Intel Announces Sunny Cove Brings 18% IPC Improvement

Discussion in 'HardForum Tech News' started by cageymaru, May 28, 2019.

  1. cageymaru

    cageymaru [H]ard as it Gets

    Messages:
    19,815
    Joined:
    Apr 10, 2003
    Wikichip has an article on the latest information about Intel Sunny Cove core optimizations. Intel Chief Core Architect Ronak described the 18% IPC increase across various workloads in this manner. “The last time we have had something in this range was during the Merom era,” Ronak, Intel Chief Core Architect, said. He added, "I have had some of you ask me, “can you actually push the IPC further anymore or are we running out of steam doing this?” And the answer is: No. And Sunny Cove is one example of this. The other microarchitectures we are working on back at home will go even beyond this. Sunny Cove is the foundation for our next wave of innovation."

    Make sure to read the article's beginning to understand how Intel improved various aspects of Skylake to create Ice Lake and then changed the game with new technology in the Sunny Cove cores.

    intel-sunny-cove-ipc-claim.png

    intel-sunny-cove-ipc-claim2.png

     
    lostin3d, DejaWiz, AceGoober and 3 others like this.
  2. Jim Kim

    Jim Kim 2[H]4U

    Messages:
    3,547
    Joined:
    May 24, 2012
  3. Keljian

    Keljian Gawd

    Messages:
    647
    Joined:
    Nov 7, 2006
    sounds like desperate scrabbling to me.
     
  4. Lakados

    Lakados [H]ard|Gawd

    Messages:
    1,720
    Joined:
    Feb 3, 2014
    Until I see “real” benchmarks, AMD and Intel are just pissing on walls asking me to look at how high it went.
     
    HAL_404, gigatexal, Roberty and 32 others like this.
  5. Pieter3dnow

    Pieter3dnow [H]ardness Supreme

    Messages:
    6,789
    Joined:
    Jul 29, 2009
    If you check the images check the cache comparison one :)
     
    chameleoneel likes this.
  6. 777

    777 Limp Gawd

    Messages:
    137
    Joined:
    Nov 8, 2015
    Hasn't Intel been touting 15% IPC improvements in Sysmark for each generation of the last 5 years?
     
  7. Red Falcon

    Red Falcon [H]ardForum Junkie

    Messages:
    10,027
    Joined:
    May 7, 2007
    So this will make up for, what, one or two of the recent hardware flaws' performance losses?
     
    Manny Calavera, Roberty, SIS and 17 others like this.
  8. sirmonkey1985

    sirmonkey1985 [H]ard|DCer of the Month - July 2010

    Messages:
    21,583
    Joined:
    Sep 13, 2008
    yes and in reality it's 3-5% on average per generation depending on the benchmark used(yes there are some outlier synthetic benchmarks that get into the 10-15% range like sysmark).. it's the standard cherry pick a benchmark and claim that's the real number that all hardware companies use.
     
  9. SmokeRngs

    SmokeRngs [H]ard|DCer of the Month - April 2008

    Messages:
    15,837
    Joined:
    Aug 9, 2001
    It's just that AMD's number with regards to Ryzen have actually been much closer to the truth than Intel's numbers. Generation after generation Intel's numbers have been almost pure fantasy.

    That's why I'd expect at best a 5% increase from Intel which will probably get walked back after the next exploit fix.
     
  10. Ikasu

    Ikasu Gawd

    Messages:
    872
    Joined:
    Jul 24, 2007
    If you read the fine print in the image as well as read other articles out there. They are comparing it to Skylake. Good job intel. Compare your "new" architecture against four generation old tech, 18% indeed. Intel's desperation is just astounding considering their laughable 28 core for HEDT, and these bloated numbers which don't relate at all compared to Ryzen 2 to 3. an 18% IPC improvement over a 4 year old architecture is just pitiful. As someone who has has intel chips for the past 12 years, so damn glad we have some competition in the market now. We'd still be on quad core i7 chips if it wasn't for ryzen, MAYBE hexcore.

    Going to be dumping my 14 core x299 rig and moving over to the updated threadripper build as soon as x499 becomes available.
     
  11. Delicieuxz

    Delicieuxz Gawd

    Messages:
    979
    Joined:
    May 11, 2016
    Good point.

    Hopefully, those flaws will be fixed on the hardware side with Sunny Cove so that there won't be a performance offset from them. But, I suppose there might not be time to make changes to the processor that will fix the most recent ones.
     
    Red Falcon likes this.
  12. SmokeRngs

    SmokeRngs [H]ard|DCer of the Month - April 2008

    Messages:
    15,837
    Joined:
    Aug 9, 2001
    Hardware mitigations do not automatically bring back lost performance. All they do is reduce or remove the flaw/exploit from being able to happen on the hardware instead of having software fix the problem with the hardware still being vulnerable. There is nothing which guarantees that any performance penalty will be reduced or eliminated.
     
  13. Araxie

    Araxie [H]ardness Supreme

    Messages:
    6,386
    Joined:
    Feb 11, 2013
    everything from Skylake to coffe lake are the same in terms of architecture and IPC. Intel has been just reusing the same uArch and optimizing the process fabrication but nothing else. they are the same.
     
  14. CombatChrisNC

    CombatChrisNC [H]ard|Gawd

    Messages:
    1,131
    Joined:
    Apr 3, 2013
    Too bad we won't be getting a [H]ard benchmarking comparison.
     
  15. N4CR

    N4CR 2[H]4U

    Messages:
    3,969
    Joined:
    Oct 17, 2011
    Intel marketing is laughable these days, sorry Kyle. Raja was pulling this shit in RTG and it got him a 'sabbatical' eventually.
     
  16. SvenBent

    SvenBent 2[H]4U

    Messages:
    3,145
    Joined:
    Sep 13, 2008
    18% boost compared to former products with Security patches in places...

    Remember the MMX benchmark with 400% performance improvement...
     
  17. THUMPer

    THUMPer 2[H]4U

    Messages:
    3,061
    Joined:
    May 6, 2008
    When will this be released? 2021?
     
  18. GreenOrbs

    GreenOrbs Limp Gawd

    Messages:
    372
    Joined:
    Jun 7, 2017
    These charts are really a bit deceptive.... the first one has 18% vs "SKL core." AMD's 15% IPC was year to year improvement not 4 years since Skylake. The 1.48x clockspeed vs 5th gen broadwell... compare it with the 1.42x of 8th gen and the year to year gain is about 4%.

    EDIT: Might not be such a bad comparison after all... Apparently, there's been barely any IPC gain over the last 4 years... maybe a few percent or none at all according to araxie.
     
    Last edited: May 28, 2019
  19. notarat

    notarat [H]ard|Gawd

    Messages:
    1,777
    Joined:
    Mar 28, 2010
    18% IPC? At the cost of what % increased security vulnerabilities?
     
  20. Araxie

    Araxie [H]ardness Supreme

    Messages:
    6,386
    Joined:
    Feb 11, 2013
    as said above, Skylake to coffe lake are the same architecture. another fact, broadwell due the huge 128mb eDRAM L4 cache pool allow better IPC than the newer Skylake but that made broadwell way more expensive.. if you think there's an architectural difference between 6700k and 9900k then you are wrong, they are the same but with more cores.
     
  21. IdiotInCharge

    IdiotInCharge [H]ardForum Junkie

    Messages:
    12,063
    Joined:
    Jun 13, 2003
    Sunny Cove was designed to compete with Skylake- and since AMD still hasn't caught up to Skylake, we really should be thanking Intel for getting their 10nm foundry sorted.
     
    Trimlock, GoldenTiger and ZeroBarrier like this.
  22. GreenOrbs

    GreenOrbs Limp Gawd

    Messages:
    372
    Joined:
    Jun 7, 2017
    ERwVRGH.jpg

    This is the graph that still bothers me... if the IPC is going up by 18%.... why does the projected single thread performance vs broadwell only show an ~4.5% increase when going from whiskey lake to ice lake.

    Edit: note that the baseline of the graph isn't zero... its 1x. By cutting of 2/3 of the range and only showing between 1.0-1.5x the improvements look larger. Everyone does this though (NVIDIA/AMD) so not unusual.
     
    Last edited: May 28, 2019
    MarkVI likes this.
  23. NKD

    NKD [H]ardness Supreme

    Messages:
    7,739
    Joined:
    Aug 26, 2007
    So intel managed to make sense for themselves about IPC lol. They are like politicians now. Just say 18% and let people figure it out hahaha. So not much compared to last gen, Also intel tends to think that increas in clock speed = more instructions per clock as well. Haha.
     
    Manny Calavera likes this.
  24. NKD

    NKD [H]ardness Supreme

    Messages:
    7,739
    Joined:
    Aug 26, 2007
    AMD has been spot on about their IPC claims. Zen was spot on. I have no doubt they pulled 15% with Zen 2.
     
    Darth Kyrie and N4CR like this.
  25. Nobu

    Nobu 2[H]4U

    Messages:
    3,279
    Joined:
    Jun 7, 2007
    clock speed is my guess. immature process and new arch makes it harder to hit the high numbers, and could be worse if temperature is an issue.
     
    GoodBoy likes this.
  26. Joust

    Joust 2[H]4U

    Messages:
    2,991
    Joined:
    Nov 30, 2017
    Maybe just one more, for old times' sake?
     
    Manny Calavera and Susquehannock like this.
  27. ChadD

    ChadD [H]ardness Supreme

    Messages:
    4,207
    Joined:
    Feb 8, 2016
    That's nice... so that makes them exactly even to where they where....

    https://www.phoronix.com/scan.php?page=article&item=mds-zombieload-mit&num=1

    16% just happens to be the EXACT impact the current generation of Intel chips looses when mitigated. So assuming these new chips have proper hardware fixes and see a 3% hit as AMD does... then I guess Intel is going to break even. Good on them. :)

    On a super cynical note I wonder if Intels IPC increase is them choosing to compare vs mitigated tests. Hey look we gained 16% after dropping 16% in a year due to their own security cheating hardware. lmao Intel
     
  28. Legendary Gamer

    Legendary Gamer Gawd

    Messages:
    679
    Joined:
    Jan 14, 2012
    That's the problem, Intel never has bench marked any of their CPU's with the patches in place. Typically the release all benches pre patch. If they actually accounted for the patch slow downs, this would be a grim picture of lost IPC. Generationally, we are likely seeing less than 1-2% gains and only in some benchmarks, with losses elsewhere. Intel's AVX512 performance will likely be stellar if their i3 10nm part (on select Lenovo laptops as of 1-2 years ago) is any indication. That's about it.

    Only time will tell, at this point IDK what I will be doing for my next build. However, given what I've seen about PCI-E 4.0 I may just wait to see what AMD does with TR. Building a nuclear M.2 Array on 16 Lanes of PCI-E 4.0 sounds like fun!
     
    Darth Kyrie and ChadD like this.
  29. Legendary Gamer

    Legendary Gamer Gawd

    Messages:
    679
    Joined:
    Jan 14, 2012
    "Intel systems all saw about 16% lower performance out-of-the-box now with these default mitigations and obviously even lower if disabling Hyper Threading for maximum security...."

    IIRC No HT results in another 5-15% loss in MT perf
     
  30. ChadD

    ChadD [H]ardness Supreme

    Messages:
    4,207
    Joined:
    Feb 8, 2016
    Very true... and perhaps its hoping for to much to think Intel has fixed their SMT implementation.
     
    Legendary Gamer likes this.
  31. Lakados

    Lakados [H]ard|Gawd

    Messages:
    1,720
    Joined:
    Feb 3, 2014
    Unless they have introduced a radically new architecture the best they can do is a mitigation of the latest attack.
     
    Darth Kyrie and Legendary Gamer like this.
  32. Elf_Boy

    Elf_Boy 2[H]4U

    Messages:
    2,352
    Joined:
    Nov 16, 2007
    Wonder how the new silicon does for tdp?
     
  33. pillagenburn

    pillagenburn Gawd

    Messages:
    1,014
    Joined:
    Oct 3, 2006

    Intel: WE HAZ ATE TEEN PARSENT MUH FR
    One must consider what security corners they had to cut in order to get this supposed 18%
     
    jfreund and Darth Kyrie like this.
  34. Jim Kim

    Jim Kim 2[H]4U

    Messages:
    3,547
    Joined:
    May 24, 2012
    Duh, that would be all of them.
     
  35. polonyc2

    polonyc2 [H]ard as it Gets

    Messages:
    16,930
    Joined:
    Oct 25, 2004
    will Hyperthreading be included on all (or most) of the chips this time?
     
  36. Dan_D

    Dan_D [H]ard as it Gets

    Messages:
    54,824
    Joined:
    Feb 9, 2002
    Intel's usually stated that performance increases have been roughly 7-11% per generation or thereabouts. However, in reality its been anywhere between 0% and 3% with 5% or more occurring in specific workloads. Of course since Sandy Bridge, we've often seen a 100-200MHz reduction in clock speed overhead.
     
    N4CR and GoodBoy like this.
  37. defaultluser

    defaultluser [H]ardForum Junkie

    Messages:
    12,634
    Joined:
    Jan 14, 2006
    this++!

    At least with AMD, we won't have to wait until the end of the year for retail release of mobile quad cores. You'll be able to test the 12 cores in July.

    And I expect AMD's Raven Ridge Zen 2 replacement will be available by the time Ice Lake is available in-quantity, so I expect quite the show in early 2020!
     
    Last edited: May 28, 2019
    IdiotInCharge likes this.
  38. GDI Lord

    GDI Lord Limp Gawd

    Messages:
    192
    Joined:
    Jan 14, 2017
    Thank you Ikasu, I missed the fine print.
     
  39. Zarathustra[H]

    Zarathustra[H] Official Forum Curmudgeon

    Messages:
    28,650
    Joined:
    Oct 29, 2000
    I hope there are no knock-outs from either AMD or Intel in this skirmish, so that they can keep leapfrogging each other and we get the same kind of advancement and pricing we saw in the Athlon vs. PIII / P4 era.

    I've been saying this all along though. AMD has an advantage right now due to process size. This won't last forever. They need to REALLY get their act together and improve the underlying architecture so they can remain competitive at the very least when Intel's 7nm process hits in 2021, if not sooner. It is unclear if 10nm will ever really provide anything of value, despite it being on their roadmap.
     
  40. Zarathustra[H]

    Zarathustra[H] Official Forum Curmudgeon

    Messages:
    28,650
    Joined:
    Oct 29, 2000
    Especially now. It's been the age old tactic when a market leader has a challenger. Create enough disinformation that buyers delay and wait for your next generation instead of buying from the competition.

    Happens all the time. Right now, its Intel, in years past its been Nvidia and their wooden GPU's.