Gadgetory

that was a good livestream so we did the i7 8086 k overclocking livestream this is a recap of the results the whole thing was four hours long so understandably a lot of you probably don't want to watch it after the fact if you do wow more power - yeah so yeah four hours long the reason it went so long is because we had a really good concurrent viewer base the entire time and I really wanted to hit a specific score and fire strike that was twenty four thousand points for physics and we got pretty damn close so I'll give you the numbers in a minute so we're just gonna recap the frequencies I hit this is a 286 k it's especially been to basically 8700 K same architecture same core account same everything same stock frequency even the difference is it over clocks higher nearly guaranteed because it is pre-selected by Intel to be one of the higher clock in 8700 KS and so we thought it would be fun to try and do that live and see how far we can push the thing before that this video is brought to you by Thermal Grizzlies high end thermal paste and liquid metal Thermal Grizzlies cryo knot is an affordable high quality thermal compound that doesn't face some of the aging limitations of other pastes on the market cryo knot has a thermal conductivity of 12.5 watts per meter Kelvin focuses on endurance is easy to spread and isn't electrically conductive making it safe to use on GPU dies thermal grizzly also makes conductor not liquid metal which we've used to drop 20 degrees off some temperatures than our dee-lighted tests by a tube at the link in the description below so first of all recapping all the Part C is what we did was first off deleted the CPU we used a rocket cooled D letter kit that I had it works for the 8086 k just fine technically the one I have for 7700 KS but it still works then we removed all of the silicone adhesive the black adhesive around the outside we removed all of it scraped it all off and then put on some liquid metal we put on thermal grizzly conductor not they were the advertiser for the stream but we were going to use their liquid metal anyway so it worked out very well so thermal grizzly conducted on something we've used a lot we have another content piece that's live already that talks about our temperature reduction by using that instead of the stock thermal paste that help what's on there so that was a major help to removing our thermal limits he really didn't have any thermal limits at all during the stream the problem was mostly it seems like the IMC on the CPU which I'll get to in a moment other parts in the build we had an asus rog maximus x or 10 motherboard did pretty well seem to have issues with the command rate of one so that was annoying we end up using 2 most the time so that was a motherboard quark but otherwise it was a good board and also used we spent a lot of time in bios of course we also used three different kits of memory so this was part of trying to figure out if the IMC was the problem or I don't know or if it was the memory and what we did was we use the G scale try and Z black kit that we used for our repeal TT stream semi-famous stream for us and that kit we know does 4,000 megahertz it's a good memory kit it does 4000 mega Hertz it does I don't remember CL 16 roughly is 16 16 26 I think we're the timings on that one on X 299 not I'm not on coffee like he can't do the same timings we were at like 16 16 28 or 30 30 or something like that and then we also had a several other timings tightened I'll list them in the article below or something like that if you're curious but ultimately even though we know the kit does 4,000 megahertz we could not get it past 36 60 something on the 8086 K it was very unfortunate so it wasn't it wasn't a memory bin the memory is fine we know it's good what it started indicating sort of unfortunately for us was that the 8086 K that we have probably didn't have the best IMC in the world the integrated memory controller and this is a problem when you talk about memory frequencies a lot of the other kit we used was a Corsair forty 600 megahertz CL nineteen kit forty six hundred megahertz pretty freakin high so the thing that you run into with those kids even if the memory at some point in testing and quality assurance did 4600 there's no guarantee you'll ever get to that number cuz you're talking about how good is the motherboard you're running and how good is the IMC on the CPU and you probably want to bin for the course so you're bidding for the core for the IMC you're getting good motherboard and you've already been for the memory by buying something that's rated as high so there's no guarantee it'll ever run at those speeds but we tried it anyway and even the 4600 kit with for it lose timings like CL 16 could not do anything over 3600 so it seems like that's just our limit for our 8086 K pushing SSC to 1.3 and IO to 1.3 didn't really help too much so anyway what we ended up with our goal was 24,000 points in fire strike physics only and we got 22,000 745 so I'll be back I'll be back to get those last three hundred or so points 250 points because it was damn close and the only reason I cut it short at the end was because we're just getting blue screen after blue screen trying to get the last couple hundred points and it was clear that you know I got a got to kind of shelve it and let it rest and figure out what's going on so way back but pretty damn close final clock on the CPU as 5.35 gigahertz tried for five point three seven tried for five point four to get it to work so we did a one-on-one BC okay we 253 multiplier 1 point 4 or 5 volts pretty stable we actually can do 5.2 gigahertz with a similar bc lk at i want to say one point four two volts or even lower sometimes depending on the that's with prime mind you with fire strike it's way less voltage so a vote frequency curve was very good on the cpu this has definitely got a good core on it we were been pretty well for the volt frequency curve towards the low end once we hit 5.3 and onward though that curve got exponentially more difficult to climb requires exponentially more voltage for each step after 5.2 on our cpu and so 5.25 was our limiter today and we might be able to do more later we tried perk or individual overclocking to 54 onto of the course even one of the core and just didn't do it would not hold kept please screening we had a one point four or five volt for most of our test we went up to one point five in some scenarios we went to one point five the thermals on the cpu are still pretty under control because it's only 30 second test so it's fine but having a lower temperature would still have helps especially power leakage so it might do a bigger radiator next time and also we did have to stick a fan on top of the vrm so that was a concern too once we start pushing 1.5 volts into the thin we tried to 102 BC okay we tried 101.5 had trouble stabilizing ultimately wasn't with the memory clocks we wanted would have liked to try 103 with a lower multiplier and we'll do that in the future and then we also spent two hours trying to Tim the RAM so at Coursera we had a coarser 32 hundred megahertz CL 16 kit that I started with honestly I thought that was gonna be the worst one but I guess that kit is pretty decent for us because I thought the g.skill kit would destroy it but as noted that g.skill kit just wouldn't go over 3600 for this memory or for the cpu rather so the courser 3200 kit we pushed it up to 3600 it was still going strong and that was with the CL of 15 so really pretty damn good for that specific kit of memory because it's it's not meant to be that good we also did a course at 4,600 kit pushed it to 3600 on CL 1415 area I think we did 14 pseudo stable at one point and ultimately again it seems like our IMC is simply not that good on our 8086 case so that's that's a sad thing for us but voltage is overall we tried pushing 1.5 occasionally was not enough 454 X on the multiplier we added an H well fans of erm when doing this we ran a system agent at 1.3 IO at 1.3 wouldn't recommend doing those voltages are higher than them for any kind of long-term use but for a benchmark who cares same for B then we had V damnit 1.85 and some tests and 1.6 and others now a couple things here V Tim you don't want to go beyond 1.5 and Rison or Intel for any kind of long-term use you'll burn up the memory controller the SOC and what we're doing is different we're doing short benchmarks the goal is to get a high score and then who cares consequences be damned we're gonna do it so don't don't just copy our numbers please because those voltages are pretty damn aggressive in some instances and could cause some long-term degradation but for one night of benchmarking it's fine so yeah 1.85 demographic we did 1.6 and other tests the RAM can absolutely take it has no problems at all at one point eight but the IMC's another question all current settings were fully uncapped so current limits on caps pushed Alcee to level six left there for the most part because at level seven tends to overshoot the voltage rather than undershoot when it dips and the big takeaway here is learning more about memory overclocking so a lot of the stream I spent interacting with the last hour so interacting with builds would be a discord and from actually hardcore overclocking I was able to get the memory improved a lot and improve our scores a lot as a result so it was a great learning experience for me my big takeaway is learning more about memory timings we're working on our secondary and tertiary timings videos Patrick's researching those that'll help me a lot too and that's definitely my weak spot in overclocking so I loved the stream I greatly enjoyed interacting with everyone thank you for being there if you were there and thank you for watching the follow up if you were not and we're going to keep doing these I'm just bear with me I'm gonna keep working on my memory overclocking skills so and not just our clocking but timings tuning so in the future hopefully that process will be faster but even then people who are skilled with it it can still take hours so we're always going to pre test some of it just to get a rough idea of where we can fall and the goal I have with these at the outset is to basically say I know I can do 23,000 points I think I can do 24 but I haven't done it yet so my goal going into these is always pick a number that's achievable and then set a goal just outside of it so that in the live stream we have some genuine real effort pushing for it it's not like it's pre-planned or anything like that it's I mean you're seeing the real blue screens as you're watching it so narrate end result here five point three five gigahertz on the core thirty-six hundred megahertz was our cap for the memory it seemed maybe we can push more with a looser set of timings but then you start dropping firestrike scores so the trade-off just isn't worth it in our instance I you could go CL 18 or 19 maybe do closer to 4,000 but I don't know that it would have gained us anything anyway we got we have a lot more to try and I'm looking forward to it so that's our stream we have thermal power and even noise as an acoustic testing on the 8086 K and another content piece it's already online go check that out if you want to see these thermal performance with or without liquid metal and how we can reduce noise by 10 DB a by going with the D lid we did so check out that content more subscribe as always get additional coverage and go to store documents access on that to pick one of our blue tear down shirts we just restocked them they've been out for a few weeks you all bought too many so they're back now if you've wanted one and it was gone it's back and patreon.com slash schemer taxes helps out directly in smaller quantities if you prefer and get access to the discord subscribe for more I'll see you all next time thank you for watching