GHz all core, while there has been an increase in the number of cpus able to hit 4.9GHz by 3% taking the number to 85%.

That's the end of the increases.

For 5.0GHz all core the percentage has dropped by 5% down to 41%. Almost 60% of 9900ks wont reach 5.0GHz all core, or roughly 3 out of every 5 9900ks are failing to reach 5.0GHz all core at the settings listed above.

Next a drop of 3% down from last month bringing the percent of 9900ks able to reach 5.1GHz all core to just 11%.
Just over 1 in 10 of the samples they are binning can hit 5.1GHz stable all core at their specified settings.

As always with these numbers there is no telling how representative of the 9900k supply as a whole the stock silicon lottery receives is. However as they bin more CPUs the numbers should more accurately reflect the silicon lottery for those considering buying retail or tray.

Finally, still no 5.2GHz bin listed. This doesn't mean they don't exist, there are probably 5.1GHz samples that will do 5.2GHz, but judging by the fact that no one is offering binned 5.2GHz 9900ks (when there are clearly people willing to pay for it) the number of stable 5.2GHz chips must be tiny.

Spoiler: Stats 06/02/19

Some updated figures regarding the silicon lottery.

Keep in mind their sample size and that it may not represent your own chances when buying a 9900k.

I think this is useful info for those overclocking. No info yet on why those particular voltages were chosen or any temperature data.

Bins are created for stability in nearly every workload and tested using a 240AIO. These percentages can be considered accurate, conservative or optimistic depending on your own expected use and cooling.

As of 2/6/19, 100% (no change) of tested 9900Ks were able to hit 4.8GHz or greater.

• CPU Multiplier: 48
• BCLK: 100.0
• CPU Vcore: 1.275V
• AVX Offset: 2

As of 2/6/19, the top 86% (+1%) of tested 9900Ks were able to hit 4.9GHz or greater.

• CPU Multiplier: 49
• BCLK: 100.0
• CPU Vcore: 1.287V
• AVX Offset: 2

As of 2/6/19, the top 39% (-2%) of tested 9900Ks were able to hit 5.0GHz or greater.

• CPU Multiplier: 50
• BCLK: 100.0
• CPU Vcore: 1.300V
• AVX Offset: 2

As of 2/6/19, the top 8% (-3%) of tested 9900Ks were able to hit 5.1GHz or greater.

• CPU Multiplier: 51
• BCLK: 100.0
• CPU Vcore: 1.312V
• AVX Offset: 2

their sample size increases both 5.0GHz and 5.1GHz samples again show a drop making them even rarer while 4.9GHz has become slightly more common.

A bin for 5.2GHz or higher is still noteably absent for the 9900k unlike the 9700k which had the 5.2GHz(10%) bin added at the end of last year/beginning of 2019.
This shows that silicon lottery is willing to update their bins when a new frequency becomes available in sufficient quantities, and that even with an AVX offset, 5.2GHz+ samples are in such short supply that there aren't enough that meet their criteria to be considered for sale as yet.
This mirrors the listings from both Caseking and OCUK that sell frequencies up to 5.1GHz with AVX offset but nothing higher.

As of 3/16/19, 100% (no change) of tested 9900Ks were able to hit 4.8GHz or greater.

• CPU Multiplier: 48
• BCLK: 100.0
• CPU Vcore: 1.275V
• AVX Offset: 2

As of 3/16/19, the top 89% (+3%) of tested 9900Ks were able to hit 4.9GHz or greater.

• CPU Multiplier: 49
• BCLK: 100.0
• CPU Vcore: 1.287V
• AVX Offset: 2

As of 3/16/19, the top 38% (-1%) of tested 9900Ks were able to hit 5.0GHz or greater.

• CPU Multiplier: 50
• BCLK: 100.0
• CPU Vcore: 1.300V
• AVX Offset: 2

As of 3/16/19, the top 8% (no change) of tested 9900Ks were able to hit 5.1GHz or greater.

• CPU Multiplier: 50
• BCLK: 100.0
• CPU Vcore: 1.312V
• AVX Offset: 2

Silicon Lottery has a very large sample size now and what we are seeing is that there are more 9900ks able to hit the lowest bin which would make sense if the yields and process were improving over time as expected. For the top two bins there is little change which could mean that the percentages are stabilising and giving an accurate representation of the state of the lottery right now. There is still no 5.2GHz or higher bin meaning that the percentage of silicon able to achieve these clocks to their standards is incredibly low.

One point of interest for anyone reading this, the voltages listed for each frequency are the voltages displayed as vcore during use, NOT the voltages that one sets in BIOS.

As always please keep in mind that silicon lottery bin for stability in a wide variety of workloads while using 240AIOs or thereabouts. This means that their figures could be considered accurate, conservative or optimistic depending on your own workload and cooling.