Haswell microcode update will block overclocking on non-Z87 chipsets

Scorpus

Posts: 2,162   +239
Staff member

With Intel's Haswell fourth-generation Core processors, overclocking on K-series CPUs is technically limited to motherboards using the high-end Z87 chipset. However, a number of motherboard manufacturers discovered a way around this limitation, updating their existing H87, B85 and H81 boards to allow overclocking.

Intel wasn't too happy that their cheaper chipsets were stealing the features of the more expensive ones, so Intel has announced a new microcode update that will block overclocking on all non-Z87 motherboards. According to a statement released by the company, the update will be released in week 30 of this year, which so happens to be this week.

While Haswell chipsets differentiate on a few specifications, such as the maximum allowed USB and SATA ports, and supported technology such as Intel's Rapid Storage Technology, mostly the chips are the same. Intel uses their best silicon for Z87 chipsets, and as they are built to take the added stress of overclocking, they are the only chipsets to technically support it for K-series processors.

It now becomes a question of how quickly these updates will turn up to kill the workarounds developed by motherboard manufacturers. It's likely that soon we'll see new Haswell processors available with the microcode out of the box, meaning if you're looking to overclock your K-series processor you should stick to the more expensive Z87-based motherboards. Intel has likely completely patched up the workaround, so don't expect manufacturers to find a second one for the H87, B85 and H81 chipsets, at least not anytime soon.

Permalink to story.

 
The overclock unlock was nothing more than a marketing grab by the motherboard vendors in the first place.
The price difference between a lower priced Z87 and a H/B board capable of maintaining a good overclock is fairly minimal, and that price becomes even more marginal if the feature sets of the boards are compared.
I'd be quite keen to find out if the motherboard vendors would honour a warranty claim if their board failed under overclocking conditions. Hard to imagine how, say, this board for example (on Asus's list) with its 4-pin EPS12V, no heatsink, and 3 phase (for CPU) power would fare overclocking a 4770K...especially using the overvolt-happy Asus AI Suite.

I'm sure more than a few people will attempt to lay the blame at Intel's doorstep, when in fact it seems more like a cheap shot by AIB's at Intel's expense. There aren't many instances where a technology partner has the option to arbitrarily alter the designs of an IHV's product/IP.
 
Plus, the Z87 boards may have a longer warranty and more features for a few bucks more.

True, but I find this to be a slap in the face to users and honestly something that should not be implemented. They have already made it hard and expensive to get an unlocked chip as it is and now your going to be required to buy the top of the line board to overclock that unlocked chip.

Now this isn't the end of the world, but the fact that they are doing this AFTER they have had the chips out for awhile and people have already purchased things, this is going to throw everything out of whack and cause some problems with users. If they had waited till Broadwell or done this at the beginning of Haswell, well people might shake one fist but it would be forgotten fast. Doing this is the middle of the life cycle feels like a middle finger to the face for some users and will harm people who may have bought an unlocked i5 and a H87 board planning on overclocking but unable to do anything if they update.
 
True, but I find this to be a slap in the face to users and honestly something that should not be implemented. They have already made it hard and expensive to get an unlocked chip as it is and now your going to be required to buy the top of the line board to overclock that unlocked chip.
First off; if you are going to over-clock, you need top of the line hardware anyway. This microcode will only guarantee, that over-clockers meet the requirement if they are going to push an un-locked chip. I personally don't have a problem with this, unless you bring the topic of what dividebyzero mentioned above into the equation.
 
First off; if you are going to over-clock, you need top of the line hardware anyway
True enough. I'd also note that Intel disclose and publicise the chipset capabilities a year or more in advance of the product going on sale. I'd sincerely doubt anyone would have bought a budget H/B chipset mobo thinking that there was a chance of 1. multiplier overclocking in the future, and 2. having guaranteed system stability with the overclock.
I'd also note, that if the mobo vendors were so interested in giving the customer a better product, why they castrate the chipsets capabilities? Why, because, if they're cutting features then it's product differentiation...whereas if Intel do it, it's....what exactly?
Case in point. Asus have probably the best appointed mainstream motherboards out, and even though the H87 chipset feature set includes 6 USB 3.0 ports (and 8 USB 2.0), Asus's own H87 Plus only features four of those six in order to differentiate the product stack ( those two missing USB 3.0 ports which are part of the chipset feature set will normally cost an additional $US10 by purchasing the H87Plus/CSM )

So you have motherboard vendors squealing about Intel locking them out of features that aren't a part of the chipset, whilst simultaneously withholding included chipset features to increase their profit line.
I also don't see a lot of consumers jumping on the mobo vendors for the withholding of features...maybe Intel make a more compelling target.
Now this isn't the end of the world, but the fact that they are doing this AFTER they have had the chips out for awhile and people have already purchased things, this is going to throw everything out of whack and cause some problems with users
The H87 chipset at launch didn't support multiplier overclocking- and officially still doesn't. AsRock made the decision to make multiplier overclocking available on a single motherboard four weeks ago (also note the disclaimers). Asus , ECS, and Gigabyte provided the same option two and three weeks ago. For this to cause problems for the users:
1. They would have had to have bought an (ostensively) non-OC motherboard and a K CPU between 27 June and 25 July (AsRock), 9 to 25 July (Gigabyte and ECS), or 15 to 25 July (Asus).
2. Been aware that a BIOS had been made available that supported overclocking, but...
3. Not noticed the disclaimers attached with the beta BIOS's and,
4. Be oblivious to the fact that Intel would almost certainly close off the feature. As noted by many sites that reported the added "feature"
There's no guarantee Intel won't issue a platform firmware update that closes the loophole allowing overclocking on other chipsets- Tech Report: July 8, 2013.
 
True enough. I'd also note that Intel disclose and publicise the chipset capabilities a year or more in advance of the product going on sale. I'd sincerely doubt anyone would have bought a budget H/B chipset mobo thinking that there was a chance of 1. multiplier overclocking in the future, and 2. having guaranteed system stability with the overclock.
I'd also note, that if the mobo vendors were so interested in giving the customer a better product, why they castrate the chipsets capabilities? Why, because, if they're cutting features then it's product differentiation...whereas if Intel do it, it's....what exactly?
Case in point. Asus have probably the best appointed mainstream motherboards out, and even though the H87 chipset feature set includes 6 USB 3.0 ports (and 8 USB 2.0), Asus's own H87 Plus only features four of those six in order to differentiate the product stack ( those two missing USB 3.0 ports which are part of the chipset feature set will normally cost an additional $US10 by purchasing the H87Plus/CSM )

So you have motherboard vendors squealing about Intel locking them out of features that aren't a part of the chipset, whilst simultaneously withholding included chipset features to increase their profit line.
I also don't see a lot of consumers jumping on the mobo vendors for the withholding of features...maybe Intel make a more compelling target.

The H87 chipset at launch didn't support multiplier overclocking- and officially still doesn't. AsRock made the decision to make multiplier overclocking available on a single motherboard four weeks ago (also note the disclaimers). Asus , ECS, and Gigabyte provided the same option two and three weeks ago. For this to cause problems for the users:
1. They would have had to have bought an (ostensively) non-OC motherboard and a K CPU between 27 June and 25 July (AsRock), 9 to 25 July (Gigabyte and ECS), or 15 to 25 July (Asus).
2. Been aware that a BIOS had been made available that supported overclocking, but...
3. Not noticed the disclaimers attached with the beta BIOS's and,
4. Be oblivious to the fact that Intel would almost certainly close off the feature. As noted by many sites that reported the added "feature"

Its still a middle finger to the face being a customer or a manufacturer. Theres no reason for this other than they are worried people wont buy Z87 (Which that's not gonna change, z87 has way more features anyway). I personally would buy a Z87, but what about those who didn't and now are going to be locked out? It does not matter how its justified, its an unnecessary thing to do, H87 or the others are not butchering or harming the sales of Z87, hardcore overclockers and most people wanting to overclock will still grab one, the fact that companies found a clever way to incorporate some overclocking on their lower boards and Intel slammed its fist yelling unacceptable (Dramatization, whatever). The problem is with doing this in the middle of the chips cycle rather than the beginning, if it was announced first that your guaranteed never to be able to do this, then fine everyone knows, but to those who bought one thinking this will be fine, they are going to be out of luck now and either forced to sell or return. Its the timing, not the whole idea that's the problem.

Timing is everything, and removing things in the middle of the chip cycle isn't something I think they should do.

First off; if you are going to over-clock, you need top of the line hardware anyway. This microcode will only guarantee, that over-clockers meet the requirement if they are going to push an un-locked chip. I personally don't have a problem with this, unless you bring the topic of what dividebyzero mentioned above into the equation.

Yes and no, some people may just want a minor overclock to like 4 ghz on a 4670k, I agree that for hardcore overclocking, you want a Z87, 990fx, or whatever for what chip you have. However, if your not going crazy with that type of stuff and want the option, then why is it so taboo?
 
Yes and no, some people may just want a minor overclock to like 4 ghz on a 4670k
Seems like a miniscule cross section of potential users just got even smaller. Even with the standard BIOS, the H87 supports all-core turbo and by extension, disabling of power saving...so your potential disgruntled user base now includes people that would want a 4GHz OC that aren't satisfied by the all-core settings in the BIOS that allow 3.8GHz.

I'm kind of wondering what the demographic is for people who'd sacrifice feature set ( 4 lanes of PCI-E 3.0 for I/O or add-in card most noticeably w/ H87, that plus removal of RST and two SATA 6Gb posts for B85) and strength of product (voltage supply, stability, and PCB layers) for a minimal cost saving all for the sake of a ~0.2GHz gain in potential overclock.
The problem is with doing this in the middle of the chips cycle rather than the beginning
LGA 1150 chipsets were launched eight weeks ago...hardly "the middle of the chips cycle"...unless you're expecting Z97/H97 to arrive in September 2013
if it was announced first that your guaranteed never to be able to do this, then fine everyone knows,
Everybody did know. Intel's Lynx Point chipset was detailed in November 2012. The overclocking exploit arose from a bug found and publicized a month after the chipsets launched. This isn't a case of Intel withdrawing a feature, it's a case of motherboard vendors circumventing Intel's specification. In point of fact it's not a great deal removed from the exploits used by motherboard vendors to unlock cores on AMD CPUs using the EC Firmware function (SB7xx)- an exploit AMD moved to shut down only for vendors to circumvent AMD with hardware switches and proprietary circuits to nullify AMDs ACC. The only real difference is that AMD chipsets don't sell in high enough volume for AMD to enforce its rule of law...unlike Intel.
but to those who bought one thinking this will be fine
Not one of the motherboard vendors publicized the exploit as part of the feature set of the boards in box art, product page, or (r)etail listings. The only way a potential owner would have known would be if they were updating their BIOS or frequented a tech site that carried the news....and of course, if that is where they got the information they couldn't help but see the disclaimers and general opinion that Intel would close the loophole.
Its the timing, not the whole idea that's the problem.
The exploit was available for between 10 and 14 days before Intel notified everyone that they wouldn't continue to sanction the vendors actions. That is actually a pretty quick response time given the time needed to research the bug and communicate with the motherboard makers
 
That does not matter, it is a quick response, but the point of this locking the chip is mute. They were worried about sales of Z87 while doing this, I mean I highly doubt it was going to harm the sales (Much at least) so it really makes you wonder why out of all things that they could be doing, this became a top priority.

What about the person who for all intensive purposes maybe bought a H87 and i5 4670k because he saved the 20 bucks extra to get that K variant and overclock it lightly. That might be a minority of people, but theres probably some people out there like that.

Now is this the worst thing ever, no, its not a big enough deal to riot against Intel or something like that, but its also just a slap to manufacturers who found a way to do this blocking out a creative way to let the other chipsets have some bonuses. A hardcore overclocker is still going to buy a Z87 and most people who buy a k variant chip are still going for a z87 (And probably were before), its just a little thing that harms those who thought that this was an option.

No, the middle of comment was not saying literally the middle, I meant that the chip had been out for awhile now and they are doing this now, even if it only took up to 14 days to stop it, it was still there and now its going to be gone.

I never said it was directly an intel feature, mainly a manufacturer add on and "Exploit" as you said to add some features to the other boards. I don't think this was a warranted emergency patch needed and worthy of making such a big deal out of. Like stated before, most of the boards don't even contain heatsinks or high enough power to handle high overclocking anyway, so it was most likely to be used for slight boosting performance.

That's my opinion on the matter, the AMD locking out the 4th core you posted (Im assuming you were talking about the 720 BE that you could unlock the fourth core on right. Well yea, im sure AMD was not happy with it and they probably could not block it as easily as Intel could have, they probably would have too and I would have been right up there with a similar complaint. Its similar to the original Refernce line of 6950's that had the dual Bios option where people would unlock the cards shaders to a 6970 (Or close at least). It was an exploit, and AMD did not make that big a deal of it or force Catalyst to block that or something like that (Later cards could not be unlocked with the removal of the dual bios option from what I saw, but still).
 
Back