Apparently Intel cherry picked the review samples for CoffeeLake.

Discussion in 'Intel Processors' started by Kato1144, Oct 11, 2017.

Thread Status:
Not open for further replies.
  1. Kato1144

    Kato1144 Limp Gawd

    Messages:
    263
    Joined:
    Oct 18, 2007
    I'm posting this here just so potential buyers of coffee lake know that they might not get close to the same performance with there new i7 or i5 CPU, unless your going for a unlocked CPU and then you probably will get close to the top performance, this is more for the ppl looking at getting a lower end part and trying to use multicore enhancement for a soft over clock.

    anyway AdoredTV explains it very well what is appears to be happening and other reviewers like jays2cents have collaborated his findings

     
    Last edited: Oct 11, 2017
    Nightfire and Master_shake_ like this.
  2. TaintedSquirrel

    TaintedSquirrel [H]ardForum Junkie

    Messages:
    8,586
    Joined:
    Aug 5, 2013
    Hasn't it always been this way? Aren't Ryzen consumer chips OCing worse than reviewer samples?
    AdoredTV is like someone who just built their first PC and decided to make a hardware YouTube channel.

    FULL DISCLOSURE I did not watch the video.
     
  3. Kato1144

    Kato1144 Limp Gawd

    Messages:
    263
    Joined:
    Oct 18, 2007
    ya but the OC difference with the Ryzen review samples to the average was 200-300MHz at worst, some of these chips like the lower end i5 have almost a 1GHz difference in clock potential from the review sample, BTW you would know if you watch the video :p
     
  4. TaintedSquirrel

    TaintedSquirrel [H]ardForum Junkie

    Messages:
    8,586
    Joined:
    Aug 5, 2013
    He does biased hit pieces so I'm not really interested in anything he has to say, since I can't trust him.
     
    Armenius likes this.
  5. German Muscle

    German Muscle [H]ardness Supreme

    Messages:
    5,383
    Joined:
    Aug 2, 2005
    Doesnt AMD cherry pick which CPU of Intels they compare it to so it makes them look better?
     
  6. Cali3350

    Cali3350 [H]ardness Supreme

    Messages:
    4,626
    Joined:
    Mar 22, 2002
    My 8700K better hit 5GHZ man... ive been waiting since the P4 days to see that reported on my system :(
     
  7. Kato1144

    Kato1144 Limp Gawd

    Messages:
    263
    Joined:
    Oct 18, 2007
    well I'm just bringing attention to this for those looking to buy coffeelake, this really only effects the low end non k skus, people looking to go with a 8400 might be in for a shock when there turbo boost only gets them to 3.2GHz on all cores but the reviewer is getting 4.0GHz on all cores on his 8400.

    It looks like intel made sure the best of the best was sent out and that is miss leading the customer if the averages are wildly different, reviews who got 8700k form there local store were getting much lower average results with multicore enhancement on then the reviews that were sent CPU's from intle, but then again the 8700k can be maunaly clocked to match so this is more of a issue with the non k CPU's

    Just to be clear this is not a attack on intel or coffeelake, this is just to keep you guys the potential customers informed of everything that is going on, if intel had coffeelake out a year ago i would be sporting a 8700k for the sweet IPC for sure but it was not to be and my R7 is more then acceptable for the next 2-3 years maybe longer
     
    Last edited: Oct 11, 2017
  8. Stryke1983

    Stryke1983 Limp Gawd

    Messages:
    181
    Joined:
    Dec 1, 2011
    Is there any actual evidence of this from reliable sources?
     
    Armenius likes this.
  9. Kato1144

    Kato1144 Limp Gawd

    Messages:
    263
    Joined:
    Oct 18, 2007
    here is 2 videos from jayz2cents about the multicore enhancement and how it is effecting the results as not every MOBD will have this option, although he is focused on the 8700k at the end of the day you can manually overclock that chip to match what you will get with multicore enhancement and he dose mention that but the real issue it going to be with the non k CPUs as the silicon lottery will determine how far you can boost and how many cores can be boosted at said speed, multi core enhancement will try and get the most out of your CPU but if it require more power and gets hotter then the boost specification it wont boost to it's max instead it will just settle down to were it is with in the spec as now i read that intel only guarantees the base clock and single core turbo and wont disclose the multicore turbo on the CPU.




    Disclaimer: This is just my take on this based on what i have read and watched, it will really be down to the ppl willing to take the chance with a non k CPU and a Z370 and see if they get a great value CPU or a meh CPU
     
    Nightfire likes this.
  10. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    47,979
    Joined:
    May 18, 1997
    Fwiw, we always test with all cores locked at our test frequencies... And I actually check those.



    Listen to folks with hands on and take the time and spend the resources to buy their own CPUs for testing to get the real story.
     
  11. SimianRob

    SimianRob n00bie

    Messages:
    52
    Joined:
    Mar 28, 2013
    Well here he did point out the weird results between reviews and a lot of that came down to Multi-core enhancement. Auto-overclocking all cores to single core boost levels and publishing those results as default clocks is obviously pretty misleading, but a lot of the reviewers themselves didn't even realize this was going on, or so they say. It's sort of funny considering there were posts on reddit from users wondering why their stock i7 8700K temps were so high, kind of wonder how many of them were using Asus boards with this setting enabled.

    Between MCE, the paper launch and Intel still using TIM on K cpu's, the more this launch really has left a bad taste in my mouth. I've got a pile of new hardware sitting there waiting on a back ordered CPU. I'm impressed with how well Coffee Lake can overclock, but I'm tempted to return the mobo and just go back to the drawing board on this new system. Feel kind of dirty thinking I've done exactly what Intel wanted in a way, rushing this release forward hoping to grab as many potential Ryzen customers as possible with little to no stock for what could be months.
     
    Last edited: Oct 11, 2017
    OrangeKhrush likes this.
  12. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    he hits home truths, the kind only fanboys are offended by, and I don't trust fanboys.
     
    Kyle_Bennett likes this.
  13. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    People will attack AdordTV because it touches points that are sadly at play again, this vexes many because of their irrational devotion to [Team], sadly this team has a history blacker than night of anti competitive and anti consumer practices, further have been able to circumvent judicial process by virtue of delaying until the party raising runs out of funds. I buy the product but care not one Iota about intel corp, nor do I trust their corporate practice in as far as I can throw it.
     
    Ranulfo likes this.
  14. TaintedSquirrel

    TaintedSquirrel [H]ardForum Junkie

    Messages:
    8,586
    Joined:
    Aug 5, 2013
    Let's put aside his off-the-rail Reddit rant where he spent an entire day personally insulting people who criticize his videos.
    Or the fact that his angry tirade caused AMD to cancel his sponsorship and pull products from him mid-shipment.
    Or the fact that he made a video bragging about living with and mooching off his girlfriend.
    Or the fact that ALL OF THIS led to him finally purging his Reddit account out of shame.

    He has a track record of making spectacularly terrible predictions. All of which are pro-AMD, unsurprisingly (except Vega recently).
    He has a track record of using misleading, or outright false, data to criticize Intel and Nvidia.
    He has a track record of continually defending AMD at every turn, giving them a pass for things he goes after Intel/Nvidia for.

    Some of us have been following this guy for his full 2 year run and he has very little credibility left, even if he does make some good points in his videos. It's hard to take someone seriously when you already know they're going to go after some companies and give another company a total pass. All I can think is, "What has AMD done that he isn't telling us about?" Because we already know that's the case -- it's inherent in every AdoredTV video. His channel exists to attack Intel and Nvidia, while also defending and promoting AMD. He is an attack dog on AMD's leash.

    The only appropriate way to watch his content anymore is to go into it knowing he's being selective about which information to spoon-feed you to push his pro-AMD agenda. Why? Because AMD fans are eager to consume pro-AMD content (clicks = ad revenue), and I've speculated before that I believe he owns a lot of AMD stock and is trying to manipulate public opinion.

    The short and simple explanation is: People don't want to watch YouTube videos defending Intel or Nvidia. And airing that kind of content hurts his viewership going forward. There are reasons why he is famously known as an "AMD shill". Nobody has ever called him an Nvidia or Intel shill, which should indicate a clear problem to you. If he were at least attempting to be fair, you would see criticisms from both sides.

    I'll wrap up by saying a channel like AdoredTV's, but actually neutral, would be an amazing contribution to tech on YouTube. We just need someone who cares more about impartiality rather than making money or trying to prove a point.
     
    Last edited: Oct 12, 2017
    Trimlock, Armenius, Nebulous and 3 others like this.
  15. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    irrespective of the past, his content on this occasion was on point, since this is the issue in debate I will forgo who fanboys for whom, anandtech are getting continually hit because they are not the pro intel anandtech of yesteryear, whose findings were correct or how steve burke got a hammering for his belief that MCE should be off when testing and Steve has a history of being pro intel.

    I don't care that people hate it because the content pokes their irrational love of a corporate, any corporate loving is just plain stupid no matter the colour.

    My conclusion is that these shenanigans were easily avoidable and not even necessary so the question is why? what else is being hidden.
     
  16. TaintedSquirrel

    TaintedSquirrel [H]ardForum Junkie

    Messages:
    8,586
    Joined:
    Aug 5, 2013
    My point is that it's impossible to know unless you do your own independent research. And also you assume all of his speculation is true, even though most of it is probably wrong.
    I'm not gonna watch the video but I will direct you to the comments here.
     
    Armenius and -Sweeper_ like this.
  17. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    47,979
    Joined:
    May 18, 1997
    Stupid is as stupid does.
     
    p3sty likes this.
  18. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    I thought it was funny how he was the first one in here claiming he did not watch the video.

    Damage control much?

    He documented the cinebench diferences nicely. 1230 to 1580 is a wide margin. Variations in turbo 2.0 make the difference.

    Take the 1230 of the OEM score and divide by 3.7 (base clock). You get 332.
    Now take 1580 of the best score and divide by 4.7 (highest possible turbo all core). You get a very close 336.

    So you can pretty much divide CB by 334 to get the avg freq.
    Eg 1300 guru3d divided by 334 = 3.9 ghz.

    Moreover, half the reviewers got very close to 1425. Divide that by 334 and you get 4.3 ghz. Yep, the all core without MCE.

    Cinebench is wonderfully consistant. We see that most 7700k got around 990. Divide by 223 (which is 2/3rds of 334) and you get 4.4 ghz or the max all core turbo for the 7700k!

    Informative that XMP automatically turns on MCE.
     
    Kato1144 likes this.
  19. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    The above cinebench factor of 334 can show if an overclock is holding its frequency.
    Guru3d and techspot had scores of 1677 and 1660 with 5.2 ghz. Those scores line up closer to what 5.0 ghz should be. 5.2 ghz should be 1700 easily. Trusted Reviews managed a score of 1684 which matches up very close with 5.1 ghz.

    I guess the point in all of this is if any reviewer had a cinebench score higher than around 1425, you can throw that review in the garbage since it really was overclocked.

    Weather Intels fault or the MB manufacture is a topic for another thread.
     
    Ranulfo and Kato1144 like this.
  20. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    Hardware canucks wasnt too smart thinking their chip was running 4.1 ghz with a 1579 cb score.

    I give them credit for having the best REAL o/c I have seen yet. They managed a 1708 cb score which lines up near perfectly with the 5.145 ghz they were reading.
     
    Kato1144 likes this.
  21. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    Jayz o/c score at 5.0 ghz of 1643 was only 5% higher than his original "stock" score of 1559. Kudos to him for catching all of this.

    Asus really looks to be the biggest dirt bags here.

    Were they hoping some unwitting reviewer would do a Z-370 mb shootout and show how much awesomer their mb was than the competition?
     
    Kato1144 likes this.
  22. Stryke1983

    Stryke1983 Limp Gawd

    Messages:
    181
    Joined:
    Dec 1, 2011
    No, they attack Adored due to his track record of unsubstantiated click bait.
     
    Armenius likes this.
  23. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    In this case he was correct, and very substantiated, so do we paint it with the same brush or do we address the issue? He also got the history of Intel's dealings correct, sure he does at times go over the top but at this point i would rather take it on a case to case basis than blanket coat it.
     
  24. Stryke1983

    Stryke1983 Limp Gawd

    Messages:
    181
    Joined:
    Dec 1, 2011
    I'll just stick with more reliable sources from the get go rather than giving clickbaiters the benefit of the doubt again and again.
     
    Trimlock, Armenius and -Sweeper_ like this.
  25. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    Ok then ignore him and listen to Jay. He comes to the same conclusions.

    BTW, did Shintai get banned or something. Haven't heard from him in a while, LOL.
     
    Aluminum, kirbyrj and N4CR like this.
  26. Vader1975

    Vader1975 Limp Gawd

    Messages:
    391
    Joined:
    May 11, 2016
    This is exactly why Kyle doesn't like "review" samples. He doesn't know what they will send him. Some cherry picked stuff or the real deal. So that is why he buys his own samples for us from RETAIL. So he is playing the loto just like us. You tell em Kyle!
     
    Pocatello, Aluminum and Kato1144 like this.
  27. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    Kyle Bennett stated from the start of the review that it was a review for [H] users and for that it would be an overclocked review, so it was in plain sight, which is why there is nothing to be said on the issue.

    The issue is some reviewers that benched the results at 4.7ghz and called it 4.3ghz stock, this is a major misrepresentation of fact, it perhaps is incompetence, I would hat to think they knew about it as that would and should be the end of all licences to review, willfully doing something vs incompetence or negligence are two different scales altogether.

    To their credit JayzTwoCents and Steve Burke found this problem and re-ran their benches
     
  28. cageymaru

    cageymaru [H]ard|News

    Messages:
    16,484
    Joined:
    Apr 10, 2003
    I can't stand AdoredTV because he fanboys too hard. I don't mind people that like Nvidia, Intel, AMD, Krispy Kreme, Ford, Home Depot, whatever. I like reading stuff from people that really like their purchase and are super eager to tell you why they love it. But when you fanboy as hard as AdoredTV does, it makes you instantly discredit everything the person says as hype. Well actually it makes me put his whole Youtube channel on block.
     
  29. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    47,979
    Joined:
    May 18, 1997
    It is not a "problem" it is the way it has worked for literally years. I worked with ASUS to get that feature put in the UEFI literally years ago. If reviewers are too ignorant to know how it works by now, that is on them. And let's just get to the meat of the discussion. How the fuck do you call yourself a "hardware reviewer" if you are not even verifying clocks BEFORE you start to run any of you canned benchmarks??
     
  30. Iching

    Iching [H]ard|Gawd

    Messages:
    1,791
    Joined:
    Nov 29, 2008
    The fact of the matter is Intel is having problems keeping up with AMD. It doesn't take a doctor to review hardware. No wonder reviews are skewed and shit.
     
  31. Nightfire

    Nightfire Limp Gawd

    Messages:
    331
    Joined:
    Sep 7, 2017
    This seems a bit harsh. It will take the new guys some time to meet the standards of the Old Guard.
    In his original review, he does mention that the 8700k was boosting up to 4.7 ghz.

    The issue with ASUS is that mce is turned on by default, at least when XMP is activated amd does not give the user the warning unless you deactivate in the bios and then reactivate.

    I think Jayz was more looking out for the consumer and not trying to raise his status as one of the reviewer elites. You might me a worthless dirtbag turd of a reviewer if you dont understand the workings of MCE and how Asus implements it, but I would argue that most users do not understand it and the stress it may put on other components.
     
  32. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    47,979
    Joined:
    May 18, 1997
    Really has nothing to do with MCE. The fact is that if you are taking benchmarks that compare performance, you should know very well what the clocks are on the CPU. How do you have any idea that your equipment is at spec if you do not check these things? I do not think it is harsh at all.

    Reviewer: Today we are comparing CPU A to CPU B. CPU A is faster than CPU B by 15%.

    Reader: What was the clock speed on the CPU?

    Reviewer: It said XGHz on the box.

    Reader: Thanks for letting me know what to spend my money on.

    Persons influencing buying decisions without doing their proper due diligence is irresponsible at best. That is what is harsh, not my opinion on it. Sure, mistakes are made, but that is not a mistake, that is simply a thoughtless process.
     
  33. Sancus

    Sancus Limp Gawd

    Messages:
    484
    Joined:
    Jun 1, 2013
    I'm just a random end user who occasionally overclocks for fun and even I know about multicore enhancement, and that it's been defaulted to on on Asus motherboards for years. When I first heard about this 'scandal' I was very confused. I would expect anyone posting reviews of hardware to know how that hardware actually works...

    I'd consider understanding multicore enhancement to be common knowledge for anyone who has ever actually read an Asus motherboard manual.
     
    Armenius, Kyle_Bennett and Brackle like this.
  34. ghostwich

    ghostwich [H]ard|Gawd

    Messages:
    1,422
    Joined:
    Sep 10, 2014
    And yet...
     
  35. rgMekanic

    rgMekanic [H]ard|News Staff Member

    Messages:
    3,185
    Joined:
    May 13, 2013
    Adored video was really good, and completely factual, dunno why you are knocking the vid

    I don't think it's Intel sending out binned CPUs to reviewers, but it's ASUS doing weird shit compared to others in the bios, and reviewers being lazy and not checking their shit.
     
  36. Brackle

    Brackle Old Timer

    Messages:
    6,962
    Joined:
    Jun 19, 2003
    Yea Why blame intel when it was a setting on the Asus motherboard? This is all on Asus, I mean TBH thats really bad because as jay2cents says they were basically overclocking the CPU to the point where an air cooler might not even be able to keep it cool. So some people think the CPU is bad since it is always hot?

    Makes me wonder why we see thread that say their 7700k is going to 100c at stock? Makes ya wonder if it was the MCE causing those issues.

    But like Kyle said, as a "good" hardware reviewer you check the clocks when running benchmarks...specially a new CPU...You want to verify stuff....crazy but makes ya wonder how many more lazy hardware reviews are out there.
     
    Nightfire, Kyle_Bennett and rgMekanic like this.
  37. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    Many here fanboy hard
    Makes you wonder why Intel publicly stated don't overclock your 7700K, Makes you wonder why in 2017 you need to delid your CPU just to stop it from failing to opperate at a optimum level. This is where I give AMD the benefit of the doubt, heat and power is not AMD's problem, it is the node and that is very fixable, for intel are the hitting the IPC brickwall and thermal/electrical conduction brickwall? Are we going to get LN2 pots with future chips, or do you need a 25K phase changer just to run your system.

    I was young, i messed around with PC's, overclocked, played around with LN2 and it no longer holds appeal to me, I just want to take a CPU out the box, plug in and enjoy rapping Battlefield, DayZ, NewZ servers and watching the tears in the chat log roll down. Also I am at that stage where 4C/8T seems boring, I want those 6/12, 8/16 parts, to actually feel like I am updating. The good thing about the CPU war is now I have a chance to buy a heavy cored chip for a fantastic price that before I needed an X platform to run, this is possible because AMD did well enough to force that stingy company to catch a wake up.
     
  38. TaintedSquirrel

    TaintedSquirrel [H]ardForum Junkie

    Messages:
    8,586
    Joined:
    Aug 5, 2013
    What's this page for then?

    https://www.intel.com/content/www/us/en/gaming/overclocking-intel-processors.html
    And here's exactly what Intel said (in response to 7700K owners reporting temperature spikes)
    Here's what AMD says in their official overclocking guide.
    Intel CPUs can reach 5 GHz even on paste while Ryzen struggles to reach 4 GHz with solder. There's also the 10% IPC discrepancy AMD needs to overcome. The fact of the matter is, Intel can get away with using paste because their competitor offers an inferior product despite having a superior thermal design. Even Intel's power-hungry 18C chip outclocks the entire Ryzen/TR line-up.

    Moreover, Intel's thermal issues can be solved with water or de-lidding. There's nothing consumers can do to fix Ryzen's clock deficiency. So when Ryzen finally utilizes its solder to break 5 GHz, then Intel will have to start worrying about their paste.

    I hope that puts things into perspective. AMD and their fans are putting the cart before the horse when it comes to the TIM debate.

    To me the issue seems to be just a huge distraction since the only other things to criticize Intel for are "bad business practices" and "too expensive", neither of which make for a very effective debate. Toothpaste memes, though? Solid material right there.
     
    Last edited: Oct 13, 2017
    Speedeu4ia and IdiotInCharge like this.
  39. OrangeKhrush

    OrangeKhrush Gawd

    Messages:
    880
    Joined:
    Dec 15, 2016
    As above Ryzen's limitation is the node not allowing high frequency but at lower frequency it is producing high output, a local reviewer with ran Intel Kabylake vs Ryzen in low clocks domain the results are stageringly different.

    IPC yes AMD has less IPC architecturally but it is not like Intel are moving IPC forward and are unlikely to ever again do it on the core I architecture, they will increase IMC performance and clock speed gains to offset this but we are looking at the pinnacle of x86 limits more or less, AMD caught up IPC wise a lot, and there is plenty for AMD to still push forward between Pinnacle Ridge and Mattise.

    What AMD have done right is SMT, it takes a offset of 700-800mhz to overcome AMD's core count, since clockspeed > cores in affecting results this is impressive from a company that nailed their first ever iteration of SMT with flying colours. I am hoping AMD answer back strongly and we continue to throw the CPU war into a state of flux.
     
    N4CR likes this.
  40. IdiotInCharge

    IdiotInCharge Not the Idiot YOU are Looking for

    Messages:
    4,476
    Joined:
    Jun 13, 2003
    You like repeating this prediction- let's see your source, or at least your explanation, if you have one.
     
Thread Status:
Not open for further replies.