Slashdot Mirror


AMD Says Upcoming Zen CPU Will Outperform Intel Broadwell-E (hothardware.com)

Reader MojoKid writes: AMD has been talking about the claimed 40% IPC (Instructions Per Clock) improvement of its forthcoming Zen processor versus the company's existing Excavator core for ages. Zen's initial availability is slated for late this year, with lager-scale roll-out planned for early 2017. However, last night, at a private press event in San Francisco, AMD unveiled a lot more details on their Zen processor architecture. AMD claims to have achieved that 40 percent IPC uplift with a newly-designed, higher-performance branch prediction and a micro-op cache for more efficient issuing of operations. The instruction schedule windows have been increased by 75% and issue-width and execution resources have been increased by 50%. The end result of these changes is higher single-threaded performance, through better instruction level parallelism. Zen's pre-fetcher is also vastly improved. There is 8MB of shared L3 cache on board now, a unified L2 cache for both instruction and data, and separate, low-latency L1 instruction and data caches. The new archicture offers up to 5x the cache bandwidth to the cores versus previous-gen offerings. However, after all the specsmanship was out of the way, AMD actually showcased a benchmark run of an 8-core Zen Summit Ridge procesor versus Intel's Broadwell-E 8-core chip, both running at 3GHz and processing a Blender rending workload. In the demo, the 8-core Zen CPU actually outpaced Intel's chip by a hair. Blender may have been chosen for a reason but this early benchmark demo looks impressive for AMD and its forthcoming Zen architecture.

22 of 188 comments (clear)

  1. Interesting... by fyngyrz · · Score: 3, Funny

    Zen's initial availability is slated for late this year, with lager-scale roll-out planned for early 2017.

    You know, although a tank lager looks big from the outside, there are usually no more than a hundred or so tanks in one. So this doesn't seem like a very large rollout.

    On the other hand, if one of the tanks rolled over the editor(s), that would be a service to humanity.

    --
    I've fallen off your lawn, and I can't get up.
    1. Re:Interesting... by PopeRatzo · · Score: 3, Funny

      lager-scale roll-out planned for early 2017

      I prefer the porter-scale roll-out. Maybe in IPA-scale roll-out on a hot day. Because that's the way I roll out.

      --
      You are welcome on my lawn.
  2. Good to hear. by jellomizer · · Score: 3, Insightful

    AMD has been behind Intel for about a decade now ever since Intel released their "Core" processors. Because back in the early to mid 00's AMD CPU finally were considered serious chips in the desktop environment, outpacing intel. Then it just fell further behind.

    --
    If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    1. Re:Good to hear. by mwvdlee · · Score: 2

      How about performance-per-dollar?

      --
      Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
    2. Re:Good to hear. by Anonymous Coward · · Score: 2, Funny

      He's probably doing it on an old Pentium with the floating point bug.

    3. Re:Good to hear. by GuB-42 · · Score: 4, Interesting

      In fact it was more like they took advantage of the P4 fiasco.
      The NetBurst architecture was a failure, it could barely compete with Intel's own previous generation. They made a few bad design decisions. Perhaps they were blinded by the MHz race, perhaps they really thought long pipelines were the future, I don't know. However, they learned from their mistakes and their next generation (Core) was a success.
      At the same time, AMD took a more sensible approach and the K7/Athlon was a worthy "next-gen" CPU. But maybe the lack or craziness also caused them to stand still when Intel advanced. Intel's commercial practices probably didn't help either...

    4. Re:Good to hear. by avandesande · · Score: 4, Insightful

      Incorporating the memory controller in the CPU and adding the 64 bit instruction set were AMD innovation and had nothing to do with Intel making mistakes.

      --
      love is just extroverted narcissism
    5. Re:Good to hear. by corychristison · · Score: 2

      This is the unfortunate truth.

      I use AMD in my personal machines (Laptop, Desktop, HTPC) because they are cheaper and I get what I need out of them. I dont play video games, and don't really do anything CPU or GPU intensive. A browser, some terminal emulators, file manager, email client, etc. all run great.

      The servers I own are all Intel Xeon E5's. You really can't beat them, and more per watt is definitely better in this space. I run dozens of virtualized servers on each machine and they just keep chugging along without issue.

    6. Re:Good to hear. by jcdr · · Score: 2

      For the 64 bit instruction set, AMD have take advantage of the Intel Itanium mistake and it was there smarter move ever as this enabled AMD to negociate the instruction set symetric cross licencing agreement with Intel.

      AMD did make a lot of technical innovations on CPU before, like copper metal interconnect, silicon on insular, integrated memory controller, hypertransport, multiple cores, exclusive caches, etc..., and lately integrated graphic. At some point Intel haved nothing to compete but there extremly agressive marketing worked long enough to develop a completely new generation of Intel core processors to gently catch up and finally surpassed AMD product. Intel HKMG and smaller gate process finally destroyed AMD that at the same time almost stopped his progression.

      The announce look like there used good design decisions. Time will tell.

    7. Re:Good to hear. by shaitand · · Score: 2

      Intel also played games with TDP while AMD gave max power consumption figures. Intel were/are still the more power efficient chips but not by nearly as wide a margin as the stated power figures lead one to believe.

  3. Re:Kind of rigged test by Anonymous Coward · · Score: 2, Insightful

    Yeah. Some rumors suggest that the Zen line will be disappointingly slow at launch (under 3 GHz). On the plus side, if we can get Broadwell-E performance out of a chip that costs $200, AMD will have a hit on its hands.

  4. Article image? by ilsaloving · · Score: 2

    Why does the article's logo show Intel?

    Has AMD falling so far that Slashdot can't even be bothered to show their logo anymore?

  5. Re:In related news by MiniMike · · Score: 5, Funny

    Bulldozers, Piledrivers, Steamrollers, and Excavators do use diesel engines. Maybe they hired the VW software engineers too late?

  6. Re: Kind of rigged test by Type44Q · · Score: 4, Insightful

    The real comparison will depend entirely on price-points.

  7. Re:"lager-scale" rollout? by ilsaloving · · Score: 2

    Will it be free, as in beer?

    Better than free hookers. I'm not sure I'd trust those.

  8. Re:Of cores not by Rockoon · · Score: 2

    Given the massive layoffs and 2-year delay on 10nm...

    Intel knows its in trouble. Their main two competitors (by semiconductor manufacturing market share) are TSMC and Samsung and both will have 10nm chips rolling out the door a year before Intel even begins testing its 10nm fab. Even Toshiba might beat them to 10nm.

    Zen will be at process size parity with Intel and that will probably equate to similar performance, but this fact is of only minor consideration to Intel. AMD isn't even close to Intels primary competitor. AMD catching up is a symptom of a much bigger problem. This time next year Intel wont be #1 in the market any more...it will be TSMC.

    --
    "His name was James Damore."
  9. Re:Outperform in generating surprise heat? by SScorpio · · Score: 2

    Not when their main competitors Intel and Nvidia beat them in performance while generating less heat.

  10. Re:Kind of rigged test by DuckDodgers · · Score: 3, Interesting

    The Broadwell-E part they benchmarked against is probably a (slightly underclocked) Core i7-6900k, and it's $1100.

    I'm taking wild guesses with the numbers here, but "15% slower than a Broadwell-E at a 45% lower cost and a similar TDP" is a valid market strategy. I haven't spent more than $240 for a CPU in over ten years, if in spring 2017 there are Zen parts at the $250 price point that are 15% behind the 2017 spring equivalent of the Intel i7-6700k or i7-6800k I will buy it.

  11. Re:Kind of rigged test by shaitand · · Score: 2

    Only if they are trying to be the highest performing chip on the block. Being the highest performing chip at a reasonable price point (the $200 mark is generally the sweet spot only the money is no object crowd wastes more on a cpu) is far more important.

    AMD needs a more efficient core. If they have done it, bravo to them.

  12. Will it have a trustzone (AMD PSP) processor? by Anonymous Coward · · Score: 2, Interesting

    I wonder if it will have AMD's equivalent to Intel AMT, the Platform Security Processor. If so, it may be a no-go for some people.

  13. Re:In related news by gweihir · · Score: 3, Insightful

    You AMD-haters are really the most stupid morons around. Don't you realize that the only thing that AMD folding will do is that Intel improvements will vanish and Intel prices will skyrocket? Or maybe you people are into self-abuse?

    --
    Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  14. Re:right.... by gweihir · · Score: 2

    You seem to be stupid, because they just _did_ match Intel clock-for-clock.

    --
    Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.