Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Intel AMD

Intel Was Rather Misleading in Its Comparisons Between the Core i9-12900K and Ryzen 9 5950X (notebookcheck.net) 103

Posted by msmash from the closer-look dept.
NotebookCheck: It seems that Intel was not telling the full story when it compared the Core i9-12900K against the Ryzen 9 5950X, currently AMD's most comparable processor. As it turns out, Intel allowed the Core i9-12900K to consume 2.4x the power of its AMD competitor and benchmarked the Ryzen 9 5950X using an older version of Windows 11 with AMD performance issues.
This discussion has been archived. No new comments can be posted.

Intel Was Rather Misleading in Its Comparisons Between the Core i9-12900K and Ryzen 9 5950X More

Intel Was Rather Misleading in Its Comparisons Between the Core i9-12900K and Ryzen 9 5950X

Comments Filter:

  • I am shocked! Shocked, I say! (Score:5, Funny)

    by suutar ( 1860506 ) on Wednesday November 03, 2021 @11:01AM (#61954297)

    Well, maybe not that shocked.

    • Re:I am shocked! Shocked, I say! (Score:4, Insightful)

      by garyisabusyguy ( 732330 ) on Wednesday November 03, 2021 @11:10AM (#61954327)

      In this tribal society that we live in, Intel lovers will quote the study without reservation, intel haters will fume and spit and... the rest of us will go, "A corporation twisting reality to their needs, meh, what did you expect..."

      Truly a sad state of affairs, what would it take for honesty to become a cherish ideal again?

      • I know people stuck with Intel because of some very specific use cases (usually around motherboards), and during the FX days you'd be crazy to go with AMD unless you were buying a 6300 for the price (the 8000 series with it's multi core focus was way, way too early, and while the processors hold up well under modern workloads/games you'd still just buy a Ryzen, especially with how silly the resale market is). And Dear God, the 9000 series. They couldn't hang with the i7s so AMD just overclocked the hell out

        • Re:Are there any Intel lovers? (Score:5, Interesting)

          by garyisabusyguy ( 732330 ) on Wednesday November 03, 2021 @11:35AM (#61954441)

          FWIW, I liked AMD because we were able to migrate from Alpha platform to multi-core, multichip AMD servers with a relatively low amount of bullsnot

          This was enabled by AMD decision to incorporate hypertransport long before Intel got their inter-processor connection scheme figured out.

          Intel's decision to shove Itanium down every organizations throat who used Oracle made it a lot easier to go with AMD as well.

          • The intel is a tiny bit faster a 4 times the power, but the really big deal is that the apple chip comes with GPU too. And that GPU runs in unified memory so it's going to be faster than an Nvidia plus intel for any operations requiring combinined CPU/GPU processing or large memory GPU. And still using much less power.

          • I think AMD was a touch more competitive in that market even during the FX days. I don't think they got anywhere in the data center is because their power consumption was too high and that meant too much heat, but if you were only racking up four or five servers the cost savings were well worth it. And of course AMD was ahead of the game on multi cord it's just on the desktop we didn't see multi-core seem much use until recently. I don't think I had a game that really needed multi-core until Far Cry 4.

            • This was an admittedly small private company data center, say 30' x 50' with raised floor, halon system, an aging PDU and cooling that was originally specced for dual mainframes (for failover). The mainframes were gone before I was there. By then it was a bunch of DEC Alpha boxes (big as fridges) tons of DEC storage and a few dozen standard racks full of every generation of Compaq gear that had been sold between 1996 and 2005.

              They tried blade servers full of Intel chips, but had a low user count per blade (

        • I got an FX-8350 for literally half the price of a competitive Intel processor and it still does everything I want it to. Games were already typically aggressively multi threaded in that era thanks to the consoles using 8 cores and everything else CPU heavy was also multi threaded by then. It has been one of.my all time best PCs.

          • Re: (Score:2)

            by gTsiros ( 205624 )

            You can probably buy a 200 USD secondhand laptop faster than that CPU.

            Not worth it. Seriously.

            • You can probably buy a 200 USD secondhand laptop faster than that CPU.

              That's literally more than I paid for this CPU when it was new.

              Not worth it. Seriously.

              Neither is reading your comments, but they still let you post.

        • Why? I still have several FX 8320 systems in use (fileservers), and they have been / are perfectly fine. Saved a ton of money compared to the equivalent i7 (2600/3770) While ST performance is inferior, MT is about on par, and the motherboards (ASUS M5A99FX Pro R2s) have all the features I need (multiple PCI-E 16x slots, high number of SATA ports, etc.)
          • I mostly thinking about desktops. For servers if it's a single server, if you're using a desktop chip in the server I'm guessing that it is, it's not that bad a deal. You probably already had software that was multicore aware back then because server software often is. On the desktop though effective use of multiple cores is a pretty recent thing outside of a few applications. And unfortunately a lot of those applications were better optimized for Intel... I've seen video of games compared on in 8350 and th

            • I got the original AMD FX G1 Gaming motherboard (the 1.0) for my FX-8350. The onboard Killer ethernet has been a big let down, it never worked reliably even in 'doze, and was unsupported under Linux as well last I looked. I slapped a 2-port pcie gige card in there and never looked back. It works so well I don't even remember which it is, but it was cheap and the drivers are mainline. It was not at all the most expensive motherboard, but it wasn't the cheapest either obviously. The onboard power button and p

        • Gamers stuck with Intel for the past 20 years because until just a few years ago AMD couldn't compete on performance due to dropping the ball with Bulldozer, Piledriver, Steamroller, Excavator, etc. i.e. my i7-4770K lasted a good 7 years and is still OK. Intel was faster but you PAID for that advantage. Ryzen and more specifically Zen 3 has been a HUGE game changer with its 20% IPC uplift. AMD is faster then Intel which we haven't seen in decades. Now Intel is the budget option. LOL.

          Consoles have been

      • what would it take for honesty to become a cherish ideal again?

        Ironically, the idea that it ever was is itself a lie.

        • Hmmm, cynical much?

          Honesty or truthfulness is a facet of moral character that connotes positive and virtuous attributes such as integrity, truthfulness, straightforwardness, including straightforwardness of conduct, along with the absence of lying, cheating, theft, etc. Honesty also involves being trustworthy, loyal, fair, and sincere.

          Honesty is valued in many ethnic and religious cultures. "Honesty is the best policy" is a proverb of Edwin Sandys, while the quote "Honesty is the first chapter in the book o

          • I don't disagree that honesty is and has been cherished by many. But, I am skeptical that the situation was much better in historical times, and think that might be a nostalgic fiction.

            Of course, I may just be cynical. :)

      • Vi wins, of course.

    • Re:I am shocked! Shocked, I say! (Score:5, Funny)

      by hey! ( 33014 ) on Wednesday November 03, 2021 @11:13AM (#61954337) Homepage Journal

      This could call into account the very legitimacy of vendor-conducted benchmarks.

      • Re:I am shocked! Shocked, I say! (Score:4, Interesting)

        by Mark of the North ( 19760 ) on Wednesday November 03, 2021 @12:03PM (#61954525)

        My favourite university instructor shared a decent piece of advice about vendor-conducted benchmarks: You can compare the performance of the vendor's products, and you can compare the performance of the competitions products, but you can't compare between the vendor and their competition's products.

        Now, I can usually wait for third-party benchmarks before making any decisions. Back then, third-party benchmarks didn't really exist. Yes, it was that long ago.

      • The only time I semi-trust vendor benchmarks is when they're comparing multiple of their own products against each other, and even then mostly if they're the same generation.

        So, if WDC releases 6 different capacity point SSDs in the same family, at the same time, and says "the 480, 960, and 1920 can hit 500MB/s write speed, while the 3.84T, 7.68T can hit 580MB/s, and 15.36T can hit 620MB/s" I'll mostly believe them, at least to the extent of "the 15.36T is around 24% higher write throughput than the 480GB."

    • When reading an article and you go to yourself, man those other guys must be morons, or look at our side we are so much better. Then chances are you are being fed Marketing, Advertisements, Propaganda, and/or misinformation. Also if you read information that is counter to your preference and showing that it has issues it isn't necessarily a shill just trying to degrade what you find joy in.

      For most of us, we often just stand behind what we are use to, the politics, religion, sports team, Operating System,

    • I for one find it inconceivable that companies with questionable morals and billions on the line would try to skew results.

  • Intel Was Rather Misleading in Its Comparisons

    No shit!

    • Re: (Score:2)

      by Pascoea ( 968200 )
      Right? Is there ANYONE on this site that is surprised?

      • Hopefully not anymore, but just a decade ago you'd have all jumped down my throat for calling Intel liars, and they're still liars, and they'd always been liars before that too. All that's changed is more people are paying attention to the same things as me now.

    • Screw that, stop with the centrist words. They deliberately misrepresented the product and failed to disclose test conditions at the time of presenting the benchmarks.

      Intel LIED!

  • So for the M1 comparisons... (Score:5, Interesting)

    by SuperKendall ( 25149 ) on Wednesday November 03, 2021 @11:09AM (#61954325)

    Recently Intel was also doing a comparison that showed itself favorably closing the gap with the M1... since that was not even a benchmark from a working system yet I wonder how much truth there is to that one as well. Will be really interesting to see real chips start shipping end being measured.

    • Re: (Score:2, Interesting)

      by ctilsie242 ( 4841247 )

      Even though the M1/M1 Pro/M1 Max chips are laptop CPUs, they do give the latest gen i9s a run for their money... at a fraction of the energy used and heat needing to be dissipated. To boot, they even do GPU stuff on die.

      I do know that the i9s need two radiators to deal with the 250 watts (or 850 BTUs) of heat coming from them, and that isn't including the heat that needs to be dealt with from the GPUs. All the while, Apple is dealing with the cooling with just two skinny radial fans for everything, and a

      • Re:So for the M1 comparisons... (Score:4, Insightful)

        by thegarbz ( 1787294 ) on Wednesday November 03, 2021 @12:23PM (#61954585)

        Correct it's not a laptop CPU, but don't take the hyperbole in the opposite direction.

        I do know that the i9s need two radiators to deal with the 250 watts (or 850 BTUs)

        They need no such thing. The rated power output of the i9s you quote is for less than 10 milliseconds. The 250w specification has literally nothing to do with your heatsink. The 125W TDP on the other hand does need a sizable heatsink, but you don't need special radiators for that either.

        Also no one has ever in the history of computing used BTUs when talking about heatsinks or thermal capacity.

    • Re:So for the M1 comparisons... (Score:4, Interesting)

      by UnknowingFool ( 672806 ) on Wednesday November 03, 2021 @11:29AM (#61954417)
      Many people have called out the Intel M1 comparisons as highly flawed including using a beta version of Adobe Premiere Pro in the M1 (as the M1 was new) while using the commercial version for Intel. Later, Adobe released beta versions of Premiere Pro that put the M1 as 80% faster in many tests. I do not think Intel highlighted that they were using beta versions of Premiere Pro.

      • Re: (Score:2, Insightful)

        by dfghjk ( 711126 )

        These people are are Apple Stans like SuperKendall. They also claimed that the comparison was not fair because the Intel processor was new and the M1's were not. Curious contradiction.

        Also, Apple, and the Stans, have claimed that the M1 runs x86 software faster than Intel because of Apple's superior "not an emulator" Rosetta technology, so these "beta M1 support" releases of software are yet another excuse. There is absolutely nothing objective about these comparisons, whether you are talking about Intel

        • Comparing M1 (with its ton of fixed-function hardware) and x86 is misleading Bit of a throwback, but it's like comparing a system with software 3D rendering + CPU-driven sound to a system with a Voodoo and sound card No sh*t that the 2nd system will be faster and more power-efficient!

        • They also claimed that the comparison was not fair because the Intel processor was new and the M1's were not. Curious contradiction.

          Citation needed. One of the complaints is that Intel switched not only CPUs but the laptops between tests [allaboutcircuits.com] but not thoroughly disclose that. Their benchmarks made is seem that their processor had better performance and it was just as efficient. The truth was that one processor had some better performance but not as efficient. Another complaint was that Intel used an x86 optimized benchmark instead of something more cross platform. Lastly Intel actually showed that a 0% for the Mac in a comparison for softwar

        • They also claimed that the comparison was not fair because the Intel processor was new and the M1's were not.

          I said nothing like that, why do you lie all the time I wonder? What is it that compels you to speak only in lies?

          I think you need help man. Like seriously. You are hurting yourself only, I normally don't even read whatever lies you write in response to my posts, but judging by this one you really have to fix whatever is going on in your head before you are permanently lost.

    • Recently Intel was also doing a comparison that showed itself favorably closing the gap with the M1... since that was not even a benchmark from a working system yet I wonder how much truth there is to that one as well.

      Didn't you hear? Benchmarks aren't important, only "benefits and impact of technology" is important https://www.notebookcheck.net/... [notebookcheck.net]

      You see benchmarks only matter if they show Intel does well.

    • Re: (Score:2)

      by dfghjk ( 711126 )

      Gotta make it about Apple, right SuperKendall? As though Apple doesn't do the exact same things.

      And bullshit, that particular benchmark, along with every possible benchmark, was of a "working system", otherwise benchmarking would not be possible. This is just how stupid you are, SuperKendall.

      "...yet I wonder how much truth there is to that one as well."

      Something you would NEVER say regarding an Apple claim. Rosetta IS NOT AN EMULATOR!!!!

    • M1max vs alder lake mobile was already reported here, it's allegedly the same story as tiger lake vs m1: intel has the performance advantage but at higher power. https://hardware.slashdot.org/... [slashdot.org]

      I haven't seen any review entity break embargo yet aside from some small scale youtube channels, will likely have to wait for tomorrow for anything better.

      • I haven't seen any review entity break embargo yet aside from some small scale youtube channels, will likely have to wait for tomorrow for anything better.

        Thanks, it's nice to know something more concrete is close at hand, I wasn't sure of the Intel release schedule.

  • No way!!!

    They are saints and we must blindly follow them and ignore all the illegal crap they have done in the past, plus all the lies that are always posted on these announcements.

    Seriously, if we had balls like customers did back in the day, we would had stopped giving this shitty company any attention.

    Same for nvidia.

    But no, we have these idiots now that are empty internally that must worship corporations and brands to fill their empty existence.

  • Between a freshly released and not yet available processor with something that's been out for over a year and is readily available?

    Pat Gelsinger needs to get the engineers back on track if they have any hope of catching AMD at this point.

  • And the lesson is, (Score:5, Insightful)

    by jenningsthecat ( 1525947 ) on Wednesday November 03, 2021 @11:20AM (#61954375)

    believe nothing a processor manufacturer says about its products, performance. Wait for the independent benchmarks and power consumption tests before choosing a processor. There are often significant advantages to not being an early adopter. Hell, there are often significant advantages to being a full generation behind the latest hotness - especially when, as in this case, "hotness" can be taken literally.

    • Re:And the lesson is, (Score:4, Insightful)

      by FallOutBoyTonto ( 6835322 ) on Wednesday November 03, 2021 @11:29AM (#61954423)
      The latest processors NDA's prevents independent benchmarks until after the processor is released. With only cherry picked benchmarks being released, they'll most likely be sold out by the time we get some actual numbers

      • Re:And the lesson is, (Score:5, Funny)

        by UnknowingFool ( 672806 ) on Wednesday November 03, 2021 @11:53AM (#61954501)
        Maybe Intel can use targeted ads to make sure these sell in Alaska. "Why buy a computer and a space heater? Just buy a 12900K and get both!"
        • 12900K? False alarm! It just seems hot because it's in Kelvin instead of Celsius. I wish people paid more attention to units...

      • The latest processors NDA's prevents independent benchmarks until after the processor is released.

        Right, and that is an obvious sign that the vendor's benchmarks are shit.

        With only cherry picked benchmarks being released, they'll most likely be sold out by the time we get some actual numbers

        So IOW I have no reason to buy one until I can't buy one? Yep, that sounds like an Intel processor.

    • Wait for the independent benchmarks

      But benchmarks are no longer relevant. Intel's CEO said so! https://www.notebookcheck.net/... [notebookcheck.net]

    • Re: (Score:2)

      by dfghjk ( 711126 )

      "...and power consumption tests"

      We know who's side you are on. ;)

      "There are often significant advantages to not being an early adopter."

      Like being able to run your software!

      "...especially when, as in this case, "hotness" can be taken literally."

      Affects almost no one, but is a major talking point of one particular manufacturer. Why pretend you. are giving advice on objectivity?

    • > believe nothing a processor manufacturer says about its products, performance. Wait for the independent benchmarks and power consumption

      Indeed.

      * In Intel's benchmark the i9-1200K was using 241W compared to the R9 5950X being limited to 105W. Also the i9-12900K was using the faster DDR5 RAM compared to the Ryzen using DDR4.

      * The Ars Technica comparison [arstechnica.com] showed the Intel's i9-12900K was using 341W compared to R9 5950X using 204W.

      Pricewise the i9-19200K has a MSRP of $669 compared to the R9 5950X having a

  • Misleading and not forthcoming (Score:4, Interesting)

    by UnknowingFool ( 672806 ) on Wednesday November 03, 2021 @11:20AM (#61954377)
    Using a problematic version of Windows 11 is misleading in their comparisons; however, Intel simply did not disclose or highlight that the performance gains of the 12900K were at the expense of much more power consumption. Many reviewers have noted those performance gains were not as great as they seem due to the power consumption. This means that the 12900K has little performance gains if overclocked while the Ryzen 5950X could get some performance if it was overclocked.

    • Re: (Score:2, Flamebait)

      by Rosco P. Coltrane ( 209368 )

      a problematic version of Windows 11

      Has there ever been a version of Windows that wasn't problematic?

    • Using a problematic version of Windows 11 is misleading in their comparisons;

      According to the article, the better-for-AMD version of Windows 11 came out in "mid-October" - in other words about 2 weeks ago, and the benchmark was released "last week." That's an incredibly slim accusation - that Intel released benchmarks that don't use a new windows version less than a week old.

      • The problem is not that Intel used it. The problem is that it was well known that up until the Oct 21 patch, Windows 11 had issues with Ryzen processors. You would think that Intel would rerun tests with the patch. But they did not.

  • Wintel monopoly strikes again? (Score:5, Interesting)

    by JoeyRox ( 2711699 ) on Wednesday November 03, 2021 @11:31AM (#61954429)
    Interesting how Microsoft released a Windows 11 build with an obvious performance bug specific to AMD processors at around the same time a struggling Intel needed to run comparison benchmarks against a surging AMD.
    • perhaps AMD wanted to give Intel a boost by providing the buggy drivers? Or are you suggesting a more nefarious scenario where Microsoft infiltrated the AMD developer team and inserted the bug in their code?
      • The problem was not just with AMD's chipset drivers, that only affected preferred core selection, it was also Microsoft's changes to Windows 11 itself which affected cache performance (and was the much bigger performance hit)
        • So AMD are allowed to make mistakes for their own implementation, but it must be a conspiracy if a vendor makes a mistake implementing code to support their hardware?

      • perhaps AMD wanted to give Intel a boost by providing the buggy drivers? Or are you suggesting a more nefarious scenario where Microsoft infiltrated the AMD developer team and inserted the bug in their code?

        And Perhaps you should've read past the headlines about the issue:

        AMD's "preferred core" technology also had issues with Windows 11, preventing it from optimizing hardware for performance. After installing Windows 11 the hardware interface to the OS (UEFI) "may not preferentially schedule threads o

  • Prepare for disappointment (Score:4, Funny)

    by Gabest ( 852807 ) on Wednesday November 03, 2021 @11:31AM (#61954431)

    I can already see kids complaining that their CPU won't perform anywhere near as the pre-release benchmarks, because they have regular cooling, not liquid nitrogen.

    • Of course, we all know liquid nitrogen is so 2020s. All the cool kids use liquid helium these days. Thanks, I‘ll be here all week.
  • doesn't run tests and post benches that make them look the best.... please e-mail me, because it would be a first. All this is old news, wait until the real benches come out. My money says- Intel will have a small advantage, at the cost of using more power, and it looks like they will be under cutting AMD prices.... for now. I couldn't care less who is marginally faster and has a bigger dick, what I like seeing is we are finally getting some real competition and innovation happening. Sounds like an overall

  • Say it ain't so! (Score:3)

    by MachineShedFred ( 621896 ) on Wednesday November 03, 2021 @11:41AM (#61954465) Journal

    Were we not all basically saying this the day that benchmark surfaced? I'm sure we'll also find that they were using cryogenic cooling with it too - something all of us have lying around the house, obviously.

  • For those that remember, "Surprise! Surpise! Surprise! [youtube.com]"
  • This is all just pointless conjecture and nonsense. It will all be revealed tomorrow, so why even bother with this article?
  • Round up the usual suspects!
  • Intel knows the bad results will be accepted long enough to make bank.
  • This has been known for at least 5 days now. Intel guidance on how to configure an Alder Lake chip: PL1 at 125W, PL2 at 241W...yep, typical operation will suck up a LOT of power. Intel said on their Q3 conference call that it would not be at process/design parity with AMD+TSMC until 2024 and Intel claims it will be superior in 2025...assuming that AMD won't be doing stuff that makes Intel designs seem old and conservative.
  • ... adviseren WC-eend. Translate at will.

  • Always just wait for reviews from independent testers.

    It's not that hard.

  • Anyone who buys Intel is getting ripped off, I guess they are to foolish to admit they have been fooled.

  • Not wanting to defend Intel, but they DID acknowledge at the time they were using win 11 with the known AMD performance issue and they stated they didn't have access to the patched version at the time as that was prior to its release. So that part was not misleading.

    • It's still misleading, in that they knew it was a flawed benchmark but published it anyway. All they had to do was wait, and they knew that, but they went ahead when they did specifically to produce a benchmark that made them look better than they deserve.

  • When is Intel going to give up these unethical business practices? They don't even care anymore that people know. It's mind boggling why any major company would do business with them.

  • Does anyone remember when Intel ran print ads touting how much faster the 8086 was than the Motorola 68000?

    • Re: (Score:1)

      by ebvwfbw ( 864834 )

      I remember lots of false claims like that one. Also, the 88000, the DEC Alpha and so on. Not to be outdone I remember Apple's Steve Jobs claimed MacOS was the first 64 bit OS when it was in fact the last 64 bit OS.

      There are Lies, Damn lies, and Computer statistics.

  • Lying with statistics and test results for fun and profit!

Slashdot Top Deals

To the landlord belongs the doorknobs.

Close