Read more.Quote:
Removal of BCLK OC feature follows reports of Intel pressure upon motherboard vendors.
Printable View
Read more.Quote:
Removal of BCLK OC feature follows reports of Intel pressure upon motherboard vendors.
Comes down to two choices then: either run your Skylake processor with known bugs and be able to OC or have known bugs removed via microcode update and kiss OCing bye bye. Who'd have thought...
Actually, everyone. I'm only surprised it's ASRock that has bitten the bullet first. Perhaps because they aren't (yet) so big to be able to run afoul of Intel.
Naughty intel. I understand the need to make money selling more advanced overclockable chips but blocking it by pushing motherboard vendors to disable features that have already been sold is not on. Hopefully someone can hack it back in so you can get the latest microcode. What's the betting BCLK overclocking will be hardware blocked for the next gen CPUs?
From what I've heard BCLK overclocking is possbile because the FVIR (fully integrated voltage regulator) has been tossed out of the processor again. Apparently, the Haifa team doesn't care for it much. The next (major) processor iteration will be made by the Hillsboro team again, which introduced (and apparently plans to reintroduce) the FVIR.
This is why we need a competitive AMD to stop intel pulling this kind of stuff.
Not naughty Intel at all, they are simply asking motherboard vendors not to enable unsupported configurations, the CPUs in question are sold by Intel as non-overclockable and they have the K range of CPUs with which they support overclocking.
It's naughty of the motherboard vendors to have enabled BCLK overclocking and advertised it without checking it would be alright with Intel - and predictably it isn't.
I wonder if anyone will care enough to demand a refund for their now non-overclockable motherboard/CPU combo? Intel's actions have made it a blatant case of false advertising
Would be interesting to know if the bclk removal is in the microcode or not.....if it isn't then I guess people can just keep their older bios and use UBU to keep the microcode and various firmwares updated.
Did anyone seriously believe the "Master" would let them get away with it? Meanwhile, the other player in the market has no such issues and that fact is not promoted often enough by the controlled media, including Hexus.
Yes, with AMD there has never been an attempt to inhibit the consumer with such trivial issues.
Well if we look back previous generations of processors have allowed bclk overclocking (nehalem architecture) which, while unsupported by Intel officially, many did. I fail to see how this is any different and not just blatant profiteering from Intel to force people to buy the K series of processors. If I'd purchased one of these boards then I would definitely want a refund. You can't market a product stating it can do something which people may base a purchasing decision on, then later remove it whilst fixing unrelated bugs.
Well, I suppose at the very least the "complex workload" bug will have been fixed. The one recently discovered by the Great Internet Mersenne Prime Search collective.
Going on the article that was fixed in the last BIOS, my Google-fu must be letting me down because i can't find any release note for any microcode updates.
I was just assuming "that" microcode update would be pushed out with the removal of BCLK overclocking. Stick'n'carrot'n'all that. I'm not sure Intel ever details what's in a microcode update. I'd wager most people wouldn't even know if and when the microcode of their processor was updated.
Because of silicon defects, not all 6950 would be physically capable of a 6970 unlock. GPUs with some parts disabled for yield are not just an artificial limitation like Intel's feature fuses - multiplier modifying can't be broken. It's likely that AMD simply had good yields early on which allowed a large percentage of successful unlocks, i.e. parts they could have sold as fully-enabled 6970 but had to sell as 6950 as there weren't enough partly-defective dies to meet 6950 demand. But it is not correct to assume that all GPUs would 'unlock' - same with countless other GPUs and CPU core unlocks. The large-die GK110 didn't even have a fully-enabled SKU for ages after release until yield improved.
It's possible they started properly fusing off cores (rather than just setting cores in firmware), but I can't seem to find any evidence to support that besides a bit of speculation on Google. And there are people with later model 6950 'unlocked' to 6970 which contradicts that. So :shrug:, maybe AMD just had a few batches with a higher defect density back from TSMC?
I'm not saying it's definitely not artificial, some of the time (Only AMD would know that), but it's definitely not artificial on all dies. And to a consumer, I can't think of a way to tell apart faulty+fused cores from working+fused cores.
I remember some people complaining that they bought a 6950 expecting a 6970. A well-known concept that some GPU parts are sold with cores disabled in order to make more dies saleable because of the fact of processor manufacturing which is silicon defects; it had been done for a long time before the 6970 and continues to this day. I gave the GK110 as an example of a part which didn't have a fully-enabled part at all on release until yields improved.
It is completely false to assume every card will unlock. I gave examples of why a large percentage of early cards (certainly not all though) successfully 'unlocked', and it's unreasonable to assume you're guaranteed an unlock upon buying a card which has no guarantee to run with extra hardware enabled. It's also easy to find reports of later production parts successfully unlocking. Assuming AMD did get extremely good yields with their 6970 and found that more dies than not were capable of operating as a 6970 then they would of course been selling lots of them as the 6950 - like with nearly all GPU families they use the same die for both end products.
Recent examples of same-die GPUs being the whole Fury family, 970+980, 980Ti+TitanX, 290+290X, etc etc. In theory a 970 could 'unlock' to a 980 if there were no real defects or hard-disabled cores.
The other point being that AMD, even if they did start fusing off cores, didn't essentially break the 'upgrade' for existing users past their return period or whatever.
Any news if the ASRock Fatal1ty E3V5 will see the light of day?
My guess is you may be able to put this back in with a bios editor (used to put my business name in, instead of award etc, not to mention using other features I'd like to expose). Been a while since I've messed with that crap, but IIRC modbin, cbrom etc would probably lead to some apps to do this (if they aren't still used). They expose whatever the bios hides ;) Useful for Dells etc, but also probably in this case for those with the affected boards. I highly doubt they yanked it out, most likely just changed default from yes to no for a feature that is in the bios. Perhaps someone can check for an FVIR option with one of these and report :)
The might of intel is enormous, its the richest, most powerfull and largest.
No company can really compete with them AMD is a dwarf compared to intel.
They use any means even the bad ones to get what they want, ofcourse making sure evidence is non existant. Asus is not small at all but against the might of intel .... nothing.
The Ati flashing and cpu stuff worked well untill amd decided, to cut the faulty parts with laser.
By that preventing them to ever work again. So if you tried to unlock what worked on the old product, you got fused parts because then you enabled the damaged cutted chips.