http://hexus.net/tech/news/mainboard/88898-locked-skylake-cpus-can-overclocked-bios-update/
So it appears that with a simple BIOS update we can now overclock all of the non-K skylake CPUs.
I guess we can finally say without any doubt that AMD IS FINISHED
F
I
N
I
S
H
E
D
Rip in pieces
At least until Intel forces the motherboard manufacturers to "fix" this, like they did with multicore turbo.
>51845898
>implying AMD wasn't finished before this.
This thread is a waste of time.
>>51845923
The community will just release bounty for custom bios.
>tfw $50 overclockable skylake celeron
AMD isn't competitive didn't have competitive CPUs as it was.
So this doesn't affect them in any way.
>>51845898
happened on z87 before and got patched afterwards...
>>51845898
Two things:
1) BLCK overclocking is shit
2) Why the FUCK would you ever buy a Z170 board with a non-K part if you had any intention of overclocking?
>>51846087
You mean just like they did when As rock enabled overclocking on non z series mobos and Intel told them in no uncertain terms to stop? Yeah, that was great.
>>51846087
>celeron
jewtel
Remember. AMD plan on releasing two graphics cards next year.
Not two series of cards.
Two. Cards.
>>51845898
>implying zen wont bring haswell like performance with sick overclocks, shitton of cores and cheaper than skyshit
>>51846309
This is what AMDfags actually believe
>>51846309
>shitton of cores
Well, 1 out of 3 is right.
>>51846301
The best cards ever, anon.
>>51846301
lolwut
>>51846087
you can't overclock extra cores onto your processor
>>51846324
>>51846309
Zen will, most likely, be competitive with Intel's offerings next year. With DX12, Mantle, and Vulkan reaching maturity, what Bulldozer could have, and should have, been is becoming apparent. AMD's offerings for the past few years haven't been bad, it's just that nobody cared to leverage the processors since Intel has always had more market saturation anyway.
AMD isn't bad at making processors, they bet on the wrong horse. The problem was not technical; it was practical. I think AMD realizes their mistake and won't make that same mistake again.
>>51846548
>The problem was not technical
Uhm, yes it is.
AMD processors clock for clock (which is all that matters right now) are on par with Core 2 Duos and Nehalem.
So yeah, it's totally technical.
>>51846597
Well, if you want to ignore the point I made in support of that statement, that's fine, but that doesn't change the fact that we can now see what an FX8350 can do under DX12.
Don't you find it interesting that AMD had to develop their own API before Microsoft begrudgingly added the features necessary to DX to make Piledriver perform? By the way, Microsoft did the same thing to ATI in 2002 when ATI pioneered tessellation. But I guess tessellation is just as shitty as being able to properly leverage anything more than three cores, right?
>>51845898
intel will patch it in no time
How convenient that you can ship microcode updates via Windows Update.
Now intel only need to make sure the users does not decline these updates... oh wait.
$270 - i5 6600k
$420 - i5 6700k
un-fucking-believable
>>51846301
>two cards
>mfw they'll just cut down the cards for the 490x and 490
>mfw the Fury X and Fury will be rebranded as the 380x and 380
>mfw not even I can believe what I'm posting
>>51846942
>$420 - i7 6700k
It's like $500 on Amazon and out of stock everywhere else.
I truly wonder who could be behind this.
>>51846301
>>51846395
>>51846475
>>51846958
They're releasing two chips, which will result in a fair range of cards from binning the chips, just as Nvidia and AMD have done for years.
If you're worried about diversity, worry about Nvidia, who have announced only one chip for next year.
>>51846199
>1) BLCK overclocking is shit
you can get good bclk overclocks on skylake, only sandy/ivy and haslel have bad bclk oc
>prevent the chip from doing something it should have been able to do all along
>hey guise now you can do this!!!
>amd chips can do that from the get-go
>amd is suddenly finished
>>51846199
>BLCK overclocking is shit
Confirmed for underage faggot. BCLK overclocking was only made garbage by Intel when they introduced their K series with Sandy Bridge, i.e. an intentional move to get people to pay more to be able to overclock. They've now untied everything else from the BCLK, presumably safe in the knowledge that you can't do that sort of overclocking on new boards. Except now ASRock have told them to go fuck theirselves and enabled it, and the others will follow suit.
Ultimately, there's nothing Intel can do if the motherboard manufacturers present a united front. They did that with the Z87/Z97 fiasco. Intel were strongly against allowing Devil's Canyon chips to work with Z87 boards, but the motherboard manufacturers told them to fuck off and added support across the board.
Chinks confirmed for alpha status over jews.
>>51846942
Literally 0 improvements.
>>51847097
From what I can piece together from Wikipedia, for the 9xx series nvidia launched seven different GPUs from two chips.
AMD, on the other hand, launched sixteen based on six different chips. A lot of those were retuned rebrands but they still offered good performance for the money.
It'll be interesting to see AMD pursue a more focused approach, perhaps fewer variations will yield more refined products?
>>51847934
>From what I can piece together from Wikipedia, for the 9xx series nvidia launched seven different GPUs from two chips.
three different chips, four if you count maxwell v1 and the 750ti
gm206: 960, 950
gm204: 970, 980
gm200: titan x, 980ti
>>51847885
>They've now untied everything else from the BCLK
But that's wrong you muppet. Go read the review, they specifically state that this gives them increased memory bandwidth and that Asrock have disabled the IGP to allow this. Things are still very much tied up with the BCLK.
>>51847974
No it isn't, you fucking faggot. Nobody cares about the iGPU.
>Thankfully with the latest generation Intel Skylake processors the base clock isn't tied to other components such as the PCIe bus, so tinkering with it doesn't overly stress or introduce instability to other system components.
Reading is hard.
>>51847974
memory clock has always been tied to bclk/fsb, the reason you couldn't OC haslel via bclk effectively is because it was linked to the sata controller and pcie bus
>>51848015
Your words are
>They've now untied everything else from the BCLK,
Now, last I checked, the IGP and the memory were part of "everything", whether you care for them or not. It's not my fault that you have a hard time making yourself understood.
>>51845898
>AMD
>competitive since 2011
:^)
>>51847885
Going back for years and years, the base clock has been linked to things in the system other than just the processor speed, usually the RAM speed or various buses. This was not something new in the past few years. As such, BCLK overclocking has always taken a back seat to multiplier overclocking. In order to get the absolute most out of your system, you'd usually want to use both in varying proportions.
Intel has now developed the Z170 chipset to allow fine grain control over almost everything in the system. This is not some kind of hack being introduced by board manufacturers; it's a feature implemented on the chipset that Intel gave them. Z170 really is an amazing chipset. You should read about some of the other features Intel has implemented on it.
But if you're going to sit there and try to tell me it's worth buying a Z170 board to overclock a piece of shit i3 through the base clock, you're an idiot. Z170 is an enthusiast platform. The feature set is designed around people who will likely be spending over $2000 on their computer-- there's no sense in buying a Z170 board solely for this feature, as you could get a much cheaper H170 or H110 board with an unlocked processor and save some money. So again I say, BCLK overclocking is shit.
>>51847912
>Doesn't even mention clocks and timings used
Unless they used memory with the same latencies, this test is entirely meaningless.
>>51847097
My guess is a very nice high end chip for their enthusiast class, possibly cutting it further for the 490 class or just rebranding Fiji and a good mid range chip to substitute Tonga and target the usual casual market.
>>51846936
Underrated post
>>51849048
Tonga is one of the freshest chips in their line-up. It'll stay.
Hawaii/Granada most likely won't see another refresh, because it's too power hungry to release it as a mid-range card (480s). Likely, Tonga will get refreshed with better binning and better performance/watt.
>>51847097
>worry about Nvidia, who have announced only one chip for next year.
So only the 1080 Ti, new Titan and related Quadros will be Pascal, and the rest will be rebranded Maxlel?
>>51850862
They both do it all the time. 600 series had Kepler, but 700 series was rebrands and big Kepler, i.e. the new Titan and 780Ti.
AMD is already dead.
Zen hasn't even taped out yet, and thus won't arrive until 2017.
Yes, go ahead and buy shitty AMD CPUs, just to feel like a social justice warrior who tries to bring balance in the game.
>>51851573
http://www.kitguru.net/components/graphic-cards/anton-shilov/amd-we-have-taped-out-our-first-finfet-products/
http://wccftech.com/amd-zen-k12-taped/
Yes the second link is wccshit, but the first one pretty much confirmed a tape-out of multiple designs and why would AMD not tape-out Zen when Keller already left (again), indicating the initial Zen design being finished.
>Yes, go ahead and buy shitty AMD CPUs, just to feel like a social justice warrior who tries to bring balance in the game.
If you're this butthurt already, at least properly inform yourself.
>>51851779
>Yes the second link is wccshit
They were bullshitting as usual, it's confirmed not being taped last week by AMD CFO.
http://techfrag.com/2015/12/09/confirmed-amd-zen-hasnt-taped-out-yet-reveals-chief-financial-officer/
