@@s1acr457 Except they at least didn’t blame reviewers. That’s why Zen5 launch got so much flak, more about the relationship with reviewers than the actual product.
Same thing happened for AMD's 9000 launch, drivers, bios, Windows updates were not launched until after release, didn't help AMD any either, but some of the fixes have been released.
Prior to the release of Ryzen, Intel either failed, or refused to innovative & were happy to charge customers for essentially what was the same processor for a number of generations. If they had made an effort to innovative, rather than count their profits, they wouldn't be in this position.
well if you build a company based on stopping competition by bribing then thats what you good at , intel innovated little and spent more time bribing OEMs to not use competition products , this is karma biting intel back , every bully gets humbled at sometime . once ARM gets foothold , intel is finished .
@@Reyfox1 That got overturned because it was only deemed illegal _afterwards_ saying Intel was too big to give discounts to it's biggest customers. You know, what every f*cking company on the planet does for their customers?
the story is a little more complicated then this, Paul Otellini the intel CEO got caught flat footed by the ipad/tablet revolution around 2010ish... he was half out the door planning retirement and working with the incoming CEO Brian Krzanich to figure out intel's direction in this new world. they had just released Sandy Bridge, and AMD had just released Bulldozer and intel had vanquished AMD from the computer desktop market. so what was the decision these two men came to? lets compete with ARM with our x86 chips. Atom was the result, and it didn't go over well, due to the atom mistep (intel had to give them away for free) Krzanich (now in charge) came to another conclusion and pivoted. the conclusion was making the desktop into mobile. and they would do this by reducing their desktop chips in energy consumption and maintaining the performance, since they really had zero competition from AMD note, these types of decisions influence 5 years of production direction. the first move in this direction was Broadwell, which was sorta a flop, but they fixed the problems with the memory control and out came skylake. sure the performance wasn't a giant step up from haswell, but the power efficiency gains were impressive. intel was counting on their vast advantage in process nodes over the rest of the world to shrink size and increase efficiency bringing the desktop down in size and increasing it's performance until there was no point in a tablet anymore, and maybe someone would hit on a way to get a windows phone to work. but intel had a problem, their 10nm node just didn't work. and intel made a tragic mistake. rather then go back to the drawing board and invent the process shrink to 10nm, someone made a budgetary decision and decided since the process wasn't making a viable chip with "1 pass" of the wafer through the machine they needed to double the passes through the machine to make a viable chip. and this worked in theory. the problem was the process was too imprecise, even a micron misalignment meant the whole wafer was ruined. and so while the fought with the 10nm disaster the started the 14+++++ odyssey. worse, right about the time they were about to pull the plug on their designs for 10nm, a 7nm production facility was opened, and they found the problems they had identified with 10nm were far worse in 7nm and they'd need 4 passes of the wafer through to make a viable chip. so intel decided to save their 7nm process they needed to master the 10nm process. About this time Ryzen launched. and while zen 1 was about 15% behind intel's 8th gen, that was competitive enough to put amd back into the game. and frankly, intel now had a new problem, unable to shrink their node, while arm was improving by performance by leaps and bounds, making it less and less likely the goal of arm competition, their vanquished foe was back offering more cores, at good performance and suddenly intel was caught in the middle of a knife fight bare handed and with their pants down. intel with no other solutions at hand was forced to aggressively develop 14nm to it's theoretical limits while poring untold millions in desperation to fix 10nm, all to compete with the foe they thought they had finished off. gone were the dreams of besting arm, now it was all they could do to keep up with AMD. so 14nm rolled on, amd grew more competitive with zen2, finally intel got 10nm to work after a fashion (there still is huge yield waste, making it expensive) and we get up to 11th gen and zen 3... it's not that intel wasn't trying its that they made some horrendous choices and pissed away all their advantage.
Clear, concise, pulling no punches. Leo just delivering time after time. Hopefully Intel can deliver. And I say that as someone not even on the platform or in the market for it.
Intel management made a decision years ago to sit back and become a cash cow using their existing products instead of investing money into engineering and the future. They're paying the price for it now. Fortunately, the current CEO and team seem to understand this and are working to fix it. It may take a while but I wish them well. The market needs strong competition to drive things forward and keep prices reasonable.
Problem is a lot of management are still in the company. New CEO doesn't mean the board of directors aren't the same people that ran the company into the ground for the past 15 years
What is this fixation on Intel being the “THE competitor”. They f up they gone. There will be other competitors comes along as long as there is a demand for it and government does not get its hand in it. It always does. Same go for the automakers.
I kinda agree on the last sentence -- market needs strong competition... But everything else... IMHO Intel is rotten from deep inside... and I don't see the competition from their side... I see it rather from ARM or RISC side of things where AMD would need to adjust and accommodate. IMHO this is the real competition for the next years till the Neural-CPUs come into the game.
Why the hell were they surprised by all of this. Why were they surprised by these outcomes. Couldn't they just build a PC with the same products available to consumer a few days before the launch, see that it's not working, and hit pause on the launch. They do more damage by releasing something bad than by waiting and releasing something good.
A few days before launch is just not viable. Theres tons of logistics to get boards and chips into people hands that's been happening for months prior. They should have been able to test most of this stuff 6 months ago, so they absolutely had the opportunity to hit pause then, but chose/hoped not to.
@@Tugela60 You're right, Intel is not responsable for MSI, Asus, or what ever mobo company shitting the bed. They are however responsable for their overall platform and the impression it makes. They need to do something.
@walter274 Well, AMD had the same problem. The reality is that the board/OS manufacturers are different companies and do their own thing. Obviously the processor manufacturers try, but all they can do really is supply specs and engineering samples to the board manufacturers, and hope they don't screw up. Intel and AMD don't get to review what their partners make. The issue with the new Intel chips is that they use very different technology than prior chips, so there is going to be a learning curve.
People have ran 1P core and all E-cores and had better performance than various configurations of P and E-cores in cyberpunk. It’s just straight up broken that much is just true. So whether or not you believe Hallock there is performance to be gained back in just having it run correctly, and at such a point with basically tweaks it will at least have a market rather being an absolute disaster.
@@rurutuM Correct. The instability issue is at large whilst the oxidation was a situation with one batch at one location. The problem is that due to the potential liability for all of the issues at large, Intel wouldn't be direct and just say how to identify and isolate which ones were potentially effected by oxidation. GamersNexus put heavy attention on the narrow issue of oxidation, so a large amount of people now conflate "oxidation" with the instability.
Sheesh! I went from Xeon E3-1240 v5 (kind of like Skylake i7-6700) to Ryzen 7 2700X and thought it was awesome. i7-4790 to 9800X3D is crazy. I actually have 5800X3D currently, but at some point, I'll built a new one with 9800X3D or maybe even 9950X3D, depending on whether it has VCache on both CCDs or not.
It will depend on what you are doing with your hardware. Only playing games then there will some improvements, but that also depends on your video card and what resolution you have on your display. For task where you can max out on processing power (multicore) like video editing or heavy calculation then you will see a large jump as most benchmark will show.
When I was at business school we had a case study of how IBM f*d up massively and went into hardcore recovery mode, selling off excess (including art pieces - though something tells me the bigwigs managed to squirrel away a few of those), concentrating on their core business etc. Looks like Intel should do something like that.
They have already started doing stuff like that, but there's probably more that they will need to do. I think it's a matter of time before their fab business becomes separated from their chip design business, but it might take a while before that happens.
Amazing how they don't care about client side. The x86 duopoly was incapable to align with Windows monopoly for a good CPU release. Kinda unbelievable... Sounds more like Windows not giving a single f and rush bios as usual.
@saricubra2867 Alder Lake are the last Intel CPUs I sell in builds, but don't kid yourself, there's issues on 24H2. The big.LITTLE design still hasn't received support like monolithic.
Him (Robert Hallock) saying they're "being honest about it", wonder where he was with 13th and 14th gen debacle, and intel kept avoiding to point out what the real problem was
Robert Hallock is just a hypocrite now, just ignore anything he says. Telling us how great 3D V-cache was when he was at AMD, and now I'm sure he will talk negative about it man just GTFO 😂
@@ms3862 on most companies the decisions are not all made according to technical specification... other reasons can verry well play into it... :) I must say I am more then extremely unhappy with our Intel servers at the moment... but that is in our use case..
Given what happened with microcode, *even if* performance was better, hitting “some snags” which have some “pretty wild unexpected effects” is personally, a complete deal breaker. They have a QA problem. That’s going stay as their problem, not mine.
AMD has had a few of these as well. There is only one solution: Don't buy anything that hasn't been out for 6 months or more. On top of a better experience you usually also get a bit of a discount.
Totally get that - no one wants to play tech support for a new CPU 😂 Feels like these ‘wild effects’ weren’t exactly the performance boost anyone was hoping for
@@VesiustheBoneCruncher Yup. But that's the thing: You can only test for so much in the lab. It is not until you mass deployment that you get "a realistic test case". And this is not just a thing with computers. Nearly every product (that isn't a pet rock) goes though design iterations based on end user feedback. But in this case, yes, Intel bumrushed the launch. Having a Windows install BSOD because there is a discrete graphics card in the computer IS something that needs to be tested in the lab.
Considering all Zen 3, 4 and 5 users recently got a noticeable performance improvement on Windows 11 via updates the statement isn't too far fetched. But we will know in a few months....
The "fix" will be buy the refresh/update of this generation. If this was a 3rd tier chip manufacturer, I would be more inclined to believe code updates will fix the problems. This is Intel, they know better. I will believe it when I see it.
Best way to fix it, is to lower prices dramatically. The Core Ultra9 285K performs a little better/worse than a Ryzen 7 5700x3D, so price it a little higher than the 5700x3D....the 285k at $350-$400, and it would be better to look at.
Drastically lower prices and commitment to the platform for at least another generation is a start. Maybe subsidizing the expensive and exclusive CU DIMMS like AMD had DDR5 thru Microcenter. It's about time Intel behaves like they're at the bottom and start providing value to end customers. They can't live off their govt and corporate prebuilts contracts forever.
They can't because of how expensive it was to develop and get manufactured. Intel can't eat the cost of lower profit margins because they're not producing the CPUs, TSMC are and TSMC 3nm is incredibly expensive.
they use the most expensive node on the market and manufacture not in house... selling that low would drain Intel rather quickly from it's last few cash reserves
@justacomment1657 Every problem you listed is of Intel's own making. How did they expect it to turn out? Basically a 15900k, a regression - slower and costlier than 13700k. If AMD cuts last gen 7800x3d to sub $350 and if 9950x3d beats Ultra9 at productivity while embarrassing it in gaming, what's left for Intel?
It is what happens when you are haemorrhaging money, stock is falling and investors are getting twitchy - you rush a product to market, trying to beat the competition. A humbling lesson indeed.
Intel didn't spend any time R&Ding on TSMC. They slapped a mess together and shipped it. AMD has 5+ years experience with TSMC including working directly with them during development. Looks like AMD foresaw REALLY, REALLY well.
@@andersjjensen Lol no... chipsets are not quite remotely the same thing as CPU.. Not even close. It's like saying Realtek audio/ethernet on an Intel board means Intel has experience outside of their node...
@@lePurpleDragon I work with discrete logic design and firmware for a living. Whether you want to make a DSP, a PCIe interface or a CPU the approach is the same. Intel uses the same tools for their own nodes as they do for TSMC (there are only a couple of vendors in each category). The PDK is basically "a plugin" that makes the programs behave the same, but outputs the masks that are correct for the specific node. And your example is stupid. Intel engineers designing a chipset that needs to communicate quite closely with the CPU is nothing like a generic vendor who only has to interface with PCIe.
BTW, Hallock did the exact same thing at AMD where he would be in chats and when confronted with a problem say, "I will look into that and get back to you in a couple of weeks or so", and he never did. So don't hold your breath, Leo.
i built a 7800x3d system in october, and got hold of one for £340, which now as the 2nd best chip for gaming after the newly released 9800xd, but now sold out and scalpers have put the price to £650-999, looks to be a sound investment until i buy the 9800x3d for 340 in a year or 2.
I purchased a 14900K in feb of this year (2024), ive had to RMA the CPU twice of the last 10 months of having this build ive had a working PC for maybe 6 months. The intel customer support has gotten worst this time around. I sent it out two weeks ago to get RMA and its been sitting there I called yesterday to get a status update and they told me they had not received the CPU, I told them over the phone its been delivered since Tuesday of last week. I honestly think im completely over Intel after they sold me a bad product, and then shipped me a new CPU with the same issue.
Good. Intel has been subsidizing board partners for a long while even though AM4/AM5 has been the better architecture. Leave those "EXTREME" motherboards to rot the shelves.
AMD isn't held to the same standards as Intel or NVIDIA. If they played some of the dumb games those two have played this year (Intel with hardware, NVIDIA with drivers), they'd be chastised to the grave.
Businesses don't care about retail customers any more. The money is in servicing other business. All that matters to intel is that the servers and data centers just keep buying and they will.
Excellent video, concise and level headed response to intel's Robert Hallock comments. I do wish other content creators in the tech space had such a level headed response to Intels troubles that KitGuru's Leo had. 👍
Intel could salvage this by becoming a budget gaming king. My Idea: Intel Pentium 185 - 1P-Core-16E-Cores - Sell it for $299.99 Intel Pentium 165 - 1P-Core-12E-Cores - Sell it for $229.99 Intel Pentium 145 - 1P-Core-8E-Cores - Sell it for $179.99 Just get those out. If someone had to choose between a Ryzen 5 9600x or Pentium 185....it would help get attention away from AMD.
Here is what I will never understand...why in the 9 hells did Intel launch it on a new socket when they had to know it was going to get curbstomped by AMD? If they would have released it for 1700 it would have made some sense as they could have at least gotten those customers who are on the platform and avoiding Raptor Lake as well as used stock to replace RMA'd Raptor Lake chips but now they have a chip that not only costs too much but requires a new more expensive platform and who is going to go through a full platform swap for an inferior product?
You decide at the outset of the CPU design whether you'll need a new pin out or not. Given Raptor Lake was a refresh of a refresh of a refresh, they just needed to do one anyway but it was likely necessary to do updated pinout for the new chiplet design anyway.
@@blue-pi2ktAgree, a few years ago the CU-DIMMs were probably pitched as one of the new platform features that would put them back in the lead. Instead it became a liability with the poor execution on the processor.
@@smmasongtCall me naive but I actually expect them to get a fix out prior to Christmas that will make it reasonably competitive for some applications. It ain't beating the X3D in gaming but I doubt it was ever intended to. Taking their suggestions at face value and it really is a multi-factor issue it almost certainly means that IF they fix it, a genuinely meaningful performance boost is coming because these issues ordinarily compound for most actions and become exponential drags in edge cases across the CPU's instruction pipeline reducing performance substantially in the process.
@@smmasongtAlso CU-DIMMs are super valuable upgrades IF the CPU is designed to leverage that faster RAM speed but with the chiplet design you can only really choose between a relatively slow ringbus or an off-chip memory controller before the latency created by the necessary switching between P cores eats up all the architectural performance uplift and at the point, no RAM can dig you out of that design decision.
@@blue-pi2kt Yeah, the CU-DIMM’s are a net add for the platform but the click bait gaming channels are doing a lot of damage framing them as an overpriced liability compared to AMD’s platform. I’m sure there are some workloads where they really help. I hope Intel makes a comeback but honestly this guy seems like another worthless Intel executive talking out his a**. The 10nm fab debacle was really a disaster for the CPU design teams and they seem to be totally reactionary at this point, always a day late and a dollar short.
This is a perfect example of why early adoption isn’t always worth it. Intel’s architecture shifts are interesting, but it’s clear they haven’t fully optimized Arrow Lake for real-world usage yet.
Totally agree-early adoption can be a gamble, especially when it feels like Intel is still ironing out the kinks. It’s exciting to see new architecture shifts, but maybe waiting for a few updates or even the next gen is the safer bet if stability is a priority. Hopefully, these changes pay off in the long run!
They were still reading off a macro sheet, just an updated one. Reading between the lines they said nothing of importance and were trying to lie about the fact that they have a lot more left in the tank.
I completely enjoy and relate to the quote at the end of the video. Perfect. It doesn't have to be this way, but it is and is not changing for the better.
looking forward to your motherboard reviews Leo - I do hope so much that you get ASROCK boards to look at as well, they are such good value on paper anyway
It’s interesting that Intel is pushing AI integration with Arrow Lake, but I’m curious how well these AI enhancements actually translate to real-world applications. Are they providing a tangible benefit to everyday users, or are they more of a marketing angle for now? AI capabilities can sound impressive, but if they come at the expense of raw CPU performance or efficiency, I’m not sure it’s a good trade-off. It almost feels like Intel is trying to force a feature that doesn’t yet have a clear, practical use case in typical consumer workloads. What’s the real value here, and does it justify the potential drop in core performance?
85,000 + Employees , and yet they missed All of these many issues ?! Mind you, ditching SMT/HT chopped out 30% of performance,and still it does ok. 9m 38s, Yes because AMD blamed the Reviewing Community, on thier blog post.
If only 1 or 2 reviewers had not so great reviews, then there would be no need for an apology. But every reviewer panned the launch. Intel knew about the performance before releasing it. All this talk about "we're going to look into it and see where we failed", well, they knew it was going to be a fail. There is no excuse. And if they think a microcode or M$ update is going to make things right, yeah... wishful thinking. What they've done is make AMD's "so-so" launch look really good.
No. They tried since 11th gen to use TSMC or a process that never received a new process. Intel was working with ASML to develop the next gen foundry equipment that ni one, including TSMC and Samsung, didn't buy. Now, many are eyeing their acquisition to have a state of the art foundry. US citizens should support our companies and us.
Never thought there would be a day that i owned an AMD chip, But here i am with a 7800X3D installed and super impressed by its performance. For me i run MSFS so the 3D vcache helps as People who use MSFS need very good single core performance
A refreshing change from Intel's usual "lets blame everyone else and throw them all under the bus", I am an AMD user through and through so these problems won't effect me as I will be staying with AMD until after Zen 6 has been released (late 2026 I believe) because I have too many AMD based systems (2 X 7950X & 1 X 7950X3D). But I will say that the X86 platform needs a strong Intel & AMD because ARM is marching forward at an alarming pace.
It doesn’t matter which side UserBenchmark takes. Their only job is to be neutral, transparent, and provide accurate information-and they’re failing at it.
@@Open6a-fx4qf funny thing is, they even removed 13th and 14th gen reviews, they are mostly saying "12600KF is best cpu ever, just buy that" nothing else matters as low resolution benchmarking is irrelevant etc... it's comedy gold
No, when you have a paper launch after your company is already under fire globally, you can't deny responsibility for anything. It's like with Arc, which is a flop with now undetectable market share, but Intel can't deny it or it makes them look really bad.
I'm interested in an Arrow Lake laptop. I don't really care about bleeding edge multi-core performance, battery life is far more important to me (so long as GPU and CPU performance is "good enough").
So if we're thinking 9900x or 265K what should we examine to sway one way or the other? They price the same with 870 and 890 mobos. Coming from Skylake and there are so many CPUs now...
I want to believe Robert that they can fix ARL and that they will communicate openly about the issues. But it's been barely half a year since the RPL catastrophe and Intel did everything but being honest about it. And there wasn't a simple fix either, just the promise of a fix, and another one a month later, and so on...
I can understand the BIOS snafu. Good way of keeping stuff locked down until it's ready for the public. Kind of weird to be shipping boards to retail with BIOS in that state though.
Hi Leo, if you do review the Aorus pro ICE that has wifi just not in the name, can you scrutinise the WIFI card compared to the Elites etc, seems to be a better module in the ICE that uses the full frequencies? I know not a lot of people are to bothered about the WIFI on boards, but to some its very important and the reason why you would even buy that board over the cheaper boards. Thanks
I like Robert, seen a few of his interviews. he does not seem to play into the whole Intel ethos of arrogance, but it is only one man, can not see him changing much
Scalpers are killing the Intel launch even more delaying people from building their pc’s allowing AMD more time to push out more chips in the meantime. This is the first time I can’t get all the parts on launch.
His Intonation and response tells me they all knew but someone higher-up pushed the button. It's clear to me this is a schedule issue, to many cores firing bits across when it should be load balanced properly. It explains why cyberpunk runs so much better with only one or two p cores enabled. Bag of money and some code to Microsoft will fix it, I'm sure 😊
What stupid logic. Intel's CPU's have been straightforward for 15 years. They change the name once, and you throw a fit and "switch to AMD". How many times has AMD changed their GPU lineups? :D Child. Child. Child.
The fact that intel stock started to drop on November 7 again after trending up for a few days. Almost feel sorry for them, but they did this to them selves.
last intel CPU I wanted to buy was the 9900k but at the time a 2700x and a AM4 b450 pro carbon motherboard were cheaper been on AM4 ever since now with 5800x on x570 still don't need to jump platform to AM5
They didn't expect the reviewer results? Do they not test this junk before they release it? The reviewers had these results within a couple of days. How long did Intel have to trial this before the release? How long was this crap in development? No wonder their shares prices have tanked. With respect to their 'honesty' and willingness to fix the issue, how long was it before they acknowledged the fault in gen 13 & 14, and how long did they spend blaming the motherboard manufacturers?
I have always been an Intel Fanboy since the days when Cyrix chips were around, I have just ordered an AMD system, because of the poor launch of the ultra. Once customers leave its very hard to get them back.
While the performance might be lacking (at the moment), with the lowered power consumption and (hopefully) no longer any of the 13/14 gen issues these new cpu most certainly ARE an improvement. Just not in ways that people might find compelling. If anything it bodes well for the next gen to know that they stopped sticking to their old designs.
big little will always have trouble on desktop schedulers and I would avoid it like the plaque in a server environment... this will never work as good as it does on a mobile. optimized system....
@@glenndoiron9317 none of those benchmarks are virtualization stuff. and if you seriously belive that intels thread director is working as intended you are delusional... 🙈 oh and BTW... those productivity benchmark you are talking about are all about tile based rendering.... this is not the only productivity application of a cpu...
It's amazing how not long ago amd was considered bad and Intel king, but to be honest it took a lot of time for both to reach this status quo, like it took Intel bad cpus time and time again to do it, which is strange. Brand name is a strong thing
So both AMD and Intel screwed up the release of their mainstream parts in the same way. Stuff worked one way in their automated test labs on engineering boards and another way when the reviewers got the parts. It sounds like both of them need to have a manual test team that tests new CPUs the way that a reviewer would PRIOR to launch. Build PC on standard OEM parts, fresh Windows install, install games, play in interesting area, report any issues or differences from automated testing. They both ended up getting a team to look into things after the bad reviews came out - seems pretty stupid to release early so you can get bad reviews and almost no sales
I like the honesty but i also am not convinced until they really do what they say. Them saying EVERYTHING can for sure be resolved… dunno… makes me feel not as great as they might think. Sometimes you f.u. and you cant fix. Maybe they can fix most of it but everything? Who knows that when they cant really tell why it even happened in the first Place? Its at least strange.
I know Intel had the top of the line chips for a long time, and that was before they went to P and E cores.. since then, they have ran into problems, may I suggest that while they work out some of the bugs in using both P and E cores, that they also create a couple of chips just using P cores.. and having a three tier lineup of chips, one with 16 P cores, one with 12 P cores, and one with 8 P cores.. no hyper threading.. 16 P cores 16 threads!! And add more Cache at all levels on Chip!! And advertise them as GAMING CHIPS!! And do so on the same new socket of 1851.. and also increase the number of PCIE Lanes on the chips, to 28 on the 8 core chip, 34 on the 12 core chip and 40 on the 16 core chip!! I have an IVY BRIDGE chip which has 40 lanes on it.. add more PCIE lanes, its a hard selling point!!
Intel got cocky in the early 2010's by dominating the Desktop/laptop and server. So they starter diversifying (SSD, memory, networking, 4G/5G, business, IoT, FPGA, autonomous driving and AI) and cut their funding to research in fabs. They also didn't think that they needed EUV for 10nm for cost reducing reasons which i believe led to the massive delay in 14nm.
I'll watch this space because I'm genuinely interested to see how it goes and I need to upgrade my CPU; it's been a lot of years. I keep swinging between AMD and Intel, but because I don't solely game and want my rig to be very good at productivity and very good at games (but not enthusiast level), my focus is on the best balanced chip. I don't give a toss about it being AMD or Intel; I really just want the best chip for the best price for my needs.
If you buy a Z890 motherboard today and immediately update the BIOS using flashback you will avoid a number of the problems faced by we reviewers. Performance will still be problematic until the next round of fixes. Leo
Even if these software bios/driver fixes are done, I will still hesitate to buy a new expensive motherboard for this new 1851 socket and some also expensive CU-DIMM memory, knowing that maybe the next Intel generation will change of socket architecture again...
I can not laugh hard enough to the Motherboard companies :P We can all see that there are FAR MORE Intel motherboards then AMD ones. Specialy the EXTREME boards from all venders you see with Intel and NOT at AMD. I think it has something to do with contracts between Intel and the motherboard companies. I can NOT believe that they are happy with Intel right now because nobody is buying those boards :P
Well you would think companies like MSI would have done some initial tests internally, no one seemed to know anything about the launch based on what Leo showed us.
The strange thing is that the speed and compute power does not match the gamaing performance. Makes me think intel focused at light weights gamer and more businesses. But I have a feeling something else else is happening which wasn’t intended. Looks like some is preventing a full access to the gpu or something is causing a limitation. Because coming from 13th gen I can say the ultra veel faster in normal working environments en rung VMware on it I could see the improvement. Something is limiting how the the cpu should feed the pci express or the gpu
that something is most likely the memory controller and the total stupid core layout where every other core complex has different latency problems... 😅 it's basically elevated zen2 problems but the competitive floor is way higher now..... running vmware on such a cpu is Alsow something I find highly unpredictable...
@ maybe your are right. I create and modify images for companies and decide to try out the new ultra 9 and was impressed by the performance because I am use to boating up 15 virtual machines and running atvthe same time and doing without having difficulty on the is cpu is very. My 13th gen fans would have kicked in like I am running cinebench 23. Anybody that doesn’t game that much will love it and not everything is about games 😁
I really wish Intel would publish a real time up to date list of known issues as opposed to going "there are issues". That would undo some of the lost good will from me with raptor lake
Not sure if an easy to point out list of issues is a good idea from a marketing perspective. It would be used by competitors as well. I'd definitely make it at least product specific. And, as a consumer, I don't care about issues on products that I do not own.
@@owlstead I agree it is not a good idea for marketing, my concern comes from intel doing a very bad job of communicating the issues around Raptor Lake and this doesnt feel that different, apart from the fact that very few people have (or are likely to have) these arrow lake parts
Looks like the available DDR5 only goes up to 8000MT with high latency when most boards will take up to 9000MT. So you are saying that the higher latency is not a concern? Is there an advantage of let's say 7200 over 8000? Also these sticks are only available in 32GB for a 48GB configuration. Why such an odd setup?
![Kodak Black - Versatile 4 [Official Video]](http://i.ytimg.com/vi/VcP_1Ua__vY/mqdefault.jpg)
I think it was a great launch. I ordered my next AMD build right after watching the reviews.
Yes absolutely
Most of us did !
I am waiting for the 285K to come in stock.
Waiting for 9950 n 5000's :-)
Great "at making us make up our minds" launch
"Disconnect" meaning they tested it, saw there was a possible problem, shipped it anyway.
They knew before shipping what was going to happen, but hoped that people would ignore the reviewers and buy anyway.
Intel's Core Ultra™. Putting the FU back in FUN
Sounds like AMD with the 9000 series initial issues.
@@s1acr457 Except they at least didn’t blame reviewers. That’s why Zen5 launch got so much flak, more about the relationship with reviewers than the actual product.
Same thing happened for AMD's 9000 launch, drivers, bios, Windows updates were not launched until after release, didn't help AMD any either, but some of the fixes have been released.
Prior to the release of Ryzen, Intel either failed, or refused to innovative & were happy to charge customers for essentially what was the same processor for a number of generations. If they had made an effort to innovative, rather than count their profits, they wouldn't be in this position.
There are those that will buy no matter what the reviewers have reported. I see it all the time.
well if you build a company based on stopping competition by bribing then thats what you good at , intel innovated little and spent more time bribing OEMs to not use competition products ,
this is karma biting intel back , every bully gets humbled at sometime .
once ARM gets foothold , intel is finished .
@@parm2-x7h the EU took them to court for that and Intel lost. We'd still be paying premium for a 4 core CPU at 14mm +++++++++ if it were up to Intel.
@@Reyfox1 That got overturned because it was only deemed illegal _afterwards_ saying Intel was too big to give discounts to it's biggest customers. You know, what every f*cking company on the planet does for their customers?
the story is a little more complicated then this, Paul Otellini the intel CEO got caught flat footed by the ipad/tablet revolution around 2010ish... he was half out the door planning retirement and working with the incoming CEO Brian Krzanich to figure out intel's direction in this new world. they had just released Sandy Bridge, and AMD had just released Bulldozer and intel had vanquished AMD from the computer desktop market. so what was the decision these two men came to? lets compete with ARM with our x86 chips.
Atom was the result, and it didn't go over well, due to the atom mistep (intel had to give them away for free) Krzanich (now in charge) came to another conclusion and pivoted. the conclusion was making the desktop into mobile. and they would do this by reducing their desktop chips in energy consumption and maintaining the performance, since they really had zero competition from AMD note, these types of decisions influence 5 years of production direction.
the first move in this direction was Broadwell, which was sorta a flop, but they fixed the problems with the memory control and out came skylake. sure the performance wasn't a giant step up from haswell, but the power efficiency gains were impressive. intel was counting on their vast advantage in process nodes over the rest of the world to shrink size and increase efficiency bringing the desktop down in size and increasing it's performance until there was no point in a tablet anymore, and maybe someone would hit on a way to get a windows phone to work.
but intel had a problem, their 10nm node just didn't work. and intel made a tragic mistake. rather then go back to the drawing board and invent the process shrink to 10nm, someone made a budgetary decision and decided since the process wasn't making a viable chip with "1 pass" of the wafer through the machine they needed to double the passes through the machine to make a viable chip. and this worked in theory. the problem was the process was too imprecise, even a micron misalignment meant the whole wafer was ruined. and so while the fought with the 10nm disaster the started the 14+++++ odyssey. worse, right about the time they were about to pull the plug on their designs for 10nm, a 7nm production facility was opened, and they found the problems they had identified with 10nm were far worse in 7nm and they'd need 4 passes of the wafer through to make a viable chip. so intel decided to save their 7nm process they needed to master the 10nm process.
About this time Ryzen launched. and while zen 1 was about 15% behind intel's 8th gen, that was competitive enough to put amd back into the game. and frankly, intel now had a new problem, unable to shrink their node, while arm was improving by performance by leaps and bounds, making it less and less likely the goal of arm competition, their vanquished foe was back offering more cores, at good performance and suddenly intel was caught in the middle of a knife fight bare handed and with their pants down.
intel with no other solutions at hand was forced to aggressively develop 14nm to it's theoretical limits while poring untold millions in desperation to fix 10nm, all to compete with the foe they thought they had finished off. gone were the dreams of besting arm, now it was all they could do to keep up with AMD. so 14nm rolled on, amd grew more competitive with zen2, finally intel got 10nm to work after a fashion (there still is huge yield waste, making it expensive) and we get up to 11th gen and zen 3...
it's not that intel wasn't trying its that they made some horrendous choices and pissed away all their advantage.
Clear, concise, pulling no punches. Leo just delivering time after time.
Hopefully Intel can deliver. And I say that as someone not even on the platform or in the market for it.
The launch was just what AMD needed at the perfect time.
Intel management made a decision years ago to sit back and become a cash cow using their existing products instead of investing money into engineering and the future. They're paying the price for it now. Fortunately, the current CEO and team seem to understand this and are working to fix it. It may take a while but I wish them well. The market needs strong competition to drive things forward and keep prices reasonable.
Problem is a lot of management are still in the company. New CEO doesn't mean the board of directors aren't the same people that ran the company into the ground for the past 15 years
What is this fixation on Intel being the “THE competitor”. They f up they gone. There will be other competitors comes along as long as there is a demand for it and government does not get its hand in it. It always does. Same go for the automakers.
Looooool
So how do you explain their R&D being higher than Nvidia and AMD combined .
I kinda agree on the last sentence -- market needs strong competition... But everything else... IMHO Intel is rotten from deep inside... and I don't see the competition from their side... I see it rather from ARM or RISC side of things where AMD would need to adjust and accommodate. IMHO this is the real competition for the next years till the Neural-CPUs come into the game.
The alarm bells have been ringing ever since CES when they were very touchy about Lunar Lake.
well, that we do remember 😉
Why the hell were they surprised by all of this. Why were they surprised by these outcomes. Couldn't they just build a PC with the same products available to consumer a few days before the launch, see that it's not working, and hit pause on the launch. They do more damage by releasing something bad than by waiting and releasing something good.
A few days before launch is just not viable. Theres tons of logistics to get boards and chips into people hands that's been happening for months prior. They should have been able to test most of this stuff 6 months ago, so they absolutely had the opportunity to hit pause then, but chose/hoped not to.
No. They don't make the boards, they don't have control over whatever issues the board manufacturers have.
@@Quasar909 So ealier.
@@Tugela60 You're right, Intel is not responsable for MSI, Asus, or what ever mobo company shitting the bed. They are however responsable for their overall platform and the impression it makes. They need to do something.
@walter274 Well, AMD had the same problem. The reality is that the board/OS manufacturers are different companies and do their own thing. Obviously the processor manufacturers try, but all they can do really is supply specs and engineering samples to the board manufacturers, and hope they don't screw up. Intel and AMD don't get to review what their partners make.
The issue with the new Intel chips is that they use very different technology than prior chips, so there is going to be a learning curve.
Same Robert Hallock that said you could OC your Zen2 with PBO to levels that never happened.
Glad I wasnt the only one thinking that.
People have ran 1P core and all E-cores and had better performance than various configurations of P and E-cores in cyberpunk. It’s just straight up broken that much is just true. So whether or not you believe Hallock there is performance to be gained back in just having it run correctly, and at such a point with basically tweaks it will at least have a market rather being an absolute disaster.
Thanks Leo, enjoyed the update and your take on it all!
That slamming into ads is jarring, but the topping out of the sound in the clips was painful when listening on mobile.
Now Intel just needs to issue a recall for oxidized 13/14th gen chips. *spoiler* they won't */spoiler*
I dont even think they could afford to replace them right now - you seen their latest financial reports?
That oxidation was the early 13th gen manufacturing problem. Separate issue from the 13/14th gen instability
yes that seems unlikely doesnt it
@@RainbowSunset-uq2hs Maybe once apple/qualcomm/amd buys them lol
@@rurutuM Correct. The instability issue is at large whilst the oxidation was a situation with one batch at one location. The problem is that due to the potential liability for all of the issues at large, Intel wouldn't be direct and just say how to identify and isolate which ones were potentially effected by oxidation. GamersNexus put heavy attention on the narrow issue of oxidation, so a large amount of people now conflate "oxidation" with the instability.
I'm upgrading from an i7 4790 to a 9800x3d. It must feel like a continental jump.
Massive, massive - from ancient ssd to nvme
Yep. The power consumption and multi thread is outrageous. I upgraded from 6700hq to 8840HS
Sheesh! I went from Xeon E3-1240 v5 (kind of like Skylake i7-6700) to Ryzen 7 2700X and thought it was awesome. i7-4790 to 9800X3D is crazy.
I actually have 5800X3D currently, but at some point, I'll built a new one with 9800X3D or maybe even 9950X3D, depending on whether it has VCache on both CCDs or not.
@catsspat that's crazy uplift almost like turbo boost.
It will depend on what you are doing with your hardware. Only playing games then there will some improvements, but that also depends on your video card and what resolution you have on your display.
For task where you can max out on processing power (multicore) like video editing or heavy calculation then you will see a large jump as most benchmark will show.
When I was at business school we had a case study of how IBM f*d up massively and went into hardcore recovery mode, selling off excess (including art pieces - though something tells me the bigwigs managed to squirrel away a few of those), concentrating on their core business etc.
Looks like Intel should do something like that.
They have already started doing stuff like that, but there's probably more that they will need to do. I think it's a matter of time before their fab business becomes separated from their chip design business, but it might take a while before that happens.
"Only UserBenchMark can save us now"
- Intel
And even they think Intel’s heading for bankruptcy while recommending 12,13 and 14 gen.
Ha
😂😂😂
Amazing how they don't care about client side. The x86 duopoly was incapable to align with Windows monopoly for a good CPU release. Kinda unbelievable... Sounds more like Windows not giving a single f and rush bios as usual.
Yeah you pretty much nailed it in a sentence man
Alder Lake is stable and no issues there
@@saricubra2867 Alder Lake is slower in games in Windows 24H2.
@Patrick73787 Only in fantasy land. It's a far simpler architecture than Arrow Lake.
@saricubra2867 Alder Lake are the last Intel CPUs I sell in builds, but don't kid yourself, there's issues on 24H2. The big.LITTLE design still hasn't received support like monolithic.
It was indeed a fluster cluck.
Him (Robert Hallock) saying they're "being honest about it", wonder where he was with 13th and 14th gen debacle, and intel kept avoiding to point out what the real problem was
Robert Hallock is just a hypocrite now, just ignore anything he says. Telling us how great 3D V-cache was when he was at AMD, and now I'm sure he will talk negative about it man just GTFO 😂
Intel has done something no other company has achieved. Making CPUs slower than CPUs from 3 generations ago.
Our IT department was already switched over to all AMD computers before the launch.
Yup. I don't manage IT atm but I'd have all Intel gone.
Interesting to hear !
same at my team
1300 staff company here and we're still on Intel, now using lunar lake products
@@ms3862 on most companies the decisions are not all made according to technical specification... other reasons can verry well play into it... :)
I must say I am more then extremely unhappy with our Intel servers at the moment... but that is in our use case..
Given what happened with microcode, *even if* performance was better, hitting “some snags” which have some “pretty wild unexpected effects” is personally, a complete deal breaker.
They have a QA problem. That’s going stay as their problem, not mine.
AMD has had a few of these as well. There is only one solution: Don't buy anything that hasn't been out for 6 months or more. On top of a better experience you usually also get a bit of a discount.
Totally get that - no one wants to play tech support for a new CPU 😂 Feels like these ‘wild effects’ weren’t exactly the performance boost anyone was hoping for
@@andersjjensen Absolutely, the Phenom TLB bug also resulted in me feeling exactly the same about AMD at the time.
@@VesiustheBoneCruncher Yup. But that's the thing: You can only test for so much in the lab. It is not until you mass deployment that you get "a realistic test case". And this is not just a thing with computers. Nearly every product (that isn't a pet rock) goes though design iterations based on end user feedback. But in this case, yes, Intel bumrushed the launch. Having a Windows install BSOD because there is a discrete graphics card in the computer IS something that needs to be tested in the lab.
Funny. Issues with Arrow Lake are at the BIOS level and at the OS level, but not at the CPU level? Well, that's one way of looking at it...
Considering all Zen 3, 4 and 5 users recently got a noticeable performance improvement on Windows 11 via updates the statement isn't too far fetched. But we will know in a few months....
The "fix" will be buy the refresh/update of this generation. If this was a 3rd tier chip manufacturer, I would be more inclined to believe code updates will fix the problems. This is Intel, they know better. I will believe it when I see it.
Best way to fix it, is to lower prices dramatically. The Core Ultra9 285K performs a little better/worse than a Ryzen 7 5700x3D, so price it a little higher than the 5700x3D....the 285k at $350-$400, and it would be better to look at.
Drastically lower prices and commitment to the platform for at least another generation is a start. Maybe subsidizing the expensive and exclusive CU DIMMS like AMD had DDR5 thru Microcenter.
It's about time Intel behaves like they're at the bottom and start providing value to end customers. They can't live off their govt and corporate prebuilts contracts forever.
They can't because of how expensive it was to develop and get manufactured. Intel can't eat the cost of lower profit margins because they're not producing the CPUs, TSMC are and TSMC 3nm is incredibly expensive.
they use the most expensive node on the market and manufacture not in house... selling that low would drain Intel rather quickly from it's last few cash reserves
@justacomment1657 Every problem you listed is of Intel's own making. How did they expect it to turn out? Basically a 15900k, a regression - slower and costlier than 13700k.
If AMD cuts last gen 7800x3d to sub $350 and if 9950x3d beats Ultra9 at productivity while embarrassing it in gaming, what's left for Intel?
It is what happens when you are haemorrhaging money, stock is falling and investors are getting twitchy - you rush a product to market, trying to beat the competition. A humbling lesson indeed.
Thanks Leo, appreciate you pointing this out !
They didn't use enough glue
Intel didn't spend any time R&Ding on TSMC. They slapped a mess together and shipped it. AMD has 5+ years experience with TSMC including working directly with them during development. Looks like AMD foresaw REALLY, REALLY well.
@@lePurpleDragon TSMC has been producing Intel's auxiliary chips for over a decade. It's not like Intel isn't used to TSMCs PDK.
@@andersjjensen Lol no... chipsets are not quite remotely the same thing as CPU.. Not even close. It's like saying Realtek audio/ethernet on an Intel board means Intel has experience outside of their node...
They must’ve run out halfway through assembly day 😂😂
@@lePurpleDragon I work with discrete logic design and firmware for a living. Whether you want to make a DSP, a PCIe interface or a CPU the approach is the same. Intel uses the same tools for their own nodes as they do for TSMC (there are only a couple of vendors in each category). The PDK is basically "a plugin" that makes the programs behave the same, but outputs the masks that are correct for the specific node.
And your example is stupid. Intel engineers designing a chipset that needs to communicate quite closely with the CPU is nothing like a generic vendor who only has to interface with PCIe.
BTW, Hallock did the exact same thing at AMD where he would be in chats and when confronted with a problem say, "I will look into that and get back to you in a couple of weeks or so", and he never did.
So don't hold your breath, Leo.
i built a 7800x3d system in october, and got hold of one for £340, which now as the 2nd best chip for gaming after the newly released 9800xd, but now sold out and scalpers have put the price to £650-999, looks to be a sound investment until i buy the 9800x3d for 340 in a year or 2.
I purchased a 14900K in feb of this year (2024), ive had to RMA the CPU twice of the last 10 months of having this build ive had a working PC for maybe 6 months. The intel customer support has gotten worst this time around. I sent it out two weeks ago to get RMA and its been sitting there I called yesterday to get a status update and they told me they had not received the CPU, I told them over the phone its been delivered since Tuesday of last week.
I honestly think im completely over Intel after they sold me a bad product, and then shipped me a new CPU with the same issue.
That’s a pretty bad thing to hear 🤨
The launch was EXTREMELY informative, clear and concise! It clearly stated that Intels processors to be avoided at all costs.
Good. Intel has been subsidizing board partners for a long while even though AM4/AM5 has been the better architecture. Leave those "EXTREME" motherboards to rot the shelves.
9:30 Reminds me of the multiple times that AMD took the blame when it was Microsoft's fault and didn't complain about Windows issues.
AMD isn't held to the same standards as Intel or NVIDIA. If they played some of the dumb games those two have played this year (Intel with hardware, NVIDIA with drivers), they'd be chastised to the grave.
@lePurpleDragon Very true.
Businesses don't care about retail customers any more. The money is in servicing other business. All that matters to intel is that the servers and data centers just keep buying and they will.
Actually, server customers prefer AMD instead. Intel is losing server market share slowly yet surely…
😂sorry mate, but what exactly is the usp of a xeon server these days?
Excellent video, concise and level headed response to intel's Robert Hallock comments. I do wish other content creators in the tech space had such a level headed response to Intels troubles that KitGuru's Leo had. 👍
What about the oxidation issues? no mention of that?
Different topic
Intel could salvage this by becoming a budget gaming king.
My Idea:
Intel Pentium 185 - 1P-Core-16E-Cores - Sell it for $299.99
Intel Pentium 165 - 1P-Core-12E-Cores - Sell it for $229.99
Intel Pentium 145 - 1P-Core-8E-Cores - Sell it for $179.99
Just get those out. If someone had to choose between a Ryzen 5 9600x or Pentium 185....it would help get attention away from AMD.
For being a budget king you may need a budget mainbord as well, 210 euro seems to be around the minimum for a good build.
Here is what I will never understand...why in the 9 hells did Intel launch it on a new socket when they had to know it was going to get curbstomped by AMD? If they would have released it for 1700 it would have made some sense as they could have at least gotten those customers who are on the platform and avoiding Raptor Lake as well as used stock to replace RMA'd Raptor Lake chips but now they have a chip that not only costs too much but requires a new more expensive platform and who is going to go through a full platform swap for an inferior product?
You decide at the outset of the CPU design whether you'll need a new pin out or not. Given Raptor Lake was a refresh of a refresh of a refresh, they just needed to do one anyway but it was likely necessary to do updated pinout for the new chiplet design anyway.
@@blue-pi2ktAgree, a few years ago the CU-DIMMs were probably pitched as one of the new platform features that would put them back in the lead. Instead it became a liability with the poor execution on the processor.
@@smmasongtCall me naive but I actually expect them to get a fix out prior to Christmas that will make it reasonably competitive for some applications. It ain't beating the X3D in gaming but I doubt it was ever intended to.
Taking their suggestions at face value and it really is a multi-factor issue it almost certainly means that IF they fix it, a genuinely meaningful performance boost is coming because these issues ordinarily compound for most actions and become exponential drags in edge cases across the CPU's instruction pipeline reducing performance substantially in the process.
@@smmasongtAlso CU-DIMMs are super valuable upgrades IF the CPU is designed to leverage that faster RAM speed but with the chiplet design you can only really choose between a relatively slow ringbus or an off-chip memory controller before the latency created by the necessary switching between P cores eats up all the architectural performance uplift and at the point, no RAM can dig you out of that design decision.
@@blue-pi2kt Yeah, the CU-DIMM’s are a net add for the platform but the click bait gaming channels are doing a lot of damage framing them as an overpriced liability compared to AMD’s platform. I’m sure there are some workloads where they really help. I hope Intel makes a comeback but honestly this guy seems like another worthless Intel executive talking out his a**. The 10nm fab debacle was really a disaster for the CPU design teams and they seem to be totally reactionary at this point, always a day late and a dollar short.
Well that was interesting to see, just went to see their channel thanks Leo
Intel's focus 4 years ago was to design a CPU that is similar to Apple's M1 or ARM-like.
AMD's focus was Zen 5.
This is a perfect example of why early adoption isn’t always worth it. Intel’s architecture shifts are interesting, but it’s clear they haven’t fully optimized Arrow Lake for real-world usage yet.
Totally agree-early adoption can be a gamble, especially when it feels like Intel is still ironing out the kinks. It’s exciting to see new architecture shifts, but maybe waiting for a few updates or even the next gen is the safer bet if stability is a priority. Hopefully, these changes pay off in the long run!
Well at least someone from the company said something 'human' rather than reading off a macro sheet.
They were still reading off a macro sheet, just an updated one. Reading between the lines they said nothing of importance and were trying to lie about the fact that they have a lot more left in the tank.
I completely enjoy and relate to the quote at the end of the video. Perfect.
It doesn't have to be this way, but it is and is not changing for the better.
looking forward to your motherboard reviews Leo - I do hope so much that you get ASROCK boards to look at as well, they are such good value on paper anyway
yeah I hope for that too, ASROCK do not get much attention lately.
It’s interesting that Intel is pushing AI integration with Arrow Lake, but I’m curious how well these AI enhancements actually translate to real-world applications. Are they providing a tangible benefit to everyday users, or are they more of a marketing angle for now? AI capabilities can sound impressive, but if they come at the expense of raw CPU performance or efficiency, I’m not sure it’s a good trade-off. It almost feels like Intel is trying to force a feature that doesn’t yet have a clear, practical use case in typical consumer workloads. What’s the real value here, and does it justify the potential drop in core performance?
85,000 + Employees , and yet they missed All of these many issues ?! Mind you, ditching SMT/HT chopped out 30% of performance,and still it does ok. 9m 38s, Yes because AMD blamed the Reviewing Community, on thier blog post.
How can you release CPUs in that state and not know about it…
If only 1 or 2 reviewers had not so great reviews, then there would be no need for an apology. But every reviewer panned the launch. Intel knew about the performance before releasing it. All this talk about "we're going to look into it and see where we failed", well, they knew it was going to be a fail. There is no excuse. And if they think a microcode or M$ update is going to make things right, yeah... wishful thinking.
What they've done is make AMD's "so-so" launch look really good.
Love me a good Leo Says when skiving at work 😅
yup, I actually trust you Leo, interested in the outcome of this. Perhaps Pat Gelsinger came in late ... or too late? (my opinion is late)
No. They tried since 11th gen to use TSMC or a process that never received a new process. Intel was working with ASML to develop the next gen foundry equipment that ni one, including TSMC and Samsung, didn't buy. Now, many are eyeing their acquisition to have a state of the art foundry. US citizens should support our companies and us.
Never thought there would be a day that i owned an AMD chip, But here i am with a 7800X3D installed and super impressed by its performance. For me i run MSFS so the 3D vcache helps as People who use MSFS need very good single core performance
A refreshing change from Intel's usual "lets blame everyone else and throw them all under the bus", I am an AMD user through and through so these problems won't effect me as I will be staying with AMD until after Zen 6 has been released (late 2026 I believe) because I have too many AMD based systems (2 X 7950X & 1 X 7950X3D). But I will say that the X86 platform needs a strong Intel & AMD because ARM is marching forward at an alarming pace.
There’s Nothing Ultra about arrow lake. Should’ve just been core i9 15900k.
What's next, the 'SUPER ULTRA TURBO 385KS!'. Just calling it 'Ultra' doesn't magically make you product better Intel...
It's not that they screwed up. The problem is the released the CPUs. They would have to have been aware of the issues but they released them anyway.
Intel f-ed so much, that even User benchmark trashes them, amd style
I thought I would never see the day
One should never mention userbenchmark, they are an absolute disgrace.
It doesn’t matter which side UserBenchmark takes. Their only job is to be neutral, transparent, and provide accurate information-and they’re failing at it.
@@Open6a-fx4qf UserBenchmark is insanity and must not be mentioned.
@@Open6a-fx4qf funny thing is, they even removed 13th and 14th gen reviews, they are mostly saying "12600KF is best cpu ever, just buy that" nothing else matters as low resolution benchmarking is irrelevant etc... it's comedy gold
Wow Intel admitting anything? I wonder will we see him being fired by the end of the year now !
No, when you have a paper launch after your company is already under fire globally, you can't deny responsibility for anything. It's like with Arc, which is a flop with now undetectable market share, but Intel can't deny it or it makes them look really bad.
I'm interested in an Arrow Lake laptop. I don't really care about bleeding edge multi-core performance, battery life is far more important to me (so long as GPU and CPU performance is "good enough").
It looks like Intel is changing... that is only good news for the future. I like what Hallock was saying and hope he makes it happen.
So if we're thinking 9900x or 265K what should we examine to sway one way or the other? They price the same with 870 and 890 mobos. Coming from Skylake and there are so many CPUs now...
I want to believe Robert that they can fix ARL and that they will communicate openly about the issues.
But it's been barely half a year since the RPL catastrophe and Intel did everything but being honest about it. And there wasn't a simple fix either, just the promise of a fix, and another one a month later, and so on...
I can understand the BIOS snafu. Good way of keeping stuff locked down until it's ready for the public. Kind of weird to be shipping boards to retail with BIOS in that state though.
Ddi you not watch the video? Thise boards never shipped to end user...
@@philipjfry628 Sure. Shipping them to reviewers is not a good idea either of course.
This month..... cant say its going to be before nov 30... what day is it? Yikes.... somebody get this man a calandar!
Hi Leo, if you do review the Aorus pro ICE that has wifi just not in the name, can you scrutinise the WIFI card compared to the Elites etc, seems to be a better module in the ICE that uses the full frequencies? I know not a lot of people are to bothered about the WIFI on boards, but to some its very important and the reason why you would even buy that board over the cheaper boards. Thanks
I like Robert, seen a few of his interviews. he does not seem to play into the whole Intel ethos of arrogance, but it is only one man, can not see him changing much
Scalpers are killing the Intel launch even more delaying people from building their pc’s allowing AMD more time to push out more chips in the meantime. This is the first time I can’t get all the parts on launch.
Which cpu are you after ?
@@KitGuruTech I decided not to upgrade this year. I'm gonna wait and save my money.
Somehow feels very similar to Intel’s Arc product launches: initially weird benchmark results, ironed out after subsequent driver updates
and it still doesn't work on some titles...
His Intonation and response tells me they all knew but someone higher-up pushed the button. It's clear to me this is a schedule issue, to many cores firing bits across when it should be load balanced properly. It explains why cyberpunk runs so much better with only one or two p cores enabled. Bag of money and some code to Microsoft will fix it, I'm sure 😊
I'll go with them, I don't like easy setup, I like challenges so I'll ignore the AMD side and I'll buy a 285k. Thank's
Would be funny if they fix everything and pretty much nothing changes performance wise.
It's a stupid name. Should be 15900k etc. Intel love confusing it's customers. I'm gonna move to AMD
get a 9950x you won't be sorry, it's really good
After so many years on intel , I also will move to AMD.
Not to mention 285K isn’t even a “good” number/doesn’t even sound like a top sku
You will get confuse with AMD laptop cpu lineup!
What stupid logic.
Intel's CPU's have been straightforward for 15 years. They change the name once, and you throw a fit and "switch to AMD". How many times has AMD changed their GPU lineups? :D
Child. Child. Child.
The fact that intel stock started to drop on November 7 again after trending up for a few days. Almost feel sorry for them, but they did this to them selves.
last intel CPU I wanted to buy was the 9900k but at the time a 2700x and a AM4 b450 pro carbon motherboard were cheaper been on AM4 ever since now with 5800x on x570 still don't need to jump platform to AM5
They didn't expect the reviewer results? Do they not test this junk before they release it? The reviewers had these results within a couple of days. How long did Intel have to trial this before the release? How long was this crap in development? No wonder their shares prices have tanked. With respect to their 'honesty' and willingness to fix the issue, how long was it before they acknowledged the fault in gen 13 & 14, and how long did they spend blaming the motherboard manufacturers?
I have always been an Intel Fanboy since the days when Cyrix chips were around, I have just ordered an AMD system, because of the poor launch of the ultra. Once customers leave its very hard to get them back.
I moved to Intel this year, can not see me going back the way that company are operated.
While the performance might be lacking (at the moment), with the lowered power consumption and (hopefully) no longer any of the 13/14 gen issues these new cpu most certainly ARE an improvement. Just not in ways that people might find compelling. If anything it bodes well for the next gen to know that they stopped sticking to their old designs.
big little will always have trouble on desktop schedulers and I would avoid it like the plaque in a server environment...
this will never work as good as it does on a mobile. optimized system....
@@justacomment1657 There's a reason Intel is murdering AMD on almost all the productivity benchmarks. Hint: It ISN'T because of the P-cores...
@@glenndoiron9317 none of those benchmarks are virtualization stuff.
and if you seriously belive that intels thread director is working as intended you are delusional... 🙈
oh and BTW... those productivity benchmark you are talking about are all about tile based rendering....
this is not the only productivity application of a cpu...
It really has been the year of bad CPU launches, That is what I will remember 2024 for. let us hope 2025 is the year of good GPU launches
Heard about the 9800X3D lately, wouldn`t call it bad.
What do you mean didn’t go as planned? Intel said pre release its inferior performance, but better power savings. Check.
It's amazing how not long ago amd was considered bad and Intel king, but to be honest it took a lot of time for both to reach this status quo, like it took Intel bad cpus time and time again to do it, which is strange.
Brand name is a strong thing
intels saving grace is their memory controller and that's not even going to be enough..
Everything is going to have to be re-evaluated once these issues have been resolved in order to get a proper results.
So both AMD and Intel screwed up the release of their mainstream parts in the same way.
Stuff worked one way in their automated test labs on engineering boards and another way when the reviewers got the parts.
It sounds like both of them need to have a manual test team that tests new CPUs the way that a reviewer would PRIOR to launch.
Build PC on standard OEM parts, fresh Windows install, install games, play in interesting area, report any issues or differences from automated testing.
They both ended up getting a team to look into things after the bad reviews came out - seems pretty stupid to release early so you can get bad reviews and almost no sales
I like the honesty but i also am not convinced until they really do what they say. Them saying EVERYTHING can for sure be resolved… dunno… makes me feel not as great as they might think. Sometimes you f.u. and you cant fix. Maybe they can fix most of it but everything? Who knows that when they cant really tell why it even happened in the first Place? Its at least strange.
We trust but we verify.
Leo
The fact they still haven't done enough about the 13/14th gen issues gives me less than no confidence in them dealing with these problems.
Robert Hallock picked the wrong time to switch from AMD to Intel.
Well done.
Good on them for owning it. Let’s see what they do.
Are they? Owning it?
@ saying it’s their fault, no one else’s, and they are working on a list of issues and solutions to publish, I would say yes. I
@@KevPez-IS talk is cheap
@@BoraHorzaGobuchul you’re right, hopefully they act next
@@KevPez-IS yes, hopefully, were don't need a CPU monopoly
I know Intel had the top of the line chips for a long time, and that was before they went to P and E cores.. since then, they have ran into problems, may I suggest that while they work out some of the bugs in using both P and E cores, that they also create a couple of chips just using P cores.. and having a three tier lineup of chips, one with 16 P cores, one with 12 P cores, and one with 8 P cores.. no hyper threading.. 16 P cores 16 threads!! And add more Cache at all levels on Chip!! And advertise them as GAMING CHIPS!! And do so on the same new socket of 1851.. and also increase the number of PCIE Lanes on the chips, to 28 on the 8 core chip, 34 on the 12 core chip and 40 on the 16 core chip!! I have an IVY BRIDGE chip which has 40 lanes on it.. add more PCIE lanes, its a hard selling point!!
Intel got cocky in the early 2010's by dominating the Desktop/laptop and server. So they starter diversifying (SSD, memory, networking, 4G/5G, business, IoT, FPGA, autonomous driving and AI) and cut their funding to research in fabs. They also didn't think that they needed EUV for 10nm for cost reducing reasons which i believe led to the massive delay in 14nm.
I'll watch this space because I'm genuinely interested to see how it goes and I need to upgrade my CPU; it's been a lot of years.
I keep swinging between AMD and Intel, but because I don't solely game and want my rig to be very good at productivity and very good at games (but not enthusiast level), my focus is on the best balanced chip.
I don't give a toss about it being AMD or Intel; I really just want the best chip for the best price for my needs.
What does it mean that it only affects reviewers and not us?
If you buy a Z890 motherboard today and immediately update the BIOS using flashback you will avoid a number of the problems faced by we reviewers. Performance will still be problematic until the next round of fixes.
Leo
Very nice, good to own it, fix it, and move on
Even if these software bios/driver fixes are done, I will still hesitate to buy a new expensive motherboard for this new 1851 socket and some also expensive CU-DIMM memory, knowing that maybe the next Intel generation will change of socket architecture again...
I think the guy that appears in your video screenshot he must have been very sorry to leave AMD right now 😅😅😅
I can not laugh hard enough to the Motherboard companies :P
We can all see that there are FAR MORE Intel motherboards then AMD ones.
Specialy the EXTREME boards from all venders you see with Intel and NOT at AMD.
I think it has something to do with contracts between Intel and the motherboard companies.
I can NOT believe that they are happy with Intel right now because nobody is buying those boards :P
Well you would think companies like MSI would have done some initial tests internally, no one seemed to know anything about the launch based on what Leo showed us.
@@FrankLeCross-fo5kr MSI never does proper testing. It's their history. I don't know how people buy their crap...
The strange thing is that the speed and compute power does not match the gamaing performance. Makes me think intel focused at light weights gamer and more businesses. But I have a feeling something else else is happening which wasn’t intended. Looks like some is preventing a full access to the gpu or something is causing a limitation. Because coming from 13th gen I can say the ultra veel faster in normal working environments en rung VMware on it I could see the improvement. Something is limiting how the the cpu should feed the pci express or the gpu
that something is most likely the memory controller and the total stupid core layout where every other core complex has different latency problems... 😅 it's basically elevated zen2 problems but the competitive floor is way higher now..... running vmware on such a cpu is Alsow something I find highly unpredictable...
@ maybe your are right. I create and modify images for companies and decide to try out the new ultra 9 and was impressed by the performance because I am use to boating up 15 virtual machines and running atvthe same time and doing without having difficulty on the is cpu is very. My 13th gen fans would have kicked in like I am running cinebench 23. Anybody that doesn’t game that much will love it and not everything is about games 😁
I really wish Intel would publish a real time up to date list of known issues as opposed to going "there are issues". That would undo some of the lost good will from me with raptor lake
Not sure if an easy to point out list of issues is a good idea from a marketing perspective. It would be used by competitors as well. I'd definitely make it at least product specific. And, as a consumer, I don't care about issues on products that I do not own.
@@owlstead I agree it is not a good idea for marketing, my concern comes from intel doing a very bad job of communicating the issues around Raptor Lake and this doesnt feel that different, apart from the fact that very few people have (or are likely to have) these arrow lake parts
Looks like the available DDR5 only goes up to 8000MT with high latency when most boards will take up to 9000MT. So you are saying that the higher latency is not a concern? Is there an advantage of let's say 7200 over 8000? Also these sticks are only available in 32GB for a 48GB configuration. Why such an odd setup?
My problem is that I can't get the thing. It seems to be selling well regardless of the reviews.
They are going to go away for a month and audit the issues, but meanwhile it’s still on sale at full price. Fixes are not guaranteed.