n4da

Buying good fans for my 8700K in a Meshify C was probably one of the most dumb investments I made. Just writing this as a warning for those willing to invest in fans. I started with CNPSx10 Optima cheap 120mm cooler, 2x120mm intake front , 2x120mm top exhaust, 1x120mm back exhaust, cheap fans from Arctic and Fractal Design (2 and 2) I ended with Noctua D15S+120mm additional fan, 2x140mm intake, 2x140mm exhaust, 1x120mm exhaust. All fans Noctua PWMs. Temps... well... the CPU gained -5C, maybe -10C at 5GHz. The GPU, a 1070ti, maybe 1-2C lower. The overall price? Around 190EU. I'd say just invest in delidding (if relevant) and change the GPU thermal paste if you're really bored, and at 1/3 price you get triple benefits, and stick with cheaper fans. Oh about noise? Yes the Noctua are quiet at lower RPM. At higher RPM (60-70% and above), they're as loud as the cheaper fans I had. The Noctua cooler and fans are great IMO, I actually like them a lot, and their support seems amazing. Their packaging was also great. The price though, holy shit.

I've went with a Noctua D15S (adding an extra NF-A12x25 PWM fan to the front, that should fit on top of the RAM and in the case still) and 4xNF-A14 PWM for the case, 2 front intake, 2 top exhaust. Decided against the AIO as I don't trust the pump reliability and its longevity in general. The big fan should last a decade with ease, can take it out and wash it, doesn't leak, and performance should be near the 280mm AIO. The big 140mm fan in the mid of the fan should blast some air over the VRM too.

Nobody? Lemme know if you need more info.

Hello everyone, First, specs: 8700k (not delidded, I really don't wanna pay 50EU just for the tools), MSI Gaming Plus Z370, Meshify C, ADATA Dazzle RAM (4.73cm height according to this: http://www.xpg.com/en/feature/376), 2 HDDs in the cage in front of the PSU (which is sadly quite long, a SuperFlower Leadex II Gold 650W, the cables kinda press on the HDD cage, and the HDDs kinda get toasty in summer (42C max). I have 2x120mm fans front pulling, 2x120mm fans top pushing, and 1 120mm fan on the back as exhaust, and a cheap Zalman CNPS10x Optima CPU cooler. There's also a 850 EVO SSD and a Gigabyte 1070ti Gaming 8G OC (the one and only factory OC 1070ti whoo) that's 280mm long. Looks like this: https://www.gigabyte.com/Graphics-Card/GV-N107TGAMING-8GD#kf OK, so it's been almost 2 years since I built this and even though the current cooling is satisfactory, in summer things get toasty. The card can get to 73C, and the 3 small fans are somewhat noisy at the approx. 50-60% rpm. The CPU doesn't really care much in games, with rare spikes towards the 65C, mostly stays around 55C. The power draw this motherboard can hold is somewhere around 150W because of the flimsy VRM, which is also why I'd love to stuck a fan on top. The 1070ti doesn't require me to push the CPU clocks for the games I play, but I'll probably buy a RTX 2070/2080 Super equivalent next year which is likely to require 5GHz for 1080p. So my plan was to buy an EVGA CLC 280 and deal with the (alleged) noise, and mount some sort of fan on top of the VRM. Meshify C is too small for top 280mm. Front yes, but to me it's kind of weird having hot air from the radiator blown in the case. I kinda don't want that. Second plan was to buy a NH D15, but it looks like my 2 RAM sticks are too high, and the case might be too small for the first fan that I'd need to elevate over the RAM. Obviously I don't want to buy other RAM. I'm also unsure about a 240mm top AIO, will it be able to deal with the 8700K properly after I OC? Any suggestions? Yes I realize I complicated things for myself with the case and RAM and even the motherboard which is likely to require a fan over the VRM, but they were cheap and I was running out of money. NZXT and Corsair AIOs cost quite a bit more than EVGA and the other weirder corporations like Deepcool and the like I kind of don't trust. Should I try a NH D15S and add a 120mm fan? Do I lose significantly on the temps? Or should I just go with one fan. Meshify C has 170mm CPU cooler height tolerance. I wonder if it would allow me to use the full D15. Any suggestions appreciated. EDIT: I wonder if it's possible to put the CLC 280 front with the fans first mounted to the case, as intake. I suppose that would heat up the GPU more, but I'll keep 3 exhaust fans, 2 top and 1 back. Is the CLC so noisy that's unbearable? An AIO would allow me the option to put a fan over the VRM.

Simply enabling HyperV, no VMs running, is taking about 5-20% CPU performance in the various AIDA64 tests. I didn't notice much of a difference in actual gaming, mostly in benchmarks, but I do have a 8700K/1070ti combo so that means that the GPU is almost never botlenecked unless I play something super unoptimized and/or old. If you're curious how much of a hit HyperV enabled is (Vt-d is active in BIOS), here's the 2 reports from AIDA64. Cinebench and CPUZ also show differences, although CPUZ multithread is actually quite a bit better WITH HyperV /shrug 3900 points with, 3820-ish without. AIDA64 Reports.rar

So I know this is a bit old, but it contains a lot of misinformation. It is extremely easy to verify if games use AVX, by adding an AVX offset in BIOS and monitoring core frequencies in MSI Afterburner, and surprise-surprise, a lot of them do. For example, modern games like SotTR, Darksiders 3, Monster Hunter World, AC:Odyssey (and probably Origins too), and even Overwatch, all use AVX, and you will see that when the frequency of the CPU goes down to respect the AVX offset you used in BIOS. So the actual truth is a lot of games use AVX, and it's also used relatively often. In Windows, you would see the CPU stuck at its highest OC frequency, but in games it will start to oscillate a lot because of some AVX being used. This is important as if you truly want to OC, you will either have to accept lower/oscillating clocks because of AVX, or you will have to stress test the OC with prime95 last versions with AVX. Considering how power hungry that can get, you might have issues. For example: - non-AVX prime95 4.7GHz 8700K, approx. 120W (depends on voltage used, of course) - AVX prime95 4.7GHz can get to 160-170W+, and requires a lot more voltage to get stable, and a lot more cooling, and better VRMs on the motherboard. The games might not consume that much watts, that is true, but you will need to test the OC to see if it's stable, and p95 is basically your only option, most other tests are trash in comparison. I've forgot how many people would stresstest with Realbench or LinX or IBT and think they are stable, when prime95 would fail in a minute, and they'd have games crashing and even BSODs at times. TLDR: most modern games use AVX and it's easy to prove.

You need the BIOS that holds the updated microcode for SSB. You want to see this line: Hardware support for speculative store bypass disable is present: False Go to true. What I assume this will do for L1TF is "teach" your CPU to flush the L1 cache (when exactly? how? it's above my knowledge). You want that so that basically malware can't "infer" (Intels's term) potentially sensitive data inside the cache. In some scenarios this would negatively impact performance. If you just play games, browse the web and do "normal" stuff, you won't notice anything, not even 1% drop (can confirm, tested with a few game benchmarks, Cinebench, AIDA64 before/after). Running VMs however might take a hit. I am also having the same issues with loading old BIOS settings after an update, not really working, so it's manually setting everything up again. I'd still update to the last BIOS. It should also contain other potentially useful stuff, like the new Intel RST EFI module v. 16.5, which should pair nicely with 16.5 RST drivers. If that's better from the 16.0, I can't say. You should also know that if you want ALL protections enabled, you'll have to manually do it for SSB. I tried it and found again no performance impact, but your mileage might vary. So: - you don't need a new microcode for L1TF, the one released for SSB should do (Intel confirmed). It should be rev. 0x96. - I'd update the BIOS if I were you. PS: it appears that the output you see in my screenshot is the best you will get for L1TF. Hardware Vulnerable will always remain True, which is kinda accurate, yet peculiar choice of words. Nobody wants to patch shit up and see "Hardware vulnerable". The output just tries to say that your hardware is vulnerable to this exploit and measures against it were applied in the OS, and for those with updated microcode, the CPU also can flush the L1 cache. Which, indeed, is not a real fix, it's just a workaround, but let's be honest... ... no Spectre/Meltdown attacks of any sort were yet detected. No PC will ever be 100% safe.

It means that: - Spectre 1-2 patched, OS and hardware (microcode/BIOS), and you still have good performance due to having INVPCID, and well, Coffee Lake helps too. - you don't have the microcode update for SSB installed, the OS is patched but it's not enough; you need to not only update the BIOS (I'm sure Asus has the new microcode in their latest BIOS), but also to manually enable SSB (it's not automatically enabled like previous Spectre/Meltdown mitigations). Doing this is a further (small) performance impact. 1% in games, if that's what you care about. - the last part is tricky. Problem is, the same microcode as for SSB should be the one needed for L1TF. But, i have that one, and well, it's not enough. Something else, somewhere, somehow must be done, I am not sure what at this moment. I have also didn't manually enable SSB mitigations. Maybe it is that? This is how it looks on my 8700K: It would be nice if somebody knew more about this, but from the 0 answers you got... I'm a bit skeptical. Edit: Alright, I enabled SSB. Now SSBDWindowsSupportEnabledSystemWide is True, as expected. Sadly L1TFHardwareVulnerable is also remaining True, so it's not that. Microsoft is saying that if you use Virtualization Based Security (VBS), you'll also need to turn off Hyperthreading. I tried that too, even though I am not using VBS, at least it's not enabled, and HT or not, you're still L1TF vulnerable. I'm guessing something is broken here.

You can run Secure Boot too. You just need to disable it once at best (or find a version that works for you), run Gigabyte's shitty program, set it the way you want, apply exit, uninstall, reenable Secure Boot.

It should be fixed in latest software. It was broken for my 1070ti for ages, until they finally fixed it, months later. Just try the latest version.

The latest version finally fixed this (at least for me). The application seems to work properly without crashing, and without having to perform any steps. That's good, as it's the only way to change the LED on my card.

The fixes suggested in the video didn't do anything here. Even worse, 1.33 removed some LED options, so after trying to see if 1.33 fixed the service fail problem (it didn't) I had to use 1.32 to actually see the 2 missing LED modes and the option to use Breathing through the colors, instead of just one. I hate this dumb application with a passion. I tried each version since I think 1.28 to see if it's fixed, and obviously it's not. If you can't fix something as minuscule as this over 6 months (about the time I have my 1070ti), I fear facing their customer support. To make things worse, there's no official support forum or place to report the issue. I tried Twitter got ignore obviously.

Anybody with a shred of brainmatter left would have already known that the 8700K cannot lose so much performance overnight. Just like anyone that actually has a 8700k/Coffee Lake CPU can test for themselves how disastrous the HPET is for framerates and frametimes. It's going from silky smooth to a stuttering mess, it's not just some performance loss. I am glad however to see the matter straightened out. Those benchmarks looked ridiculously dumb. I would have never published them before understanding WHY it is just my website that has different results. Also Intel are barking mad with their recommendations for HPET. I hope they change that crap for their own good, if they care about good press.

I can't believe that after 2 days of people suggesting me all kinds of expensive crap that I don't want for my purposes, here we are, after hours of troubleshooting, disabling audio inputs and outputs, plugging off USB devices, unplugging the front panel from the motherboard, using the iGPU and taking out the 1070ti, here we are. Ended up using a longer, very thick extender cable I had to the other room, plugged in it another extender, and in it the 2 monitors as far from the PC as possible. Clean audio. Miracle. Even in games. It looks like crap and my GF will be definitely unhappy with the cable mess, but I really don't care now. I guess I'll have to figure out a way to bring power to the speakers from the other room or even the bathroom. Out of curiosity, is there any explanation how an old Zalman Bronze crap-tier PSU resulted in less electrical interference compared to a new Superflower Leadex II Gold that's apparently tier I and praised all over the web? Many thanks Belzebuth. Clean audio is orgasm to my ears, after the insane noise I experienced, this is a miracle.

Sadly doesn't have any effect, tried just now to use a monitor in an electrical outlet in another room, I can still hear the mouse and whatever else is going on inside my PC case. Actually. I was wrong. I played a bit more and I could find an outlet with much less noise, it's exactly the same noise, but significantly less. It's ALMOST OK, almost. What can I do to improve the situation more? I have only one electrical outlet in this room and I don't really want to drag a very long, ugly cable extender to the other room just so I can connect 2 monitors. Thanks for the suggestion btw, it definitely worked, now if we could only refine it a bit.