Gadgetory

welcome back to harbor unboxed recently we released our core I $9.99 oh okay a review and sided thermals as being a pretty major issue for the new 8 core CPU in fact familes was so bad I was worried I'd done something wrong but prior to release fellow other youtubers Brian from Czechia city and Jarod of Jared's Tech confirmed the horrendous thermal results then about 20 minutes before the NDA lifted overclocking God devourer confirmed what we'd all found once all the reviews will live I then jumped around all over the place checked out what other media outlets were reporting however the most part it looked pretty hot there were however a few outliers OC 3d TV being one they reported shockingly low operating temperatures up to 65 degrees and then when overclocked just 79 degrees both were lower than my stock temperature this didn't make sense until I dug into the system specs and discovered the issue an issue I actually touched on in my own review though not in relation to power consumption and thermals anyway I had no intention of addressing the results despite a few people requesting that we did in the comments section on various social media platforms that were on and whatnot the review police and chances are they're being harassed by quite a few viewers so they'll probably look into themselves so that was that but then the following day our discord chat blew up quite a few members who subscribe to floatplane and that's the lioness tech tips early access platform we're ranting about there 9900 k review our long story short linus was claiming that the peak load temperature was just 58 degrees for the 9900 k and the total system consumption was just 139 watts now both these figures are absurdly low and despite countless viewers pointing that out Linus's stabbed his foot down and declares that he stands by them unfortunately though they do fail the common sense test more than once and despite being technically correct this should have set off alarm bells I know that statement contradicts itself but bear with me his own results show the eight core 19900 K running cooler and consuming less power than these six core 8700 okay the 8-core model of course packs 33% more cores and is also clocked higher and it still uses the same 14 nanometer process so it should be impossible for the 9900 k to use the same amount of power as the 8700 K let alone less so at this point you're probably thinking okay Linus got it wrong here his team messed up well no not exactly Linus address to viewer concerns in a pinned comment saying the following the performance numbers we got were double checked against intel's engineers and they are in line this suggests that our thermal tests are indicative of stock performance with motherboard vendor optimizations disabled the crazy part being that this is actually correct all their performance numbers are technically correct we could argue that as a more correct but technically new the harbor on box nor Linus tech tips got it wrong despite publishing vastly different numbers intel's hell-bent on declaring the 1900 k as a 95 watt part which is ridiculous and it's the reason why our results differed so much to those published by Linus tech tips we used two motherboards to validate the 99 0 k numbers the MSI z3 90 godlike and the as Roxanne 390 Taichi ultimate both boards are more than capable of extracting every last bit of performance from the 8 core processor Linus like OSI 3d TV used a motherboard packing a measly 4 phase VRM a verum that's not nearly capable enough of getting the most of the 9900 K but I should note does officially support the 8 core processor so there is no reason why they can't test with such a motherboard in our day 1 review we compared the Zed 390 godlike to a Zed 370 board using a fourth phase vrm the same design used by most entry levels at 390 boards and this is the reason why we used it and when doing so we saw a very heavy vrm throttling which reduced the clock speed and therefore performance what we didn't show was thermal and power consumption numbers when throttling so I've gone back and retested the MSI z3 70 PC Pro now this isn't the same board used by either Linus tech tips or OSI 3d TV but it does pack a four Fay's vrm and crucially has the same 95 watt limit i've also dumped my open-loop cooling setup with the three hundred and sixty millimeter radiator in favor of the corsair h 100 i program previously we saw a stock operating temperature of 85 degrees with the h 100 i but now with the 95 watt limit in place the temperature only peaked at 60 degrees before dropping down and holding steady at 55 degrees right in the range of the 58 degrees linus reports power consumption also dropped from 249 watts right down to 160 watts a little more than what linus reported but then we are using a different board memory power supply and so on point is that's a 36% reduction in power consumption now you might think well that's great the four phase board will add the ninety nine hundred K to run much cooler and consume less power sounds like a win to me however it's only doing this because it's limiting the ninety nine hundred case clock speed whereas in our review at maintain an all core clock speed of 4.7 gigahertz with the 95 watt limit in place it ran all course at just four point two gigahertz now Intel will be quick to point out the base clock is 3.6 yigit's so we're still within spec but with the limits removed the 9900 K does target a times forty seven o'clock multiplier for a four point seven gigahertz operating frequency linus also reported the 99er EK is having issues with DirectX 12 in games which is on as we saw no such issue and while that would just be odd in general however linus was led to believe this as the DirectX 12 API better utilizes the 99er case eight cores placing it under more load and therefore forcing the clock speed down this is why the 8700 K was slower than the 99 okay using DirectX 11 but faster using DirectX 12 they witnessed this in Deus Ex mankind divided and shadow of the Tomb Raider if you recall I mentioned in my day 1 review the four phase board would allow the 9900 k2 hit 1900 points on the first Cinebench run and guess what lioness reported 1911 points and that's 7 percent lower than our score however if you run the test at least three more times back-to-back the score will drop down to under 1800 points or at least that's what we saw and that's a 13% decrease and that makes sense as the 99th okay is running at least 11 percent slower in terms of frequency so mystery solved and technically the Linus tech tips numbers are correct as the Intel spec is very broad now so we're not calling Linus up we're not saying they did anything wrong as technically they didn't for those of you wondering the 1900 order for it to maintain 4.7 gigahertz the all core frequency the 95 watt limit needs to be raised to 150 watts and that's a 58% increase so that means boards aspect to only handle 95 watts are running massively out of spec in order to get the most out of the 8 core processor it also means if you're going to run the 9900 K at the maximum spec like we did then you are going to need a serious cooler and even then expect things to get pretty hot anyway I hope that clears things up and explains why our results were so different to those who used boards with the 95 watt limit in place and with that I am going to end this one if you did enjoy the video please hit the like button subscribe for more content just like this and if you appreciate the work we do home box then consider supporting us on patreon you will gain access to our discord chat where you can harass us about our review and other reviews and stuff like that and you can also join you can get access to our monthly live stream we're gonna do one of those this Thursday so coming up very shortly on the channel that'll be a lot of fun anyway I'm your host Dave see you next time