Intel Core i9-9900k 8c/16t, i7-9700K 8c/8t, i7-9600k 6c/6t 2nd Gen Coffee Lake CPU's + Z390

They have it defined for p1 to be 95W TDP. The entire point of this and people talking about this is that you cannot honor both. Because of that, you have variance on the basis of where you buy your machine, cooling solutions, and for custom builds you have MB mfrs ignoring the TDP rating of the chip. I'm sure you remember Sandy Bridge throttle down after a period of time, and I damn well know you know how to set the boost period higher than the 100 sec max of spec, or locking all cores above that.

Considering some marketing materials say Intel uses less TDP than AMD as a selling point, when it actually uses more when used by consumers, especially in server settings or as anyone using normal MB settings or using MCE, then the TDP is misleading. Period. Just because they allow partners to exceed it doesn't make it not a spec.

 

TDP depends on your luck and silicon lottery chip. Better Si generate less heat and deliver better power on air.
These days going just from TDP is simply BS, they measure TDP at certain workloads which we never do at all. Same is true for Nvidia GPU.
I really like AMD GPU since I can push them more beyond target TDP.

 

The only time TDP is "bad" is when the number is too small to be awesome and/or it is locked and cannot be exceeded, and performance is limited due to lack of it. The only time TDP is "good" is when it is unlocked and unlimited. Otherwise, it's just a number that is neither good nor bad, and there are far more important things worthy of being cared about. Likewise, I think any marketing that represents AMD TDP is "too high" compared to Intel is also a lame excuse from Intel fanbois. If that is the best they can come up with they need to find something else to be a troll about, because they are not doing themselves or Intel any favors with their ludicrous nonsense. Putting too much focus on that instead of results is retarded no matter who is doing it. Stupid is always stupid, and it is not a respecter of personal bias or special interests.

 

Hey, what are you worried about? You've got the Asus Maximus Hero XI with the golden 95w TDP profile, the out of the box "turdboard" of z390's.

Your CPU is safe from overheating, and outperforming the 2700x...

All of those poor All-Core-4.7ghz Profile motherboards need to use 360mm-420mm water cooling to reign in the thermals.

Then again the 2700x runs at spec, and is 70% cheaper while being more than fast enough to not have a noticeable difference in everyday use against either profile z390 running a 9900k.

FTFY

Cost vs performance is still with AMD vs 9900k @95w or @295w.

 

Can you guys recommend me a nice MATX mobo for the 9900K?
Even Mini-ITX would suffice, but I need the 4 RAM slots.

I'll probably go with the Asus TUF Z390M Pro. Seems like it's the one with the least bells and whistles for the price, which should mean more money was spent on higher quality components, one would hope.
The MSI Z390M has too much stuff I don't need. Also haven't owned a MSI motherboard in over 15 years.
Gigabyte has a nice looking Z390M model which boasts 10-phase power modulation on the CPU and RAM, slightly better than the Asus or MSI at 8 phases (I think, some sources say 4 with phase frequency switching which I have no idea what it is but doesn't sound too good.)

I just need a nice, simple board, good overclocking potential and a stable BIOS.

Thanks!

 

Check the thread for reviews of z390's, they are really just getting started reviewing them effectively, more than just powering them up and booting into Windows.

That's where you are going to get the real data, from end users and reviewers actually putting the boards through it's paces.

Try posting a request in an " Actually Hardcore Overclocking" motherboard review video, or one of the other channels covering the z390's and ask them nicely to review the MATX boards you are interested in buying.

Maybe even donate a buck or two via Patreon to them to get their attention, I've seen that work many times.

The Asus boards come with the 95w profile out of the box, unlock it for full performance, but cooling will be a problem. I generally prefer the Asus BIOS and motherboards, but have used the other brands without issues after a few BIOS updates.

The Gigabyte boards as Aorus are getting the best reviews, but the BIOS even though recently re-done is still not getting good comments.

https://www.google.com/search?q=matx+z390
https://www.youtube.com/results?search_query=z390+matx
https://www.reddit.com/r/intel/search?q=z390+matx&restrict_sr=on
https://www.overclock.net/forum/6-intel-motherboards/

https://www.reddit.com/r/intel/comments/9m8s7y/which_known_z390_matx_motherboards_are_likely_to/
https://www.reddit.com/r/intel/comments/9qi1p0/which_matx_z390_board_for_i9_9900k/

https://www.overclock.net/forum/6-intel-motherboards/1638955-z370-z390-vrm-discussion-thread.html

Why do you need 4 slots? The new 32GB per socket ram is coming if not here already. Better OC for RAM / CPU with 2 sockets. Unknown how that 32GB ram will OC though...but 4x slots is typically bad for OC.

 

I have a 4 x 8GB setup. All TridentZ 3000MHz. RAM is way too expensive to upgrade or even sidegrade right now. This setup has served me well and I typically don't overclock RAM.

There's really not a lot of stuff around there about mATX offerings.
Seems like the form-factor has taken a nose-dive in terms of number of users in the last few years. Everyone's crazy about mini-ITX right now.

I love mATX setups, always where my go-to motherboards. And I quite like my Phanteks Enthoo Evolve mATX. One of my favourite cases ever. 240mm rad up top, GTX1080TI in the middle, happy gamin'

 

Well, now's a good time to sell RAM and then wait a tick to buy it over the holiday sales. RAM prices have been dropping, and are expected to continue to drop.

 

Yeah I'm hearing about that. Still need to see it though.

If RAM does get cheap then I'll sell mine for cheap and buy new for cheap. I'm not about to sell my RAM and then sit waiting on good deals.
With that said. RAM is the only issue making me NOT go with the new Asus Gene. But if the TUF Z390M has the same core capabilities then I'll buy it gladly. Again, not into lots of bells and whistles. I don't even use onboard audio. Rather the purchase cost go into better quality caps and components.

I'll continue fighting the good fight for mATX setups! Not too big, not too small, just right

 

Z370 / Z390 VRM Discussion Thread
https://www.overclock.net/forum/6-intel-motherboards/1638955-z370-z390-vrm-discussion-thread-315.html

Telstar
"Add the Gene, which is not available in USA. EU prices are around €330."

Posted by AlphaC:
Quote: Originally Posted by SpeedyIV
" I don't get why Asus would put this beefy VRM on the Gene but not on the Code or Formula (or Hero).

I also don't get why they would not plan to sell the Gene in the USA. Seems like it is the only alternative to the Extreme (and Apex?) that does not have the new Asus 4-Phase ("Twin 8-Phase") VRM design. I suspect there will be a lot of people in the US market that will want a better VRM but can't or don't want to spend the $ on the Extreme (or the Apex). I guess the jury is still out on this new VRM design used on the Xi Hero, Code, and Formula."

"There's an easy explanation: ROG Gene is a 2 DIMM board also meant for overclocking on LN2, ROG Formula/Hero/Code are boards meant for everyday use.

Buildzoid PCB Breakdown: ASUS ROG Maximus XI Gene
https://www.overclock.net/forum/6-i...0-vrm-discussion-thread-315.html#post27708566

"A typical Asus XOC board but more expensive (310 GBP) than before
Dual 8-pin EPS, only one needed except for sub-ambient OC
More accurate voltage reading on Maximus XI with redesigned monitoring circuitry:
https://www.overclock.net/forum/27686004-post2664.html

VRM:
Better than Maximus 9 & 10 Apex
PWM ASP1405I (5+2), 5-phase Vcore, 2-phase Vgpu,
10+2 IR3555 60A PowerIRstage: OTP, current monitoring, Body-Braking mode, Diode emulation mode.
No doublers: no current balancing, but have a sort of self-balancing mechanism between a pair of mosfets based on temp, Rds(on) and current.
Good efficiency: Fsw 400kHz, Vout 1.3V, 5Vdrv- Iout 150A, power loss 17.5W
- Iout 200A, power loss 21.5W
- Iout 250A, power loss 28W
- Iout 300A, power loss 36W
- Iout 350A, power loss 45W
- Iout 400A, power loss 58WVccsa: PWM APW8723 (1) at 300kHz, 2x 4C10B (1H1L)
Vccio: TI TPS51362 10A integrated FET converter
Vmem: 2-phase, redesigned output filtering with extensive amount of 150uF caps

Debugging features:
MemOK switch useless
Color-coded troubleshooting LED
LN2 mode jumper: lifts all security limits
Slow mode switch: forces CPU running at 8x ratio
RSVD switch (for LN OC): removes cold bug
Retry button
Safeboot button
Probe-It connector: dumb design
Clear CMOS & BIOS Flashback on the rear panel
Last edited by eric98k; 11-09-2018 at 12:13 PM."

http://forum.notebookreview.com/thr...-lake-cpus-z390.811225/page-110#post-10818636

 

Your CPU is safe from overheating, and outperforming the 2700x...

All of those poor All-Core-4.7ghz Profile motherboards need to use 360mm-420mm water cooling to reign in the thermals. [/QUOTE]

What? In English next time.

https://overclockers.ru/lab/show/94...fi-luchshaya-model-dlya-coffee-lake-refresh#9

Philosopher's Stone found. The processor’s impressive appetite is quietly digested by the motherboard’s power system.

Peak power consumption with VRM losses is more than 300 watts. And eight cores 9700K at 4.9 GHz cannot even be cooled by a custom dropsy. The processor is really hot. But the result is obvious, VRM without cooling only heats up to 84 ° C.

Pros ASUS ROG Maximus XI Hero (WI-FI):

• Ready set of settings for overclocking any level;
• Extended voltage setting range, including PLL Voltage;
• Separate Load-Line Calibration setting for main acceleration, AVX block and Turbo mode;
• Exact automatic tuning Load-Line;
• Eight fan connectors;
• Ten phases of processor power;
• The ability to withstand loads of more than 300 W without forced blowing;
• Open PCIe 1x slots;
• Lack of cyclic restart during acceleration and cold start.

 

^^^ This. By quoting both a "95W TDP" and letting board partners drive it well beyond that by default, they are deceitfully mixing claims of efficiency AND performance and most reviewers are too superficial to delve into what is actually going on in their tests.

Board partners are all in a race to tweak performance relative to each other with power limits and bclk to be "the fastest" in testing

It's all set up to mislead and induce a buyer into buying something that will then be unfit for purpose. If they buy a 95W capable cooler for their 9900K... They're gonna have a bad time. Misleading and deceptive advertising is supposed to be illegal under consumer law.

 

Beating the Intel 9900k Liquid Helium CPU World Records using Liquid Nitrogen.. Possible? Lets Try..
Bearded Hardware
Started streaming 51 minutes ago
My buddy Jason aka (shadyreaper) was blessed by the Viking Tech Gods, he found a chip so good that we will come close to beating the Liquid Helium cooled CPU World Records. Liquid Nitrogen goes to -196C, Liquid Helium goes to -230C.
2nd session - restarted after internet drop...

1st session - internet interrupted

Spoiler: Click to See Video

 

Some of you folks might be interest in this. Free is a really good price.

 

Now this is a guy I would listen to. He knows his ****.

 

My i9 arrived today and a 1 TB 970 evo arrives later.
A Vega 64 will probably arrive Friday. Then I need to wait for some deals on DDR4 RAM and a Gigabyte Aorus Master and then find if I want to use a Noctua DH-15 heatsink or something else. For now I'll have fun rolling a d12 around.

BGA turdbook for posterity because I don't have a cat to look at my .....8 core pretending to be a 12 core CPU?

And those max temps were from a 4.7 ghz small FFT AVX disabled prime95 run @ 1.276v, because BGA must die. Not from current clocks/volts.

 

I'm a current fan of air cooling (given I have NH-D14), so I'd be interested to see how your i9 gets along with the NH-D15.

Some quick thoughts on your build if you go NH-D15:

• try to have NH-D15 exhaust lined up with exhaust fan as best possible so heat goes straight out of the case. 100% CPU load doesn't increase case air temperatures in my case at all because of this - it just goes straight out the case.
• NH-D14 & NH-D15 are massive tower heat sinks that take up a lot of space in the case near the exhaust fan. I have a feeling it can block hot exhaust air from the GPU from leaving the case to some extent. Best results for me came from having 2 exhaust fans: one on the rear panel in the usual place, and then another exhaust fan in the roof sitting directly above the NH-D14 (towards back of roof panel). This helps get rid of that hot GPU exhaust air. I haven't tested GPU temperatures with NH-D14 vs a closed loop liquid cooled CPU (because I don't have the latter), but this is just my intuition on the subject. (Biggest positive effect on GPU temperatures though is having an intake fan that feeds directly to the GPU fans with no impedance or seperation to other areas).

 

Are Budget Z390 Motherboards a Scam? VRM Thermal Test
Hardware Unboxed
Published on Nov 14, 2018

 

What's going on everyone, been a while since i posted but figured it was time since i built a new rig.

Just wanted to share some temperatures with my new rig and 9700K. Only ran Prime95, V27.9, Build 1 for 13 minutes as temperatures reached near throttling. I have a second EK 240mm PE Radiator I am going to add to the loop in the front of the case on the way and will re-test to see if it makes much of a difference in Prime95 once it is set up. Playing COD Black Ops 4 for 2 hours the CPU reached a max of 71 C which is fine with me but would like to get the temps down some and am hoping that second rad will help. Let me know if you all have any suggestions or requests to additional settings. I had to run Vcore at 1.35 (was the minimum), ran a blend of Aida64 for a few hours and was in the low 80s. I also had to run CPU VCore Loadline Calibration on turbo as I saw crashes on auto or lower settings. This processor did come from Silicon lottery though and was tested there at 1.35 on there rigs as well. I am currently running the D5 pump at 3096 RPM (60%) and the EK Vardar fans at 1305 RPM (50%) which was about my personal tolerable range from where the computer sits. Computer fans are running around 1200 RPM.

System:

Case: Fractal Arc Midi R2 with black window with 2-140mm (Front), 1-140mm (Bottom), 1-140mm (Back) and top 2-120mm with 240mm EK PE Radiator
Power Supply: Corsair AX860
Motherboard: Gibabyte Z390 Aorus Master
CPU: I7 9700K @ 5.0 GHZ (Delidding with Conductonaut (73 W/m·K) – Gelid GC-Extreme between block and processor.
Memory: G.Skill TridentZ Series 32GB (2 x 16GB) - F4-3200C14D-32GTZSW - Timing 14-14-14-34
Hard Drives: Samsung 970 Evo M.2 1 TB & Samsung 870 Evo M.2 500GB
Video Card: MSI Geforce RTX 2080 Gaming X Trio
Cooler – Open Loop, D5 Pump, EK 240mm PE Radiator in top of case

Bios Setting:

XMP Enabled – 3200 - 14-14-14-34
CPU clock -5000
Uncore – 4700
AVX Offset: 2
CPU VCore – 1.35
Disabled – Speed shift technology, CPU Enhanced Halt (C1E), C3/C6/C7/C8/C10 State Support, Ring Core offset (Downbin) & VT-d.
CPU Vcore Loadine Calibration – Set to Turbo

 

Been a while but does not look like those images posted. If you share the URL from imgur through the image icon they are supposed to post correct? Never mind think i figured it out.

 

Congrats on the build, it's a tasty set of components, CPU/GPU! You say you're running 1.35V Vcore, but in reality according to HWInfo you're all the way up to 1.416V, so I'd actually say you're at 1.416V rather than 1.35V, because that's worst case scenario and you'd probably want to base your max usage of voltage around the highest peaks seen. 1.416V is quite high, if I was you I'd run something like 1.37V as a maximum peak. I ran 1.4V on my Skylake 6700K and saw 10mv of CPU degradation over a period of 1 yr. I've since dropped my VCore to 1.368V and backed down to 4.65Ghz. I don't think you'd see any degradation at 1.37V maximum. As a bonus this will lower your power consumption & therefore lower your temperatures. You've got your CPU delidded, so that's ideal already.

(Why are there 2 lines/entries for VCore in HWInfo, one reads higher than the other?)

 

Thanks, I need to figure out how to dial it in a little bit better. Here is a photo of what the two different lines represent. I am unsure why they are listed twice but the voltages are different. One is listed as (ITE8688E) and the other (ITE IT8792E).

 

Brought my loadline from Turbo to high and going to see if i can get it stable with some more Tweeking, that should help bring the vcore down some.

 

Two different sensor chips. The GB rep replied on OCN about that.
I believe the second one is a bit closer to direct multimeter readings.

 

Here was a second test. Looks like this should be better, showing max at 1.368. Temps are still high but came down as i ran this test for over 2 hours and they were still a little lower than the quick 15 minute test. Will be interesting to me to run the same test when i get the additional radiator in and see what kind of difference the temps are going to make.

 

That looks a lot better in terms of a safer long term VCore, and by the looks of it you're at the same CPU frequency, and temperatures have come down a bit too. One thing that concerns me a bit is that the VID for the cores (far left side of your HWInfo Sensors) is at 1.44V as a max. Does that mean you're getting vdroop at load, and your VCore is actually higher at around 1.44V when the CPU is only lightly loaded and/or idle? If that's the case then I wouldn't be comfortable with that higher voltage when lightly loaded, and that will be the majority of the uptime for your PC too. You'd be able to tell by just leaving HWInfo Sensors open while doing light stuff like web browsing, virus scanning, etc, and then see if your VCore ever went up the the 1.44V that I was concerned about? I'd try to tweak the voltage setting and loadline calibration so that in any given situation VCore would never go above 1.37V, that would be my goal.

 

Thank you again. Tweeking the loadline calibration from turbo to high to medium does not really change the outcome of the maximum VID. Which voltage should i be tweeking to try and get the maximum VID down to and what should be my high level while doing normal tasks? I looked around the bios but was not finding what i needed although this bios is kinda messy.

 

I was reading changing IAC and IDC to 1 can reduce the core VID. I am trying that now as well as lowering my Vcore a little bit. So far it is is running in Prime 95 at max 1.319 core VID. Is this more in line with safe? i will post after i have run prime for an hour or so.

 

Starting to get a little safer now i believe. Thanks for the help!! Do not want to degrade the chip in the process. I will run some aida64 tests and run prime95 longer when i get some time. I ended up changing IAC and IDC to 1 (not exactly sure about this setting but it seems to lower the Core VID), bumping down core voltage in bios to 1.34 and changing CPU Loadline Calibration from turbo to high. The change in settings netted a max 1.324V on the core VID and 1.344 on the Vcore.

 

Core VID is not core voltage.
The VID is only used when using adaptive (may be called DVID on your motherboard) voltage, as the VID is the target voltage that adaptive voltage uses.
Setting a manual voltage overrides the VID with cpu vcore settiing.

On MSI jokebooks, the IA AC DC setting affects the VID and the cpu vcore (which is not shown as there is no vcore sensor), that is why its required to set AC DC loadline to 1 on MSI jokebooks to stop the VID overboosting. It should not be necessary on desktops unless using adaptive/dynamic voltage rather than static.

Did your cpu vcore, power draw and temps go down and change when changing IA AC DC loadline from auto to 1 (without readjusting cpu vcore manually?)

 

They went down about 2 to 3C but that could have just been within margin of error. The only difference i saw was a drop in voltage or at least what is being reported (which shows to have dropped from 1.44 to 1.324 with setting IA AC/DC from auto to 1). Should i set it back to auto and try and find the setting to manually set the VID to around 1.3 or 1.35? not really sure what i should shoot for. I was reading on overclock.net for gigabyte to leave IA AC/DC on auto but then in the other menu to change it from 0 to 1 for both and then once you find your fixed voltage to use VID+/- offset to match the setting for core voltage. This i am guessing is for adaptive though like you stated above? Thanks

 

It's your motherboard. Don't be afraid to experiment.
Pay attention to POWER DRAW (watts) and CPU VCORE when you set it back to auto.

NEVER EVER EVER under ANY circumstances go anywhere near the higher regions for IA AC DC. Ever. The reference value i think is 1.60 mOhms, which in MSI and Gigabyte Bioses may be a value of 160 (check the divider) while on Asus it may be 16. Assuming the range is 0 to 62.49 on asus, and 0 to 6249 on MSI/GB.

 

As Falkentyne said, VID and Vcore are 2 seperate things. The VID is just a guideline for the motherboard to 'work from' and isn't what is being fed to the CPU, VCore is what you need to focus on & is the voltage that is being fed to the CPU. I only commented on the high VID being a potential indicator of a high VCore during idle seeing as you VCore was ok at load, I figured you had vdroop, that's why asked you what your VCore was during light loading when idle or when doing stuff like web browsing - because you don't want your VCore to ever go over 1.37V (at least in my view/experience). So, what's your max VCore when idling or web browsing? Forget about all the VID tweaking nonsense you seem to have started doing, go back to what you were doing before with VCore and loadline calibration. Your last setting that I replied to where you had a max VCore of 1.368V was fine, but I just want to ask you if your VCore goes above that when the CPU is lightly loaded?

 

With those settings i must have had a vdroop as when doing standard surfing, etc my readings were peaking around 1.410 at times. That is why i went to i had set AC/DC to 1 was trying to lower the vdroop as i really was not sure how to do so.

 

Thanks, if i use this to tweek the settings i looked and a value of 100=1.00mOhm on this board.

 

Ah, ok, so my suspicion was correct then about the higher voltage during light loading & vdroop at high loading. I've never really had to combat vdroop, as my motherboard does an excellent job of keeping it stable through the load ranges when set to Auto for load line calibration. In your situation, my first thoughts are to go back to the same settings you initially had when we last talked, your VCore 1.368V at load, and then tweak the following:
-try lowering your voltage input: so for me and my PC as an example I use Offset Mode for voltage and use +150mv. I would try lowering that input voltage until you get an acceptable low load voltage, and then I would increase the load line calibration to counteract your v-droop so you were still seeing the 1.368V at high load (which you know to be stable). So you're trying to keep voltage below 1.37V at all times, and trying to keep a relatively flat curve of voltage as load increases. That would be my approach to solving it. (I don't think you'd need to get involved with the whole VID thing.)

It's normal to see slightly higher VCore when running programs like the later versions of Prime95 which use AVX instructions. If you use an older version of Prime95 v26.6 which doesn't use AVX then it won't kick up the VCore as much. This is all assuming the load line calibration is working properly and you don't have vdroop. CPUs ask for a few notches more voltage when running AVX instructions. Mine goes from 1.344V on non AVX Prime95 up to 1.368V for AVX Prime95. So, depending on what you're doing the voltage isn't gonna be completely flat.

 

Thanks, will have to figure out what the setting is called in gigabyte in order to use the offset mode as i could not seem to locate it in the bios.

 

I don't think you have to use the offset mode, I was just saying to reduce the input voltage and increase your load line calibration. Some people use a manual static voltage, others use offset or adaptive voltage modes. After doing research I decided on Offset Mode, Adaptive seemed 'too harsh' with it's voltage control, and manual voltage I didn't like the idea of the voltage not decreasing during idle.

I use:

• CPU Ratio Mode: Dynamic
• Offset Voltage
• Loadline Calibration Auto
• C-states enabled with CO as the minimum Package state: which allows the CPU cores to enter sleep states but the package remains at CO, which is still very responsive but low idle power.
• Windows High Performance Power Plan: which has Minimum & Maximum CPU clocks set to 100%, which means the frequency stays at max all the time, no downclocking. But the C-states enabled above allow for the low idle power consumption & also allow the voltage to decrease at idle. I did some testing, and for me there was zero performance hit with the combination of C-states enabled combined with High Performance Power Plan, it was the most efficient option.

In your situation I would be tweaking the input voltage (Offset Voltage or Adaptive Voltage or Manual Voltage - whichever one you choose to go with), and tweaking the Loadline Calibration settings. Given your description of your CPU behaviour, I would lower the input voltage & increase the load line calibration - hopefully giving you a relatively flat voltage curve at low load & also at high load.

 

Ok i understand what you are saying now. I will give it a try. I already tried increasing the load line calibration a few settings with the same vcore but it seemed to have minimal impact on the vdroop i will give it another shot though. Thanks.

 

(You could try Offset Mode to see if loadline calibration somehow works better with that mode, that's if you hit a dead end with the load line calibration testing on your current voltage mode. I don't know if it will, but might be worth a try.)

 

Well increasing loadline and decreasing vcore did not show any improvements, nor did using offset mode. Offset mode showed about the same results. If i downclock the ratio from 50 to 48 without changing anything the VID drops to a tolerable range. Maybe i will try and jump on some gigabyte threads and see what other changes i can make.

 

Yeah, see what you can find out from your motherboard specific threads. But, you mention VID, don't get hung up on that, it's not important in & of itself, it's the VCore that you need to be paying attention to, like we mentioned earlier.

 

Left a message there last night, still waiting on response but i did read on reddit where buildzoid said "Vcore to 1.35V and LLC to turbo" was the best spot for settings (he did a review on youtube). So if i can get a solid OC down to less than 1.36 Vcore. Do you think the random spikes on the core VID will be fine? From what i can tell it is normally around 1.361 but if i close or open a program is seams to peak up to 1.408 and every once in a while a little higher. I am thinking maybe this bios needs some tweeks and hoping future bios will fix the issue.

 

As long as the VCore is not spiking above levels you are comfortable with then you are fine. We talked about maximum allowable peaks of 1.37V, that would be my preferred maximum.

 

https://siliconlottery.com/collections/coffeelake-r/products/9900k51g?variant=15392436093014

Silicon Lottery 9900Ks, and they're sold out. Apparently I have a CPU worth $970 according to them lol. These prices are just stupid. Go get a regular off the shelf 9900k, leave it stock 4.7ghz for the hell of it and save $400.

 

My BGA Killer parts arrive today. I wonder how good or bad my 9900K will be?
Just waiting for the motherboard, B-die RAM and heatsink to arrive sometime today.

 

Congrats! What GPU did you go with? Good luck in the silicon lottery! I hope you got a good one.

 

Rx Vega 64 firestarter edition
@Papusan

 

So i am not really that concerned anymore after our conversation but the answer i got on Gigabyte website was: "The Core VID is only what is being requested. If you are hard set then you only need to pay attention to your Vcore readout". Does this sound correct that this readout is only being requested and not being applied?

 

Ya i bought my 9700k from siliconlottery as well but did not buy the 5.1 version just the 5.0 version. They had a 4.9, 5.0 & 5.1. The 5.0 was just a little more expensive than retail but the main reason i wanted one from there was to pay a little extra for the delid. Worth the price to me but maybe not the much bigger price for a extra 100mhz but all relative to what your willing to pay. Looks like the 5.1 9900k is currently in the top 14 percent (no clue how many he tested) but that is pretty good. I would never have the luck to get one in the top 14 percent if i bought it off the shelf, lol.