Intel Officially Kills “Tick-Tock”

Discussion in 'HardForum Tech News' started by HardOCP News, Mar 23, 2016.

  1. HardOCP News

    HardOCP News [H] News

    Messages:
    0
    Joined:
    Dec 31, 1969
    According to financial documents spotted by the folks at Motley Fool, Intel has officially killed its yearly tick-tock cycle. Intel even provided a little graphic to explain the new three step process.

    We expect to lengthen the amount of time we will utilize out 14 [nanometer] and out next-generation 10 [nanometer] process technologies, further optimizing out products and process technologies while meeting the yearly market cadence for product introductions."
     
  2. Dr.Diablo

    Dr.Diablo Limp Gawd

    Messages:
    178
    Joined:
    Nov 21, 2011
    Does this mean that we are going to stop seeing a yearly socket?
     
  3. Armenius

    Armenius I Drive Myself to the [H]ospital

    Messages:
    17,297
    Joined:
    Jan 28, 2014
    I guess that is a sign that a silicon replacement is pretty far away.
     
  4. Zigmorph

    Zigmorph Limp Gawd

    Messages:
    240
    Joined:
    Jan 29, 2015
    [​IMG]
    R.I.P. Tick Tock, you will be missed
     
  5. Modred189

    Modred189 I'm Smarter Than You

    Messages:
    14,430
    Joined:
    May 24, 2006
    So, We get a new architecture. Then new versions of the same architecture with better optimizations (like Haswell E?). Then that architecture shrinks to a new process size. Then a new architecture at that new size... and so on?
     
  6. Hornet

    Hornet [H]ardness Supreme

    Messages:
    6,607
    Joined:
    Oct 4, 2005
    No surprise there. It's not realistic to expect a new process node every other release, as it gets ever more difficult to shrink them.
     
  7. Stoly

    Stoly [H]ardness Supreme

    Messages:
    6,141
    Joined:
    Jul 26, 2005
    It was good while it lasted
     
  8. natos

    natos Limp Gawd

    Messages:
    442
    Joined:
    Oct 1, 2010
    AHHhahahh, that's awesome!
     
  9. defaultluser

    defaultluser [H]ardForum Junkie

    Messages:
    12,288
    Joined:
    Jan 14, 2006
    Yeah, it's been clear it's been dead for a year. Every Broadwell release was fashionably late (6 months) before you could buy it in quantities, and Skylake has shown the same issues. I expect similar if not longer teething periods for each new node.

    We already knew that the cost of new process tech is rising so high it costs more per-transistor. Higher costs = more time.
     
    {NG}Fidel likes this.
  10. THRESHIN

    THRESHIN 2[H]4U

    Messages:
    2,946
    Joined:
    Sep 29, 2002
    i say its about time. calling their product cycle 'tick tock' was nothing more than a marketing gimmick that we can do without. i don't know about you guys, but i'd far rather a product misses schedule a bit if it means releasing something of quality.
     
  11. Ultima99

    Ultima99 [H]ardness Supreme

    Messages:
    4,887
    Joined:
    Jul 31, 2004
    Yearly socket? Go ahead, entertain me and attempt to prove this with facts...
     
  12. More like....
    New Node size
    New architure
    Refinement of node to allow larger die sizes (What intel calls optimization)

    Intel is REALLY REALLY good at optimization out of the gate. What they have done with Atom is amazing, given the overhead of intels legacy CPUs. They are only limited in the number of transistors they can lay down to speed things up (reduces number of compute units possible)
     
  13. Red Falcon

    Red Falcon [H]ardForum Junkie

    Messages:
    9,865
    Joined:
    May 7, 2007
    Roughly translated, Intel realizes that Moore's law is slowing down, and they won't have much further down in die-size to go before there is no where else to go with existing technology or the x86 architecture as it is now.
    So, this will allow them to spread out their new processes over and even longer period of time (allowing more sales and gives them more of a buffer before shit hits the fan and/or something completely different must be done [fibre-CPU or quantum computing]) and will allow them to optimize and refine their CPUs even more so than usual to squeeze as much performance out of a generational architecture as possible (which might actually be a good thing).
     
  14. MrPatate

    MrPatate Gawd

    Messages:
    1,023
    Joined:
    Jan 4, 2007
    Back in the days when James Hyneman was still robot looking.
     
    Santa, {NG}Fidel and Armenius like this.
  15. DukenukemX

    DukenukemX [H]ardness Supreme

    Messages:
    4,388
    Joined:
    Jan 30, 2005
    This is what happens when AMD isn't doing well in this market.
     
    sleepeeg3 likes this.
  16. DeathFromBelow

    DeathFromBelow [H]ardness Supreme

    Messages:
    7,118
    Joined:
    Jul 15, 2005
    You know what happened when AMD was top dog with the A64? Intel raised their prices. AMD is too small to matter.

    This particular issue is about physics, not market conditions. The cost of die shrinks is skyrocketing for everyone.
     
    Armenius, {NG}Fidel and N4CR like this.
  17. kbrickley

    kbrickley [H]ardness Supreme

    Messages:
    7,514
    Joined:
    May 13, 2012
    I think it is more about the approach of a fork in the road, as others mentioned. It is ARM and the mobile processor makers that now dominate the market, not AMD and Intel. Intel dominates a market that is saturated and gradually shrinking. They are also approaching a massive technology shift where they will be forced to abandon the silicon technology they have used for the last 40 years. Stretching out the silicon development cycle while they prepare for the Sea Change coming is only prudent business. Especially since there is no indication that PCs will ever return to their boom cycle levels again. Mobile based ARM technologies are likely here to stay and will continue their expansion into the PC territory. PCs won't die as prognosticators like to periodically prognosticate but they will diminish to some commodity level and probably stay there, pending another form of Sea Change that is as yet not on the horizon.
     
    Pieter3dnow likes this.
  18. raz-0

    raz-0 [H]ardness Supreme

    Messages:
    4,496
    Joined:
    Mar 9, 2003
    It was for a while, until 2013. It still gets petty close, and definitely feels that frequeunt when you add in that they often drop a couple at a time and have product delays/availability issues on CPUs.

    CPU socket - Wikipedia, the free encyclopedia
     
  19. NickelFinger

    NickelFinger n00b

    Messages:
    12
    Joined:
    Mar 13, 2016
    Silicon = 27% of the earths crust. Refined Germanium = 21x rarer than refined Gold. You bet their asses they better slow down if they have to change materials to shrink beyond 10nm.
     
  20. chenw

    chenw 2[H]4U

    Messages:
    3,977
    Joined:
    Oct 26, 2014
    Half expected it, considering we are seeing Skylake in force, yet Broadwell is still absolutely nowhere in sight.
     
  21. LightningCrash

    LightningCrash 2[H]4U

    Messages:
    2,412
    Joined:
    Dec 29, 2000
    I like the useless visual aide.
     
    {NG}Fidel likes this.
  22. Comixbooks

    Comixbooks Ignore Me

    Messages:
    12,873
    Joined:
    Jun 7, 2008
    Moors law is the stuff keeps getting smaller but the chips are still hot.
     
  23. LightningCrash

    LightningCrash 2[H]4U

    Messages:
    2,412
    Joined:
    Dec 29, 2000
    Coors law is that the beer tastes twice as good every 1.8 beers.
     
    Emission, Armenius and {NG}Fidel like this.
  24. DukenukemX

    DukenukemX [H]ardness Supreme

    Messages:
    4,388
    Joined:
    Jan 30, 2005
    To be fair, Intel was pushing AMD out of the market through unlawful tactics. So of course they can charge a boat load when everyone is buying from you anyway. Probably the only people who didn't buy Intel were DIY builders.

    Also, it's only a physics problem when you're trying to shrink the die of the CPU. You can still add more cores. Intel Xeons are 28 cores I think. And while non of us here could make good use of that many cores, that's also a software issue. Developers need better tools to use that many cores, like DX12 and Vulkan for example.

    Not the way I see it. Though ARM will definitely move onto the desktop market eventually. The mobile market is dying just as much as the PC market. Both are "dying" because of lack of new things. Why don't 4690K owners go buy a 6600K? Better yet, why doesn't a 2500K owner go buy a 6600K? There's no good reason to do so. Why doesn't a Galaxy S6 owner go buy a S7? You don't, at least if you're sane you don't.

    Intel's big money maker is the server workstation market right now. Cause there's enough idiots in that market that will pay 10x more than the consumer, for the hardware equivalent of the same thing but with ECC memory. Which ECC memory pisses me off cause every PC should have that by now, but whatever. If Intel were to make 8 core CPUs for the same price as a 6600K, then Intel would lose money and raise standards for no reason. With no AMD to compete against them, there's really no reason.

    There's also the lack of software utilizing the hardware. Until open source AI, games that aren't built around Xbox and PS4, and decoding your own genome is more common at home, then nobody needs more than 8 cores. I would also argue there's a lack of software tools as well.


     
  25. krotch

    krotch [H]ardness Supreme

    Messages:
    4,509
    Joined:
    Aug 12, 2004

    Huh? Intel's prices didn't change when AMD was top dog with their A64 and X2s. I wanted to go with an X2 during then but it was about $500 for their cheapest X2. I opted for a 3.4 ghz dual core Prescott for $300 during then. Ran hot as balls, but I was watercooled, so I didn't care.
     
  26. hardboner

    hardboner Limp Gawd

    Messages:
    235
    Joined:
    Feb 23, 2016
    they're so far ahead of AMD that they can just not produce anything new for 5 years
     
  27. Pieter3dnow

    Pieter3dnow [H]ardness Supreme

    Messages:
    6,730
    Joined:
    Jul 29, 2009
    Why don't they produce a chip that can beat AMD APU on both fronts , lets see them do it in 5 years :) Same price category same manufacturing process ..

    Don't worry about ARM either ...
     
  28. krotch

    krotch [H]ardness Supreme

    Messages:
    4,509
    Joined:
    Aug 12, 2004
    Intel already beats AMD's APUs and in the same price category. Not the same manufacturing process, cause AMD is slow. Maybe AMD needs to catch up to Intel's manufacturing process in 5 years instead, so they can actually compete with the same process. I don't see Intel wasting their time waiting.
     
  29. DukenukemX

    DukenukemX [H]ardness Supreme

    Messages:
    4,388
    Joined:
    Jan 30, 2005
    The problem with AMD APU's is why? No seriously, why buy one? The 7850K is currently $115 and that's not a bad price, but you couldn't really game on that. At least not games in medium or higher settings anyway, not including AA. And while no Intel can do that for $115, you could just grab any $100 graphics card and get a 10x better experience.

    We all also forget that all of Intel's consumer CPU's are APU's. They're crap but not totally useless. You could also buy a 860k and throw a $50 graphics card and still get a better gaming experience for about the same price of a 7850K.

    No they don't. A 7850K will pretty much beat any Intel CPU alone. Not with a graphics card. But in my opinion, not enough to make it valuable for gaming. Cause you're going to want a graphics card regardless.
     
  30. krotch

    krotch [H]ardness Supreme

    Messages:
    4,509
    Joined:
    Aug 12, 2004
    The i3 6100 beat the 7850K, not so much in the graphics dept. You have to move into the budget FX and even those get trounced, except in multi-core benches/tests.

    Also I can't imagine anyone wanting to buy those budget CPUs for gaming. Some obviously will since that's what their budget affords, but I would hope they'd pick the i3 or FX procs over an A-series.
     
  31. Uvaman2

    Uvaman2 2[H]4U

    Messages:
    3,010
    Joined:
    Jan 4, 2016
    I don't know i think amds apus are decent.
    AMD A10-7890K APU Review - Introducing The A10-7890K
     
  32. krotch

    krotch [H]ardness Supreme

    Messages:
    4,509
    Joined:
    Aug 12, 2004
    Never said they were bad procs, just don't see them as being all that good either. They're just there. I just think ppl should shoot for the i3 or FX, so they are on an upgradeable socket for future proofing. Granted in the case of the FX proc, you'll need to get a cheap GPU. To try to keep the cost equal, you'd end up with a real crappy GPU. One the APU is going to easily stomp.

    If the user never plans to move away from the APU and stay budget, than I see it being perfectly fine to pick. I'm just not that kind of consumer. I upgrade. I was on Socket 478 for the longest time. As friend's upgraded, I picked up their old procs for like 10-20% of their original price. Went from 1.8 ghz P4 to 3.4 ghz P4, right up until my mobo broke. Then I moved to Core 2 Duo.
     
  33. groebuck

    groebuck 2[H]4U

    Messages:
    2,415
    Joined:
    Mar 9, 2000
    Ke$ha will be suing intel now...
     
  34. Kardonxt

    Kardonxt 2[H]4U

    Messages:
    2,864
    Joined:
    Apr 13, 2009
    New Plan:

    Tick-Whenever the hell AMD releases something competitive.