Intel Shutting Down Skylake Overclocking Loophole

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
According to this article, Intel will be shipping a new microcode update for Skylake processors that will put an end to overclocking non-K CPUs.

”Intel regularly issues updates for our processors which our partners voluntarily incorporate into their BIOS,” an Intel spokesman said. “The latest update provided to partners includes, among other things, code that aligns with the position that we do not recommend overclocking processors that have not been designed to do so. Additionally, Intel does not warranty the operation of the processor beyond its specifications.”
 
And so continues the vicious cycle of Intel trying to prevent overclocking on most parts, and the motherboard manufacturers figuring out how to bypass Intel's controls. Same shit, different day.
 
Aka intel doesn't want you doing what you have been doing with their processors for the past 2-3 decades: getting extra performance for free.
 
how long before we do not actually "own" the part, but merely license its use?
 
getting extra performance for free.

I thought that was the idea. Not spend more for a K chip. It's basically just a more expensive, faster chip then that lets you feel like you're doing something.

According to this article, Intel will be shipping a new microcode update for Skylake processors that will put an end to overclocking non-K CPUs.

”code that aligns with the position that we do not recommend overclocking processors that have not been designed to do so. ”

Thanks for looking out for me. Whew, I might have wrecked the darn thing otherwise!
 
Aka intel doesn't want you doing what you have been doing with their processors for the past 2-3 decades: getting extra performance for free.

*nods*

Intel design board: Let's milk that 2% of the market that overclocks for that couple hundred extra dollars. Bad publicity? What bad publicity. They are 2% of the market. Yet them yell.

And on the same note:

Intel design board: Why actually make something faster when we can just give you small base clock speed increments and modest IPC improvements with a lower overall top overclock speed as compared to the previous generation.

And they wonder why PC shipments are in the toilet.
 
how long before we do not actually "own" the part, but merely license its use?
It's already here. Pretty much everything that has software in it is that way. Sure you "own" the hardware you bought, but the code that let's you run it is licensed.

There is a reason why they want all the OS software you use to be able to be "updated" without your permission. It was first due to being altruistic in fixing bugs, now it's a full on attack on the people when they do things outside the norm. Even if you don't update your bios so you can keep the OC-ing functionality, you be sure MS gets an update from Intel that breaks your system with Windows 10's forced update ensuring you will update the bios.
 
According to this article, Intel will be shipping a new microcode update for Skylake processors that will put an end to overclocking non-K CPUs. [/I]

So how long until someone compares the microcode in a ne wBIOS with an older one and figures out how to delete the new microcode section or replace it with an older microcode section that allows the overclocking? OR even learns how to re-enable overclocking by editing the new microcode?
 
This is what happens with a lack of competition. That standard used to be cranking up the FSB and tweaking settings to make it work which sounds pretty similar to what these Skylake CPUs require to overclock.
The only difference I see is the increased market dominance Intel has now and the complete lack of AMD mindshare.
We just have to use what they are willing to sell us. There is no alternative so the company is going to focus on profitability and not increase our bang-for-the-buck. I half expected the Pentium anniversary edition to start something but it must not have been as popular as I thought it would be.
 
I'm going to be the Devils advocate here. To the best of my knowledge, lower end chips make up the majority of PC sales. Some small PC retailers have been know to sell OC systems and then blame viruses/malware/etc as the reason for instability. Not only does Intel loose out on selling a higher priced chip but their brand name gets a bad reputation. Might not be such a big problem in the US but I've fixed several fly-by-night systems that had me shaking my head. But that's not the only reason. Getting a worth while OC pretty much requires a better Heat sink, motherboard, RAM and PSU. A conservative 50+ USD per item puts you at 200USD more...
Humm what cant I get for 200USD?.....A better processor? Or maybe a SSD and Video card that would make a more tangible difference.
 
I'm not sure what to make of this.

Current K series are not as overclockable as the allmighty SB. And for the most part overclocking does very little nowadays. Certainly not worth it for gaming.
 
Intel design board: Let's milk that 2% of the market that overclocks for that couple hundred extra dollars. Bad publicity? What bad publicity. They are 2% of the market. Yet them yell.

Intel design board: Why actually make something faster when we can just give you small base clock speed increments and modest IPC improvements with a lower overall top overclock speed as compared to the previous generation.

And they wonder why PC shipments are in the toilet.

This.

I always used to buy the lower/mid range CPU's and over clock them to match the performance of the higher-end chips. With the overclocking lock down and the introduction of the K chips, they raised the cost of buying a chip than can over clock too much. Non over clocked chips are not much faster than what I already have.

I used to upgrade every few years, but now I'm still using my 6 year old i7, and other family members are still using overclocked i3's from the same generation. The system still do what we need, and it's not worth spending hundreds for a small boost, or a lot more for a K chip to get a larger boost.
 
I guess I must have missed it. Are they patching all existing CPUs to lock them down (software), or just new ones going forward (hardware)?
 
im sure this looked good in the board room, on paper, but i see AMD being able to capitalize on this. The Black addition comes to mind.
I can see it now ZENFX BLACK! :cool:
 
Aka intel doesn't want you doing what you have been doing with their processors for the past 2-3 decades: getting extra performance for free.

People posting this either haven't been around that long or they have short memories. Intel has NEVER liked the idea of anyone gaining performance for free. Overclocking via Extreme Editions and K SKUs has always been more about marketing toward the enthusiast than anything. Overclocking lower end processors to higher end equivalents allows people to buy less expensive and therefore lower profit margin CPUs. At the other end of the scale, its a different story. Overclocking at this level probably benefits Intel more than it could hurt ever hurts them. It entices more users to buy those high end CPUs and it showcases what Intel's products are capable of. Lots of people getting good results from these CPUs on forums can equate to more sales. More importantly, Intel tolerates it on the highest echelon of CPUs because these are at the top of the product stack. Overclocking these CPUs doesn't cannibalize its own sales.

Throughout the bulk of Intel's history, it has NEVER wanted anyone overclocking. In the 1990's they had problems with people reselling processors over seas after remarking them. Back then all you had to do was set your motherboard's jumpers for a different processor and pop yours in. Sometimes it worked, sometimes it didn't. This led to things like certain revisions of P54CS Pentium CPUs which were multiplier locked. A system running one of these not so bad boys wouldn't POST with the CPU configured incorrectly. Aside from a few examples, Intel hasn't had much choice in the matter because motherboard manufacturers had control over external inputs such as voltages and bus clocks.

Intel always makes moves which people believe are designed to prevent overclocking. Consistently, motherboard manufacturers have found ways around the limitations put in place by Intel. It's all about market segmentation for Intel. They want you to spend as much money on their computing hardware as possible. This is not only done by giving you superior products to its competition (which frankly barely exists now), but by making sure that there is a product in every price point so that people of various income levels can contribute to it's bottom line.

Intel does a lot of things to ensure this happens. Joe Sixpack buying a $200 processor and overclocking it to the same performance levels as a $1,000 CPU goes against that. Sure Joe Sixpack bought something from them, but he spent $200 instead of $1,000. So Intel's got to ensure that you buy the products you can afford (or not) at the highest margins possible. Beyond segmenting through price, they segment through features. Reduced cache, lack of VT-d, or reduced core counts, PCIe lanes, or whatever helps heard people toward the more expensive offerings.

This is how Intel has pretty much ALWAYS worked. This isn't a sudden move, or something more underhanded or devious than they've done in the past. Just like any other business. It has never wanted anyone overclocking. Intel would probably put an actual end to overclocking if they didn't think it would hurt their business by doing so. Even with this microcode update, overclocking is in a better place than it's probably ever been. Intel allows it on K and X SKUs, Many Xeons are also unlocked allowing users flexibility in the clock speeds of their CPUs. Intel even offers extended warranty coverage for CPUs that are going to be overclocked. Internally, to vendors and reviewers Intel openly discusses overclocking. 10 years ago, that wouldn't have been the case.

UEFI / BIOS rules are better than ever. More motherboards are better made and are designed to allow for overclocking by having the necessary supporting voltage hardware for it. AIO water cooling units and higher end heat sinks allow for easy cooling of substantial heat load.

Doom and gloom about the death of overclocking today is as premature as it was when Haswell was released, or the half dozen other times we heard it in the media.
 
You guys crack me up. If Intel had corrected the "loop hole" before it was publicly known, you'd have nothing to nerd rage about.
 
There are two sides to this:
1. People buying low end chips expecting top end performance when they overclock. They probably end up burning the chip down in the process and then try to submit it for RMA expecting a new chip - rinse and repeat.
2. In some ways its a dick move by Intel because so many are accustomed to overclocking chips of all sorts, but again it comes back to point 1.

I dont see what the big deal is. I buy chips based on what they offer, not so much what I can OC them to. 10-20% overclock on my 4790k isnt that big of a jump from stock speeds, so overclocking is not really a concern at the end of the day.

If you are a true enthusiast you will get the chip that is meant to overclock, simple as that.
 
There are two sides to this:
1. People buying low end chips expecting top end performance when they overclock. They probably end up burning the chip down in the process and then try to submit it for RMA expecting a new chip - rinse and repeat.
2. In some ways its a dick move by Intel because so many are accustomed to overclocking chips of all sorts, but again it comes back to point 1.

I dont see what the big deal is. I buy chips based on what they offer, not so much what I can OC them to. 10-20% overclock on my 4790k isnt that big of a jump from stock speeds, so overclocking is not really a concern at the end of the day.

If you are a true enthusiast you will get the chip that is meant to overclock, simple as that.

Solution to #1, have the processer log into specialized memory the last speed it ran at. If people try to RMA it, they're shit outta luck.

Samsung did this with their bootloader on their phones. They didn't want you to overclock them, but they saw the backlash from when HTC started locking their bootloaders. Their answer? Install special fuses that blow when the bootloader modified and would display on screen at bootup in the corner. Prevented users from doing a warranty return but also allowed them to do what they wanted with the device. Win/win.
 
I have not heard of one person who burnt out a CPU recently.

That said

It's quite simple. Intel has a disclaimer on the warranty for non-K parts. If you overvolt, undervolt, or overclock, your warranty is VOID. But keep the overclock available.

And this information can be BURNED into the CPU once you try to overclock.

Intel can then provide a validation utility part of CPU-Z that identifies if the chip has been overclocked/overvolted and the warranty is void. If you are a motherboard vendor and supply software which overclocks via software or BIOS, then through license agreement with Intel, you must warn the user that overclocking/volting non-K parts will void the warranty on the CPU. Also the BIOS boot screen will display if the CPU warranty is void with "CPU WARRANTY VOIDED: Overclocked"

So if customers @#%@# about Intel not honoring the warranty, they really can't complain.

If you are a good OEM System vendor, you disable overclocks with your own custom BIOS that locks that stuff out.
 
...If you are a good OEM System vendor, you disable overclocks with your own custom BIOS that locks that stuff out.

Pretty much my attitude. If you want to overclock buy a board with a BIOS that supports it. Ain't that what we've been doing for years?
 
1. People buying low end chips expecting top end performance when they overclock. They probably end up burning the chip down in the process and then try to submit it for RMA expecting a new chip - rinse and repeat.

I've been overclocking Intel and AMD chips since my first 386/25 cpu.
I've NEVER tried to RMA a CPU because it failed under over clocking, because I've NEVER had one fail. I did have a couple old AMD CPU's that where terrible over clockers (366mhz out of a 333mhz cpu), but they just weren't stable or wouldn't boot.

I do miss the old days, like when I bought a Celeron 300a and over clocked it to 450, making it faster that one of my fellow gamers 400mhz P2, that he paid over 4 times as much for :)
 
I've been overclocking Intel and AMD chips since my first 386/25 cpu.
I've NEVER tried to RMA a CPU because it failed under over clocking, because I've NEVER had one fail. I did have a couple old AMD CPU's that where terrible over clockers (366mhz out of a 333mhz cpu), but they just weren't stable or wouldn't boot.

I do miss the old days, like when I bought a Celeron 300a and over clocked it to 450, making it faster that one of my fellow gamers 400mhz P2, that he paid over 4 times as much for :)

I had an old intel pentium that i could run the internal clock @ 183 mhz or something crazy like that. It was not always stable, but it never died on me. My old opteron would crank up too, those server class processors could really handle it. Still have it in a box somewhere. We are spoiled today when it comes to overclocking, compared to years ago. Having to use jumpers was a pain :D
 
I have not heard of one person who burnt out a CPU recently.

I've worked with thousands of CPUs and several dozen of those or so were overclocked. I've seen two Intel CPUs that actually failed under normal circumstances and three that died due to power surges that took out the motherboard. I've seen maybe one or two Intel CPUs killed by overclocking. Even then they didn't die immediately. They both became less and less stable, and requiring more and more voltage to run. The last time I saw this happen was with a Core 2 Duo, the first one was a Pentium III.

I've seen far more dead AMD CPUs, but only four or five failed due to being defective. The other dead AMD CPUs were either cracked T-Bird cores due to the heat sink being put on improperly. I 've seen probably three or four Thunderbirds burnt up through overclocking while being insufficiently cooled. I've intentionally fried a couple of Duron's because I found it amusing. The rest of the dead AMD CPUs I've ever seen died as a result of the motherboard failing and their shitty voltage hardware taking out the CPU as well.

The point is, CPUs are actually quite difficult to kill. They always have been. Overclocking became far more common when motherboards began allowing tuning through the BIOS. Even In the early days of overclocking people killing CPUs was relatively rare. With thermal protection coming in multiple forms now, it's damned near impossible to kill a CPU on most motherboards.
 
I've worked with thousands of CPUs and several dozen of those or so were overclocked. I've seen two Intel CPUs that actually failed under normal circumstances and three that died due to power surges that took out the motherboard. I've seen maybe one or two Intel CPUs killed by overclocking. Even then they didn't die immediately. They both became less and less stable, and requiring more and more voltage to run. The last time I saw this happen was with a Core 2 Duo, the first one was a Pentium III.

I've seen far more dead AMD CPUs, but only four or five failed due to being defective. The other dead AMD CPUs were either cracked T-Bird cores due to the heat sink being put on improperly. I 've seen probably three or four Thunderbirds burnt up through overclocking while being insufficiently cooled. I've intentionally fried a couple of Duron's because I found it amusing. The rest of the dead AMD CPUs I've ever seen died as a result of the motherboard failing and their shitty voltage hardware taking out the CPU as well.

The point is, CPUs are actually quite difficult to kill. They always have been. Overclocking became far more common when motherboards began allowing tuning through the BIOS. Even In the early days of overclocking people killing CPUs was relatively rare. With thermal protection coming in multiple forms now, it's damned near impossible to kill a CPU on most motherboards.

I remember when Patric Norton "Screen Savers" forgot to plug the CPU fan in on an Athlon processor, smoked that sucker in 30 seconds. :D
 
Back
Top