*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

I don't see Intel giving away the 8 core units just yet. The game right now is where Intel is giving away a little to AMD in the productivity side, not a lot mind you. They are keeping the high IPC for gaming and desktop performance. I think the 8 core is being held off too look at the Zen-2 first.

At that point they will know how much of the overclock headroom they need to bite into. AMD is sure to not leave much if anything. The other issue is having an 8 core Intel out and available would yield the benchmarks for AMD to shoot for on Zen-2 right up front.

Unfortunately for you guys it looks like extreme overclocking will become less of an issue. That is unless AMD misses their own predictions and 7nm just does not live up to its hype.

 

That would not be unfortunate at all. It would be great if AMD can correct that shortcoming. The lack of it is why I am not able to view AMD Ryzen and TR CPUs as acceptable options. That would make me happy. I hate monopolies and I don't care about brand name. I just want the best extreme overclocking results. Everything else is secondary. If they were to fix the same issues with their wimpy GPUs and release an extremely overclockable 1080 Ti killer, I'd kick NVIDIA to the curb and take a dump on their face without batting an eye. I'm just not going to settle for parts that do not overclock worth a darn. I would be happy to give AMD my money if they could give me what I want and require in return.

 

What I mean is that AMD will use up all of its overclocking headroom at release. You may see 5.2 GHz at the turbo boost of the CPU's, and I think multicore of 4.8 GHz or so. This will force Intel to do the same and other than silicon lottery push their stock speeds to the extreme at stock as well to try and stay ahead. In the end this may take the easy overclocks away along with the extreme ones as well.

 

i agree. sure its more fun to choose only hardware that overclocks farther. but in the end its all just perspective: if manufacturers decide to clock their chips at stock what they would usually achieve via overclock, then one wouldnt be able to say that "this" or "that" chip is bad because it lacks OC headroom. Because: what is it really about in the end? Having more overclocking headroom to play around with or having an overall higher performance? Zen+ beats CFL left and right when it comes to raw power, but CFL overclocks better (but still doesnt reach Zen+ performance levels due to lesser core count).

At extreme point people will be like: yeah ill rather get myself a 3rd tier CPU and overclock the crap out of it just because it still has OC headroom! Instead of picking the 1st tier CPU which has already been overclocked at stock to the bleeding edge, so no manual OC possible anymore. Still 1st tier will beat 3rd tier any day, any bench even at stock.

Id even go so far as to say that overclocking headroom is rigorously tested and planned ahead of any CPU launch. So our manual testing and OCing here is just an illusion because were just repeating what Intel & AMD already know anyways.... Just that they let you pay more for the "privilege" to be able to manually reach the actual max clocks.

Yeah, getting a bit philosophical here... probably cuz i just finished Bioshock Infinite (that ending! wtf! )

 

Well, that would totally suck. Hopefully, it would never play out like that. What that would ultimately mean is the people that once enjoyed overclocking as a sport and spent lots of money on it will no longer be supported and they will find something else to do. I know I would not spend any more money or time on powerful computers if they do that. No point in having a belly-button bubble gum gamer PC if everyone gets a trophy for participating. In that scenario there is no longer any incentive because it will amount to status quo mediocrity. The higher clock speeds and performance will almost immediately become irrelevant as the new normal. I have zero interest in being numbered among the sheeple, and I am pretty sure most people like me would feel the same way. Socialism is invading the personal computer space, run by Nazi dictators.

 

Didjya pop the lid yet?

That is one heck of a chip if its stable 5.4Ghz @ 1.28-1.3V .Say.. would you be willing to sell it ? xD

 

Yeah. I just did it Friday after I got back from my trip.

Been doing some fitting testing as well...

 

Right now it is not uncommon to see 10%-15%, or even more, on an Intel overclock with some money and effort. In the near future it will be more likely AMD and intel will only leave 2%-5% on the table and some of this may be by silicon lottery. So your sport will take somewhat of a hit and the benefactor will be those sheepeople. I feel your pain but I am all for the general consumer benefitting from this.

 

It would not be "somewhat of a hit" LOL. It would be more like a complete and utter annihilation of the sport. It would cease to exist as far as I am concerned.

I run the 8700K in my desktop with at least a 43% overclock 24/7. (Sometimes 45% when the house is cooler.) Gaming, benching, web surfing... everything. Never misses a beat. If that goes away, I'll have to go find something else to do. Not interested in participating as a general consumer. No point in bothering with it at that point... worthless and boring endeavor if that happens. I'd probably just buy a $500 (or less) BGA turdbook for basic computing tasks and give it up. So, I'm definitely not supportive of the idea of the general consumer benefiting from that (if it happens) because general consumers will be the only thing left. Until then, I'll continue to be optimistic that we won't have to take a knife in the back from Intel like that. I only enjoy participating in things where I can achieve extraordinary results. I am never content with being a follower and almost immediately lose interest in anything that I am not able to stand out as being superior to the status quo. That applies to work and leisure. I'm just wired that way.

 

10-15% overclock on Intel based systems is a gross understatement.

My 5820K had a base and all core boost of 3.3ghz. I ran that thing 4.4ghz all core ( and I could have easily gone further) until I sold it 2.5 years later. That’s 25% difference.

My 8700K has a base 3.7ghz and I run 5.0ghz all day. 26% difference.

I’m sure there are plenty of other examples of higher overclocks such as @Mr. Fox.

I have a feeling Intel may copy AMDs current play of pretty much max out of box performmace to an extent, but they won’t entirely void overclocking. How else will they sell the “K” SKU. Right now Intel still has the clockspeed and IPC advantage. AMD had to throw all they could out of the box to compete on those all important synthetic benchmarks and left little to nothing in the tank for their “unlocked” processors.

 

Some good overclocks you guys are getting on your 6 core 8700K's in terms of percentage increases in performance! I've got a 17.5% overclock on my 4 core Skylake CPU. I think one of the reasons for the large percentage overclocks you're seeing is because the base clock is quite low at stock - I think Intel set it low for CPUs with more cores, because it's got to fit within a certain TDP cooling window. The good thing is that for us overclockers with good cooling we can break those TDP limits with ease (especially with delidding too), and then the limiting factors are just the quality of the silicon and the ability of the motherboard to deliver the power. It's probably more fun & satisfying to overclock these high core count CPUs (as opposed to low core count CPUs) for the challenge of keeping them cool as well as seeing the potentially very large percentage over stock performance increases that can be seen.

 

Should I go ahead and buy a build for the 8700k or should I wait until the next gen comes out? I'm about to hit the trigger on $3,500 worth of parts.

 

You've already got your awesome laptop, I'd probably wait for the next gen of NVidia (or AMD) GPUs and probably combine those with an upcoming 8 core Intel CPU that is based on the 8700K (in as much as having a ring bus rather than the mesh of the current 10 core+ Intel CPUs - ring is better for gaming).

 

this is getting quite ridonculous @Johnksss i love it!

btw, was CB11.5 so harsh on the RAM or why did u have to drop it down to 3600?

 

I say to 10-15% for the general overclockers group, and also why I say even more as this is a truism. I obviously can not address all circumstances and results vary widely. It sems though everyone gets my point though.

 

Looks like people might have to wait a little for NVidia's next gen GPUs!
http://www.guru3d.com/news-story/nv...w-gamers-geforce-is-a-long-time-from-now.html

Basically, NVidia guy was vague & didn't really answer the question, but he did say it would be "a long time from now". I don't know if they'll do a Pascal refresh of some sort in the interim, but that was a comment on the next new architecture. What a snore fest hey! I've been reading the comments on that article and one poster thinks they'll just wait until 7nm is ready for their next gen GPUs, sounds likely to me. No new GPU overclocking fun for a while then - maybe early'ish next year!

 

note that his answer "a long time from now" was specific to a Volta-related question regarding consumer cards. So a Pascal "refresh" wouldnt be out of the question. Maybe on 12nm with higher clocks or some more CUDA cores...

 

Yeah, perhaps, but any Pascal refresh is just gonna mean that proper next gen cards would be pushed back even further, which I think is unlikely. Also, a Pascal refresh isn't gonna be giving us the 50% or so increase in performance that we might see from a new architecture & die shrink - so a refresh is not much to get excited about. Well hopefully they'll just release the next gen cards rather than bothering with a refresh, getting out the next gen faster - fingers crossed!

 

Jensen's response was as expected. June/July was just the rumour mill going wild. I'm still expecting a new GPU announcement come August. That time slot would align perfectly with new gaming GPUs for laptops arriving by the end of the year, as stated by a Gigabyte representative. The main counterargument would be that just because it was scheduled doesn't necessarily mean it still is.

 

EVGA Unveils Closed Loop Liquid Nitrogen Cooling System For Extreme Overclockers-Hothardware.com

We should mention that this is by no means a consumer-grade system. This is just something that KINGPIN and TiN have been working at EVGA on to help further advance their overclocking achievements with minimal fuss. It most definitely is a rather impressive achievement, although we doubt you'll be see anything on the commercial front for a device of this type anytime soon.

 

I'd much rather have that closed loop liquid nitrogen setup that a new GPU.

Hell, I be happy to see my 1080 Ti reign another year on the throne without a GPU refresh if the NVIDIOTS would just knock it off with their retarded Nazi vBIOS bull feces. I'd be happy to ride it even two more years. We have not been able to really exploit all the performance Pascal has to offer thanks to their cancer firmware. I suspect the next gen will be even more screwed up and limited than Pascal, 'cause that's how they roll. Sometimes you get what you ask for, jumping from the frying pan to the fire. Based on observations of their behaviors the past few years, the Green Goblin should never, ever, be trusted to do the right thing for their customers. They'll release the next product when it benefits them the most, and not a day sooner. I suppose it must be pretty easy to reign as the supreme dictator when you have a one-horse race and enjoy a year-over-year-over-year-over-year monopoly.

 

How did you work with the 8700K processor in the P870DM3 model?

on your signing system

 

@Prema to the rescue

 

Anyone using P870DM3 can do it?
Thank you

 

Nope. It takes a special firmware mod, and it is not available publicly. It also requires minor modification of the CPU. Folks that want an 8700K will have to buy a new Clevo (from a @Prema partner shop if they want it to perform correctly) or build a desktop.

http://www.hidevolution.com/evoc-p870tm1-dual-gtx-1080.html

I am sure that @Donald@HIDevolution would be more than happy to help with that.

 

Where can we get firmware? I have no intention of leaving this business

 

It is not available. Not even to @Prema partner shops. It's a proof of concept and for demonstration only.

 

okay. Thank you.

 

I just realized my GPU no longer boosts to 1.093v which now I cant get 2100mhz stable on core, what's up with that? No matter what I do with the curve, it will only go to 1.081v. I even tried going back a couple drivers, still same thing, and I'm tying this while temps are starting out around 33 degrees.

I haven't try an earlier version of afterburner though. Maybe this latest version caused this?

 

Will it let you press "L" and lock it instead of letting it dynamically adjust the voltage? I am using the latest Afterburner and have not noticed any issues so far.

 

Yep that's using L but it still starts at 1.081v and never goes past that. The red lock line is on 1.081v, when I press L on 1.093v, the line turns yellow and never locks there.

 

If you haven't already, try moving ONLY the dot for the voltage you want. Drag it up to the clock speed you want, then "L" and while leaving that window open click the "apply" button. If I try to adjust multiple points on the voltage curve, I literally NEVER get what I want. I just accidentally found that moving only the dot for the voltage I want works perfectly just about every time. I don't adjust any of the other parts in the table. If that doesn't work, not sure what is going on.

 

Yeah I tried that too. I hit refresh, went to the 1.093v dot moved it straight up to 2100mhz, pressed L, then the check mark. The curve went up to the 1.081 dot. I will do a video later when I get home and upload it here so you can see what its it's doing.

Maybe DDU is causing problems? I've been using it quite a lot lately testing different drivers.

I will go back into afterburner and make sure "Force constant voltage" is checked. Even though this was the first thing I checked. I might have to go in and change the bits inside the file. I've never had to do that before though since my GPU is MSI.

I wonder if my Vbios got corrupted somehow and changed my voltage limit on account of that blue screen problem I had about a month ago from the MSconfig problem. That's when I was trying to go to safe mode using msconfig and it was causing all sorts of problems.

 

I should also say that my GPU will try to run 2100mhz at 1.081v but wont hold that clock due to Voltage problems. It will stay there for about 5 seconds then clock down to 2088. Never says thermal related, always voltage related.

 

I'm not sure. It's kind of weird. We have already seen NVIDIA manipulate things and pack a secret payload of filth in their drivers that can have persistent and lasting harmful side effects. Brother @Johnksss and I started noticing that funny business back in 780M/880M days. It is very possible that something extra got quietly added to the vBIOS chip.

 

You said exactly what I was thinking. I wouldn't be surprised if they're doing this deliberately to push everyone towards the new architecture coming out soon what ever name they decide to call it.

What's it take to flash a Vbios on these cards? Do you have to use a programmer to do it now or did someone figure out a way to so it with NvFlash?

 

God, I remember those shenanigans!

I think the NVidia driver influences the clockspeed & voltages seen differently given how quickly temperatures are rising. It's not always related to the ultimate temperature the GPU ends up running, but if temperature rises quickly then it tells the GPU to scale back the core an extra bin than it normally would for that given temperature. I've seen this on my GPU, sometimes it stabalises at 2050Mhz, sometimes 2037Mhz, sometimes 2025Mhz - I used to think it was due to random differences from reboot to reboot, but I was testing some front intake fan cooling mods on the same reboot and different stabalised frequencies were seen even though ultimate temperatures were the same - the cooling mod resulted in the card heating up slower and it didn't step down as far in frequency. Hard to explain in words here, but it does seem to me that the rate of temperature increase of the card is also a factor in frequency & voltage drops, and not just the absolute temperature of the GPU core. Maybe try experimenting with max fans to slow down the rate of temperature increase - it might then decide it can hold max voltage & frequency - might be a long shot, but hey!

 

Now NVIDIA is saying the next GPU architecture will not be released for "a long time" (which is not defined). Who knows what that means. But, NVIDIA always seems to start artificially manipulating (reducing) existing GPU performance when their next thing to sell is getting closer. Their idiocy and treachery was why I began selling my Alienware systems. They totally annihilated functionality of 780M SLI when Maxwell became their new cash cow. Unless I used drivers that were more than a year old, the M18xR1 and M18xR2 and Alienware 18 severely malfunctioned. Yet, those stinkers claimed that they could not replicate the clearly documented problem.

As long as you are flashing a stock vBIOS that has not been tampered with the signature will be valid and you can flash it using NVFLASH. You only have to use a programmer to flash mods because of the Falcon Security Nazi cancer.

 

Back and forth with rumors on next Nvidia... http://forum.notebookreview.com/threads/nvidia-thread.806608/page-33#post-10740604

 

Umm what are your temps @ 2100 Mhz.. ?

 

I used to be able to run superposition all the way through and only get to 48c max (Ambient temp was around 65f) starting at 2100mhz with about 28c reading on and ending up at 2088mhz sometimes 2075. Right now, I cant get no more than 3 seconds at 2100mhz @ 1.081v before it immediately drops to 2088 but I never see anything thermal related, just VOP,Vrel in gpuz.

Im going to mess around with it in a little bit, I will report back here if I find anything.

Here is a pic of HwInfo64 and my Gpu power is grayed out but still taking a reading though?

 

I don't normally see the GPU power reading grayed out like that unless the GeFarts driver is unstable. You might be on the verge of a TDR if the driver is taking too long to communicate with the OS. I finally got fed up with these crappier newer drivers and went back to a @j95 387.92 driver mod. How far back did you go in driver versions?

 

Yeah as @Mr. Fox said, thats because of the driver crash.

Can you take a screenshot of your VF curve?

 

I got it now.

I uninstalled AB and RT then reinstalled them and all is back to normal now. I guess they got borked back when I had to do a clean install the other day maybe?

Also, I noticed RivaTuner was in F drive and Afterburner was in C drive. I dont know if that might've been a factor as well. I got them both in C now.

 

Yeah sounds the curve was stuck or something.

 

Edited my post above your last with Curve.

 

I just went back to 391.24 =when testing, but now all is back to normal. Thanks guys for the help!