Originally Posted by cletus-cassidy
Helpful post as always. Any thoughts on the VRM throttling Tom's noted here: https://www.tomshardware.com/reviews...ew,5838-2.html
Trying to decide if I should should upgrade my z370 Taichi to a Maximus Formula, as I have a custom loop, and want to get at least a 5.1+ ghz OC.
Tom's hardware USA is terrible at testing things as OCN users have noted in the past and they don't provide any evidence of where their numbers came from or how they derived their load, I've seen inconsistencies between their reviews and every hwbot overclocking team's site (i.e. lab501 , overclockers.ru, overclockers.ua, etc.). You should look toward Tom's hardware Europe (FR / DE) for stuff from Igor Wallossek as he usually uses FLIR thermal image pictures and shows the PCB shots as well as temperature / noise graphs with thermal imagery video.
The takeaways from the Toms hardware "non-article" (using Prime95 small FFT) with a i9-9900K are : all boards reached relatively the same frequency ("129%" - "131%") Z370 Gaming 7 heatsink isn't scaling great with airflow (not surprised), Z370 Taichi heatsink isn't great in general because 158CFM doesn't seem to affect it, Z390 Taichi has voltage fluctuation suggesting immature BIOS, Z390 Pro Carbon tripped Overcurrent protection, Z370 Pro Carbon unusable.
It could just be early pre-embargo BIOS might not have proper power limits. I9-9900k testing is still under embargo which is why I'm hesitant to say anything definitive as Z370 Taichi is a solid board. Basing it mainly off power draw, feature-set, and what was obviously removed for cost savings that is my impression. YMMV depending on BIOS implementation (version and bugs) , chip quality, and motherboard anomalies.
You should also note which version of the Z370 Taichi you have, apparently there's two versions: a Fairchild FDPC5030SG and Sinopower SM7341EHKP version. You can tell by the memory section's powerblock mosfets.
Originally Posted by Pyromonkey83
How well do you expect the Maximus X Hero/Code to fair on the 9900k in overclocking? I purchased it about a month ago for $200 new on Amazon during a sale, and was hoping the higher end VRM would support the 9900k, but the Toms Hardware report that VRM thermals and throttling are concerning with Z370 boards. Do you think I'll be limited on everyday overclock levels around 5-5.3Ghz? I'm using an AIO 280mm cooler as well, so the only airflow is the 120mm Maglev exhaust fan above in my H500i case. I'm starting to think I'll have to pony up another $100 for a Z390 Taichi or Maximus XI Hero instead (and I only have ~6 days left in my return window to come up with a decision).
see above reason why I dont trust Tom's hardware USA. If you aren't
using AVX instructions it should do fine as it uses 8x 60A powerblocks. At worst you can sell the Z370 off and get a Z390 but I highly doubt that it will be a large issue.
Originally Posted by jasonwaterfalls
May be 10+2 phase since 12 chokes are used as opposed to an odd number such as 11 or 13. This assumes that IO + System Agent (SA) are not routed at that area. Mosfet quality is unknown , 10K FP caps from Japan are used. Impossible to tell anything definitive without heatsink removal or the back of the board.
Originally Posted by Hackslash
same (except the return possibility)... would love to know if my MXH is good enough for 5Ghz 9900k (with AVX)
i9-9900k performance (power scaling with voltage, AVX power usage) is under embargo so anything to do with AVX
or AVX2 enabled is truly unknown or a guess based off the 14nm i7-8086K at this point.
, it looks like the Z390-I STRIX is using the same Onsemi NCP302045 powerstage (rated 45A) as the other STRIX boards. I truly doubt it can effortlessly handle over 160W or so , it's half the powerstages as the STRIX Z390-E and Z390-A. Peak efficiency for the powerstage is roughly 91% and that's without a ~1.1X scaling factor for 1.3 or 1.4V output voltage.
For example at 150W output and 1.3V you would have roughly 4.5W of power to be dissipated for each of the 4 powerstages (see figure 7 and figure 12). Even using the highly optimistic 12.4°C/W thermal resistance figure you'd be over 100°C quite easily ; using a realistic 20-30°C/W from an average VRM heatsink and you'd likely be well over the recommended 125°C. http://www.onsemi.com/pub/Collateral/NCP302045-D.PDF
Unless you have
to buy an ASUS board I'd probably look into the Asrock Z370 Fatal1ty ITX or Asrock Z390 Phantom ITX as they are proven to be able to push older i7 hexcores with airflow due to using 5x thermally enhanced 60A ISL99227B powerstages that are more efficient , are normalized against a much higher 1.8V, and have lower thermal resistance due to the exposed junction (get hot less easily when heatsinked). It also comes with Thunderbolt 3 built in.
There's been ROG Maximus XI Hero testing vs the older Maximus X Hero :
The newer board actually fared worse at times with the i7-8700K either due to vulnerability patches or due to early firmware.