I feel kinda weird inside...

[ArbitraryAffection]

I would prefer to buy AMD over Intel and Nvidia if given the choice between the companies. Nvidia is a scummy company with things like GPP and such, and Intel is no stranger to anti-competitive BS. Besides, AMD actually offers very competitive products even GPUs. Well, if they were anywhere near MSRP.

I just really don't like NVIDIA or Intel as companies.... But even then I still bought an Intel/Nvidia Laptop because i needed one and AMD still have absolutely no decent gaming lappies in the £600-800 range.

8700K is a really nice chip, though, Can't fault it. It'll serve you well. I plan to stick with my 1800X until Ryzen 3000.

 

[R-T-B]

No, they delid to do extreme OC's[LN2] generally.

Considering liquid metal isn't usable on ln2 due to it's high freezing point and it's the delidding tool of choice for many, not sure that is true...

 

[Basard]

I'm using the HE01....

I think deliding would give a good drop in temps, even if you just put some quality paste in there. It's not like I would glue it back down or anything, I'd let it go commando so I can repaste it whenever. That's pretty far down the road I hope.

I'll set voltages manually in a few hours, got some family in town I gotta go see. Here's where I'm at now with these VIDs....

 

[cucker tarlson]

do you have the enhanced turbo boost option enabled ? if so, turn it off.

 

[Vario]

I'm using the HE01....

I think deliding would give a good drop in temps, even if you just put some quality paste in there. It's not like I would glue it back down or anything, I'd let it go commando so I can repaste it whenever. That's pretty far down the road I hope.

I'll set voltages manually in a few hours, got some family in town I gotta go see. Here's where I'm at now with these VIDs....

View attachment 99077

Interesting how it shows seperate VIDs per core. VID doesn't really matter on these actually, has very little to do with the Vcore applied. Aim for something like 1.2V-1.25V Vcore and see where you end up. If it is a good chip that should get you around 4.8-5.0. It is as simple as bumping your multiplier up a bit. Select offset mode for your voltage and leave it on 'auto' and see what voltage it gives you. Then give it + or - offset instead of 'auto', until you end up in that 1.2ish range. That should result in reasonable air cooled temps. Once you figure out how high it can go at a reasonable temperature and voltage, then you try raising your uncore frequency, you should be able to do about 4.5 uncore easily. What is your dram voltage? Usually DDR4 is 1.2 to 1.35 volt. If you want to get your temps down consider adding an additional fan to your HE01, it might help.

Here is the coffeelake discussion thread: https://www.techpowerup.com/forums/...-overclock-results.240682/page-6#post-3820751

 

[Deleted member 67555]

I was getting stupidly high temps and lockups on my system and I got sick of it and went to Intel to start the RMA process..
I ran the tests they wanted and was still getting stupid temps and lockups even though it was passing their tests..
They ended up running some toolkit on my PC via remote desktop and it's like a whole new beast... My temps aren't going above 70°c anymore.. in gaming I'm not going over 50°.
I haven't bothered to overclock anything since...
Ruining at stock clocks and memory at 2133 I'm getting a single core score of 530 in CPU-Z benchmark.... I have 3200mhz memory but as if yet I don't need it any faster...
I have a feeling Intel has a better Microcode update on the way that's going to address the problems these chips are having with temps and it'd be nice if this was covered under the next BIOS release.

 

[Vario]

I was getting stupidly high temps and lockups on my system and I got sick of it and went to Intel to start the RMA process..
I ran the tests they wanted and was still getting stupid temps and lockups even though it was passing their tests..
They ended up running some toolkit on my PC via remote desktop and it's like a whole new beast... My temps aren't going above 70°c anymore.. in gaming I'm not going over 50°.
I haven't bothered to overclock anything since...
Ruining at stock clocks and memory at 2133 I'm getting a single core score of 530 in CPU-Z benchmark.... I have 3200mhz memory but as if yet I don't need it any faster...
I have a feeling Intel has a better Microcode update on the way that's going to address the problems these chips are having with temps and it'd be nice if this was covered under the next BIOS release.

How odd. Can you tell me your microcode revision? You can look it up in your registry, it is in HKEY_LOCAL_MACHINE\HARDWARE\DESCRIPTION\System\CentralProcessor\0 and is the number in your Update signature.

 

[Basard]

Interesting how it shows seperate VIDs per core. VID doesn't really matter on these actually, has very little to do with the Vcore applied. Aim for something like 1.2V-1.25V Vcore and see where you end up. If it is a good chip that should get you around 4.8-5.0. It is as simple as bumping your multiplier up a bit. Select offset mode for your voltage and leave it on 'auto' and see what voltage it gives you. Then give it + or - offset instead of 'auto', until you end up in that 1.2ish range. That should result in reasonable air cooled temps. Once you figure out how high it can go at a reasonable temperature and voltage, then you try raising your uncore frequency, you should be able to do about 4.5 uncore easily. What is your dram voltage? Usually DDR4 is 1.2 to 1.35 volt. If you want to get your temps down consider adding an additional fan to your HE01, it might help.

Here's where I'm at now. Instability in Prime95 on the 12th thread. 1.21v idle, 1.11v under load, a little lower than you suggested (it's not that I don't trust you, I just don't trust myself, lol). I was getting a quick blue screen when starting up P95 at slightly less voltage. It's a work in progress:

I had an add an additional FHP141 on the tower for a while, but then it turned my Fractal 140mm exhaust into a saxophone. I was going to put my third FHP141 in the exhaust port, but i need another MM for clearance... SO, I just settled for FHP141 exhaust fan and a single FHP141 in the middle of the tower. I guess now would be the time to install the one FHP141 that I have left, and put the Fractal fan back in as well.

I did some work on the fan curve in my BIOS... boy that's a far cry from the last few boards BIOS's. Still stuck in Gigabyte land... (my last 10 boards, JEEZ) I hate them so much but it's the enemy I know, I guess. I've built tons of others for other people though. The good deals always seem to come from Gigabyte though (gee, I wonder why sometimes.)

Their app center software could use some work. When @bios gives me three errors and crashes before it even finishes loading, how then do I go about updating from F4 to F5, lol..... So I'm still on F4.

So, onward to add some more voltage and maybe get some BIOS updates!

I'm going to reboot and see what RAM voltage is being read, can't seem to find it yet in HWiNFO.... App center's SIV is telling me 1.344

@Vario When you say "Select offset mode for your voltage and leave it on 'auto'," would that be the same as "Dynamic voltage control?" I see that under my vcore in BIOS, but it's greyed out, something to do with turbo or something I'm guessing.

 

[Vario]

Yes I think thats what gigabyte calls it.

 

[trparky]

Have you tried downloading the UEFI update file and apply it in UEFI?

 

[Basard]

@trparky I'm gonna be working on that pretty soon.

I've got all my fans installed. Three FHP141, one exhausting behind CPU, one on the cpu, and one on intake on top of my case, blowing down on the side of the CPU tower. An additional three of the Fractal 140mm's are also on intake. Temps seem to have dropped about 4C all around and things are a little quieter.

Running all cores at 4.4Ghz under load vcore is at 1.133, 1.2 idle. 70C on the cores after about 30 mins of prime.

Gonna go after that BIOS now and restart.

 

[Deleted member 67555]

How odd. Can you tell me your microcode revision? You can look it up in your registry, it is in HKEY_LOCAL_MACHINE\HARDWARE\DESCRIPTION\System\CentralProcessor\0 and is the number in your Update signature.

So now I have a better idea of what happened...I think..
As I've reported prior my problem started after I started getting a firmware error in Windows... I then updated to MSI BIOS 1.2 released on December, 17 2017 with a microcode update...
My system went super wanky and locked up a lot...
I then went to start an RMA with Intel...jumped through their hoops and still had a problem...Intel then ran a program on my PC via remote desktop my PC restarted and was fine...
After talking to you I found out I was back to the release day microcode 7c...
I have since updated to microcode 84 and my machine now runs like it should have from day one...
Voltage numbers are fantastic and MSI added some really good options in the BIOS...

 

[Basard]

So.... basically I had everything still installed from my old motherboard..... I think it's time for a clean sweep of everything.

I thought to myself "I still have @bios, from my old motherboard's software," installed. Uninstalled that, and a million other things, and it works fine now. The new BIOS didn't seem to help much of anything noticeable for me.

I've got these two thermal probes that came with the motherboard.... How cool would it be to get a temp of the intake air and a temp of the exhaust air on the computer? So that will be a thing pretty soon.

Everything is basically at optimized defaults right now, with exception of the IO and SA volts.... Not a whole lotta time to mess around with this thing, and I wanna get it crunching at a respectable temperature and noise level.

Fan curves in the BIOS is my new favorite thing. At least I can tell my fan to shut up if I want it too, but then temps seem unreasonable.

You guys say that 80C and below is good for 24/7 operation?

And for the life of me, I can't figure out when the dynamic vcore is ever NOT greyed out for me...

 

[lexluthermiester]

I'm using the HE01....

If you're talking about this one; http://www.silverstonetek.com/product.php?pid=341
That's damn good heatsink and you really shouldn't need to delid your CPU to get good OC's and performance with it. That's just me though.

 

[Basard]

@lexluthermiester That's the one. I can get good temps from it, but it's LOUD. I like to keep the fan below 1300rpm. My wife and I have our computers in the same room, and I like to run Boinc during the day--without driving her insane.

After setting Boinc to run with three of the six cores, temps seem to be maxing out at about 73C on the used cores, with fan speeds maxing at 1200ish rpm.

 

[Vario]

So.... basically I had everything still installed from my old motherboard..... I think it's time for a clean sweep of everything.

I thought to myself "I still have @bios, from my old motherboard's software," installed. Uninstalled that, and a million other things, and it works fine now. The new BIOS didn't seem to help much of anything noticeable for me.

I've got these two thermal probes that came with the motherboard.... How cool would it be to get a temp of the intake air and a temp of the exhaust air on the computer? So that will be a thing pretty soon.

Everything is basically at optimized defaults right now, with exception of the IO and SA volts.... Not a whole lotta time to mess around with this thing, and I wanna get it crunching at a respectable temperature and noise level.

Fan curves in the BIOS is my new favorite thing. At least I can tell my fan to shut up if I want it too, but then temps seem unreasonable.

You guys say that 80C and below is good for 24/7 operation?

And for the life of me, I can't figure out when the dynamic vcore is ever NOT greyed out for me...

Might be you have to change multicore enhancement or something to adjust that dynamic vcore I am not sure.

Would be better if it was like 60C if it is crunching 24/7 but 80C is probably okay. Have you tried remounting the cooler? I found that turning my cooler 180 degrees made a huge difference on the thermal paste imprint, strange as it seems. I would be really reluctant to delid if you want a reliable 24/7 don't mess with it machine.

 

[Basard]

I remounted it last night, spread was damn near perfect... a work of art lol... I'm gonna post some pics of it tonight, super high res!

 

[lexluthermiester]

I can get good temps from it, but it's LOUD.

Have you thought about an AIO water cooling unit? There easy to install, cost effective and are very quiet even load.
There are a bunch of really good ones. I lean towards the Cooler Master's currently because the performance vs cost balance is in a sweet spot. But there are a lot of companies that are making good AIO's.

 

[EntropyZ]

You can't go wrong with 8700K, especially on gaming. I slightly regret getting the Ryzen 5 1600, because most of the games that I like.. they run faster and more stable on Intel, and developers for those games aren't making that change. I banked on it being a good budget performer (which it is) for the next 3-4 years at least. That is the biggest con for the Ryzen series to me. I can surpass non-overclocked Skylakes easy with SMT off and a healthy overclock. But the system... I've had so many app crashes and blue screens, and I wish it would just end.

I don't know what happened, one day it ran fine, the other when I updated the BIOS, Windows to Creators Update and chipset drivers, all hell went loose. Been suffering for 2 months.

I have never had anything like this with Intel rigs I built, very rare issues, and they solved quickly. Usually be themselves.

 

[lexluthermiester]

when I updated the BIOS, Windows to Creators Update and chipset drivers, all hell went loose.

So roll back to your previous config?

 

[Basard]

@lexluthermiester Well, I was eyeballing a few, that's why I finally decided to upgrade my case, I could only fit a standard thickness 280mm radiator and standard fans in the front of my Monster II SE. I really want to make use of my two "spare" FHP141 fans, and this case will let me mount them on a thick 280mm. I was looking at a 280mm Phobya kit, or just building my own....

Onward with the computer porn! Documenting my original mounting of the cooler, followed by my application procedure (blob, mush, scrape, and plop), and finally my temps and whatnot today while running 50% cores today for 9 hours on Boinc.

I could do with some lower voltage, actually, as I've ran all cores for a few hours at 4.5Ghz with 1.13-ish volts. DOH!

The statistics chart in Boinc looks promising, sadly, I didn't do a benchmark on my old system. I could make my wife do one on hers though, it's almost identical.

 

[EntropyZ]

So roll back to your previous config?

Isn't it just supposed to work? Especially when there are so few people having these same sort of crashes. Everyone else seems fine.

I even went with a clean install of everything. Which was a waste of time, because the result was the same. Memtest64 shows no errors when I ran it for 5 loops.

Consulted in the AMD forum, no help. Everyone is blaming my RAM, how can it be? If every stress test is showing it to be OK. It's just those applications that crash with memory exceptions and windows kernel errors that never point to drivers, I wish I had another rig or RAM kit to test with, but I don't.

I lost access to all of my resources recently. Which makes troubleshooting very hard and time consuming.

 

[Basard]

I would get a lot of totally random lock-ups with my FX chip... my wife does as well (FX8350+Gigabyte 970A-UD3 with latest BIOS) We blame Windows 10 like 99% of the time, never used to happen on 7, but then we both upgraded from Phenom to FX right around the same time Win10 came out. Oddly enough, the only lock up I've gotten so far with this setup is the self inflicted blue screen. So who knows.... pray to rnjesus. @EntropyZ

 

[EntropyZ]

I would get a lot of totally random lock-ups with my FX chip... my wife does as well (FX8350+Gigabyte 970A-UD3 with latest BIOS) We blame Windows 10 like 99% of the time, never used to happen on 7, but then we both upgraded from Phenom to FX right around the same time Win10 came out. Oddly enough, the only lock up I've gotten so far with this setup is the self inflicted blue screen. So who knows.... pray to rnjesus. @EntropyZ

I've sorted out the kinks when we had the taste of AGEGA 1.0.0.6 update, which allowed my RAM to finally run at 3200MHz, even at being dual-rank with only 1T command rate. Which was awesome btw.

Everything was fine up until recently, I had to change out my AsRock AB350M Pro 4 though because of front audio issues, I got an identical board from the RMA. So I fired up my rig after 3 months absence, everything updated, the BIOS, the drivers... And then I got cucked.

And I have no idea what went wrong. Because so many things happened at once without any testing before-hand. Which was a big mistake.

Stress tests run fine. I get good temps on the VRM and the CPU runs on an AIO. Running with a completely stock system still has random crashes.

Just bought an ultrawide monitor to take advantage of that GTX 1060 I bought. You know what happens 10-15 minutes after gaming? Hmmmm.

I already destroyed my tempered glass table and my previous monitor from this, when it finally started to happen.

 

[Basard]

Oh yeah, in other news. I get a text from my wife today, while I was at work. She writes that my mom's sound doesn't work on her computer. So I drive up there, in this bastard snow storm for 25 minutes, walk inside, give the dog a treat, reboot the computer and it works. That's and old 780G system with an Athlon II x2, rocking some RAM I won from crunching prizes TPU... but WINDOWS 10!! LOL...