@@szlaven8254sir are you aware of the multiple bsod and chip destruction by faulty 14700k. Once the chip is destroyed its lost. They are not even honouring their warranty
@@szlaven8254 its not alway about performance, in this case, it has to do more with how long it will last. I already seen issue with the 13 and 14th gen intel cores even in business landscapes that dont even care about performance. They simiply just stop working or started having other issues. Even had a few just pull more power to everything causing the whole system to over heat until some of the microcode stuff got release to fix some of that lol Like Intel side of things is just a mess atm. The main reason we got Intel was so we didnt have issues or problems like we saw with AMD stuff. On the flip side now, we're seeing more stable designs with AMD over Intel. Intel would still be in the lead even if it lost the performance crown if they could just make sure everything is stable enough.
@szlaven8254 i don't understand the coping, you can go watch from actual benchmarks like gamer nexus, the 7800X3D is 4% better on average than the winter heater 14900k with 200w
Bro ... Same. My last Intel CPU was the 6600k. It was great but I was so tired of Intel making me buy a whole new board for like 5% gain. Sold my 6600k and got a 1700... Never looked back.
The thing is Intel and AMD do SMT differently, although I don't know how they differ. And when Intel came out with it first it was called Hyper-Threading, but then AMD came out with their SMT it was called something else (although it still had the initials H.T.) but we are going way back to the Pentium 4 days. ALSO, when I was reading up on WHY Intel decided to ditch SMT is wasn't about performance - they KNOW that SMT increases performance for sure - it's about the amount of REAL ESTATE adding SMT to the chip takes up vs. the performance gains it brings. Intel figured (just as an example, I don't recall the numbers they used) for that 18% average gain from adding SMT, it took up something like 60% the size of one of their little core. and so using that space for SMT on all their big core to add a few more little core would bring more performance. This would NOT work for AMD tho, since they don't have "little" cores, they have 'DENSE' core... which are essentially THE SAME as a normal core, minus 100 to 200 MHz (I think) AND they are multi-threaded... So it would be pointless for them. :)
How do you get extra space when cutting hyper threading ? You get extra space cutting cache size that is why efficient cores have small caches compared to performance cores.
That was only during covid. People are broke broke the economy is hurting. They can try to scalp but they won't have any buyers and inventory will quickly grow. This ain't 2020.
But Puget said it had to subpoena - lol - Intel for the default setting, where as the MB makers were following the vague guidelines set by Intel, ie, Intel is still at fault if they say it CAN do it, but latter find out it CAN'T actually do it long term.
You can't "subpeona" a company for that sort of information. 😂 What they did was ask and then did as Intel said, whereas the MB vendors are coming up with their own settings to maximize bench,arks in reviews.
Hyper threading does improve performance but slightly and increases temps considerably. Removing hyper threading is correct approach for Intel to provide headroom to the main thread not the fake thread.
Only problem with your analysis: the Windows and Linux schedulers distribute the work on real cores first and then starts using the "fake" HT cores. So as long as you are not going full MC, activating or deactivating HT has no impact on temps. And it seems that the HT approach Intel uses is much weaker than the AMD approach.
Bro reported about nothing in regards to the 5090. Congrats, someone put 5090 in the description, like who cares bro. It told us nothing, and then ur argument for why cooler master ought to know like wtf. I care that’s why I’m mad. You need to improve your reporting skills pls
Not just that man but how can this coolermaster story "prove" it when multiple other stories that mention the 5090 don't? I agree man. It's also based on his opinion because he even says it in the story where he really doesn't think so about the typo. Still no OBJECTIVE proof. It's all subjective.
@@danielhulan3058 guy is just a clickbait monster, him and graphically challenged really never report anything of merit and have weak supporting arguments etc. idk why I keep watching these clowns.
Bro this channel has been putting 5090 in the thumbnail since like a month after the 4090 dropped and literally never says anything relevant about it. For a while i thought this was just a gossip channel that never reported anything real, just total guesses and assumptions with clickbait. But they’ve gotten a little better
intel isn't going bankrupt, no one in the tech industry would be allowed to buy it because of anti-trust so worst case the US government bails it out as a national security interest.
@@TheworthsearcherI think they're going bankrupt either. But their internals are actually weaker on servers than they are on the consumer market. They are barely making a profit, for the huge gross, in servers. That shocked me!
that 16 pin GPU connector that nvidia uses is way too undersized for use at 600W and beyond.. They should have used the regular ones that all other GPU uses.... the 8+8 power connectors... these have bigger terminals that can handle higher power without melting the connectors. just imagine, putting 600W of power on 16 connectors the size of the 1mmx1mm terminals (almost the same ones used on the motherboard panel connectors.)
@@jonasmartinez1739 that connector is only rated for 600W anyway. And tbh it should be rated lower. It's probably a little too unstable for the 4090 but anything below like 400W should be more than fine
I think one reason why Intel is trying to drop hyperthreading is due to various security flaws involved in Intel's implementation of hyperthreading. You can't have those security flaws if you don't have hyperthreading. I think Intel really needs to redesign their architecture, and dropping hyperthreading to restructure their cores is a good first step. Once they have the single threaded cores figured out, they can work on adding hyperthreading/SMT back into the cores or provide other enhancements.
Its so lame. They mentioned 5090 because its obviously coming out in the future. Its like saying GTA 6 will run on ps 5 and PS 6 and 7. We know its going to happen. Such click bait shit. Unsubscribe
On another note: The whole thing about big and little cores, or 'performance and efficiency' cores as Intel wants to call them... It's a dumb idea that seemed like a good idea, that needs to go away. Why it's a dumb idea: You are effectively causing a new bottleneck in the chain of processes going on when you are doing anything. The computer is having to decide which tasks need to go to which cores based on their complexity and other factors. Where as before it was more or less "If core available, take work and do it", it is now "If appropriate core is available, take appropriate work and do it". That's a pretty huge nuance of a problem, because if the cores that are capable of the work are not available... where does the work go? Sure enough, Intel ran into this problem basically right away with having to make a different version of the task scheduler code that handles this. So now it's fixed, sure... but as I understand this caused some new problems for cpu's that still do things 'the old way'. Which of course hurt AMD at the time, and so a new fix was released to fix the old fix. Unless I experienced some sort of other universe temporarily before writing this in this one... this is a problem that will keep on popping its head up any time they try to make a 'new' cpu with a different variation on how they handle their performance and efficiency cores. Honestly, I wouldn't be surprised if Intel's current problems with 13th and 14th gen also somehow stem back to this, just not realized yet. How? Well, different cores need different voltages depending on their capabilities, right? And most of the problem revolves around voltages spiking, and causing the minimum load to increase. I'm not sure, but I don't believe anyone has bothered to see which cores are failing first. Just that they are failing. I wonder if it's the efficiency cores that go first. They are after all effectively laptop cpu cores being given the lesser of desktop cpu core tasks. If the performance cores are spiking the voltage in a way that raises the minimum that affects all of them... Well... I wouldn't be surprised if it's ruining the efficiency cores first. And that's aside from the copper oxidization issue. The only reason it seemed like a good idea at the time, was because from a power efficiency point of view, it doesn't make sense to run power hungry cores running tasks that don't require a lot of power in the first place. But due to Intel having to push the power meter to keep up with AMD, this very thing might be part of what is biting them in the ass so to speak. Then again, this all could be wrong. Perhaps I am not remembering some aspect correctly. Perhaps I am wrong about the cores. But it wouldn't be the first time I've been pretty close to the target when it seemed like I was way off. I bought 10th gen, instead of 12th or 13th, because I suspected problems from voltage increases, and all of this was part of the logic going on in my head at the time. 10th gen was slightly better than 11th gen in some ways, but the motherboard I was looking at could take 11th gen Xeon later when I got the Arc card for AV1 encode. This would, as I figured at the time, allow me to side step all the potential issues Intel was about to face from my point of view. And I figure they are still in for a wild ride yet. I probably won't be buying anything brand new from Intel for at least 3 or 4 more generations, possibly as far as til 20th gen; if they aren't bankrupted at that point. Even if they get this voltage issue under control, they have found out the hard way they need to double down on their quality control too. The fact this was able to happen at all to them, means that they have some seriously deep rot going on in the company that needs to be rooted out. That takes time. Thankfully, some of that process has already started to occur at the time of writing this. One of the people in charge of the wafers that oxidized, has resigned. I forget his name, but I doubt he's getting that sort of job again in the future by any reputable company. Also, Intel will need time to figure out it actually needs to stop trying to compete with AMD, and instead focus on the things that people prefer to use them for the most. That's not gaming... that's servers. The folk running servers need those parts to be super reliable. They cannot have any issues. If they do, they will be replaced if erroneous enough. Like we are seeing with some game devs now with their servers. And Intel has new players trying to enter this market as well. Intel needs to focus on servers, not gaming rigs. Not laptops either. Those are secondary. They need to focus on servers, and to a lesser extent, workstations; as workstations are the second form of computing that needs to be absolutely infallible as well. The result of this will be like in the old days to some extent. Intel will release new server hardware, and in due time, it trickles down to the consumer market for desktop use; perhaps with a few tweaks to make it better for gaming. This always got them a respectable turnout, while securing their profit margins with the server and workstation clientele. But to do that, means going back to the old ways of cores as well. They'll have to drop the big/little setup. For a server or workstation, this doesn't make sense to have. All tasks are important tasks for these units. Sure, saving on power usage is nice still. But not part of the equation in these units cases. They are meant to be hungry. We must fill them. And to their credit, it does look like they are sort of taking this approach again. But they are trying to keep their big/little setup in some way. Understandable, since it means losing lots of invested time and money and labour by dropping it. A shame, really. But they dropped Optane, and they should have kept that one. Should have kept that one, and dropped big/little instead. Big/little makes sense for mobile devices. They don't do super powerful tasks usually. Desktop... this applies less, but can still apply in some scenarios. Anything workstation or server related though... and that's where Intel's bread and butter came from for a long time... well, that's not a good idea to mess with. You don't mess with your bread and butter. And they messed with their bread and butter. If they don't reverse course soon, or find a safe path through these treacherous waters; they will go bankrupt due to it. Their profit markets aren't the consumers that normally need this sort of CPU.
e-cores are a stupid gimmick anyway. Ryzen CPUs don't have that BS yet they still use less power than an equivalent Intel chip (and they don't degrade).
My God! 40fps @17W . Is AMD sleeping? This is what we get after so many years of "improvements"? I have to give it to Valve! They did a great thing with their APU!
So the argument is... Intel motherboard vendors have been putting dangerous voltages, going over Intel recommended specifications for a long time over multiple generations, and now that they have gone too far that's now completely Intel's fault? Even though Intel released correct voltage specs from day 1, they're still responsible for third party manufacturers dodgy products?
Confirmed my ass what a clickbait title. A PSU listing that it supports the 5090 isnt confirmation we all know its gonna launch i just want to know when.
This is kinda dumb. Ryzen is designed to work best with hyper-threading. Intel is not. If they gave an Intel chip that does have hyper-threading enabled the same test they will find the same results
Oh god... disabling HT and the e-Cores does not do any harm to gaming performance as of right now. Disabling HT and adding 8 e-Cores has way more multi-threading performance than 8 p-Cores with HT on. So what Intel does is: 1) retaining single core performance crown with ht disabled p-cores, that need less voltage and less amps therefore run cooler 2) reitaining multi core performance crown with added e-cores, because 8+20 is better than 16+HT. What is not stated in this video, but is already leaked info: the upcoming e-cores are going to be on par performance wise with current p-cores.
You must understand Hyperthreding. It was a trick implemented on Pentium 4 to gain some processor speed. "Virtual" core provided by hypertreading is ~20% speed of physical core. So, 12/24 core means 12 x 100% + 12 x 20% performance = 14.4 physical cores. Then, 16 physical will be faster and simpler than 12/24.
6:50 the issue is that the VID table is all wrong though, this isn't something they can change, isn't it baked into the micro code, and added to that post vdroop the CPU requests too much and the voltage spike is insane 1.5-1.6v+
Testing SMT on a Ryzen to compare it with an Intel, which uses a markedly different design and is an expert in single thread performance but historically was lagging behind AMD with multithreading, is a bit of a stretch. Also, what it effectively did and still does is use potentially unused resources to run a secondary task, having an indirect optimization effect compensating for applications not making full use of a core. So doing away with SMT makes sense only if you also reduce the number of processes a CPU pipeline can handle in parallel therein reducing the potential of SMT, such as with compact efficiency oriented CPU designs. If Intel wants to drop production costs or simply outdo the competition in core count possible, i don't know. I do know though that virtual / shared cores have little chance compared to the real deal.
The 5090 isn't exactly "coming soon". They will announce it in what, January 2025; it will just be a paper launch with only privileged content creators and reviewers having access to it at first. Then give or take 2 weeks later the only way to buy that thing will be from resellers on eBay at about 200% MSRP, or more. It won't be regularly available on Newegg or the likes until maybe March 2025.
I'm excited going to the beach and relaxing in another country enjoying the sights and great things to do along with the culinary dishes. I still love to edit and make videos, but I think most of the "work" I will be doing is camera work. No deadlines, so I can get buy with some server GPU or whatever is out there right below the cutting edge.
The 5$$$ series is yesterday's news. It's time to talk about NVIDIA's 16,000 series GPU, which can power your games, be removed from the computer, and transform into an AI friend and companion! It even knows how to file your taxes!
Dont know if this has anything to do with it or not but I've been running a 13900K for over a year now and have not had these problems. I have the aftermarket cpu bracket which helps keep temps down far better than the OEM bracket , along with an ASUS ROG STRIX LC II 360 AIO which hardware kanucks showed with testing that it was one of the best ones, a ton of fans as well in my case BUT Im not running z690 board which is probably the biggest culprit of the crashing issue. Instead Im on a rog b760i motherboard and Im still running the bios released from June of 2023 IIRC . Ive played countless hours beating Cyberpunk at max settings several times and never had a crash to give you some usage perspective. Im sure intel is partially to blame but my advice for anyone looking to run this cpu or a 14900k is to stay away from the z690 boards, run it with the oem cpu limit settings, do the aftermarket cpu bracket, a good liquid AIO and have plenty of airflow. Yeah the issue is over volting however my experience so far has been great so I think the combination of not running a board engineered to use as much power as possible and keeping the cpu as cool as possible helps tremendously with stability. On a side note : I'll either wait for their true next gen CPUs or go for Nvidia's if its out by then but until then Im not updating anything my voltage is already well within Intel's acceptable limit and again I believe its because of the board and the oem settings .
They got a lot to work out unfortunately. I started looking into some financial stuff about Intel and their stock downturn wasn't even on this issue. It's about falling profits, loss of sales in several sectors, huge expenses building out their Fabs, and stopping dividend payments. They're one of the few in the chip Industry to *not* have growth over the last few years. Most investors aren't even factoring this 13 and 14 gen issue in yet! But I don't think they're doomed. They'll likely do what AMD did and sell the Fab. Besides the US is currently highly concerned about this area so the government isn't gonna let them fail. But they need to skin down and focus on good designs. Hopefully in a few years we'll have great competition from both(Intel and AMD) and neither will be in a dominant position.
Intel is 100x bigger than AMD. They have a lot of time. Even if Intel was behind AMD for ten years, AMD still might not reach the same market share as Intel. Plus, ARM is trying to take some of the pie now too.
SMT adding performance is nothing new. We all (should) have already known that SMT being enabled on AMD but not available on some Intel CPU's due to the vulnerabilities they suffer just a few more of than AMD. Reason being that back when that was all news for us all, the main topic was that the only 'fix' for the security issue was to turn off SMT or any form of hyperthreading and predictive options in the bios. And this was hated by basically all back then, because it meant taking a whopping hit to the performance. All to avoid a security flaw. So of course if AMD is still using SMT and other predictives, they are going to have more performance than Intel's cpu's which do not have it available. Honestly, it's one of the most 'well F'ing DUH!" kind of moments in tech right now. The only real neat thing about it, is the lack of power consumption increase. That power consumption barely budging is really nice, because otherwise the tech does normally increase power consumption a bit more than that. Not a lot, but more. Enough that turning it off is a legitimate way to reduce temps on a hot day where the extra threads are just not required. (Like playing an older game for example that is gonna be just fine with 8 full cores, cause it was programmed when 2 cores was the norm.)
To be honest.. I'm not excited about next gen AT ALL. I'm burnt out on trying to keep up with $2000 graphic cards. I'm at a 7900xtx liquid devil and perfectly happy with it and will be sticking with it
I've just decided now that even if AMD put AI into their notebook CPU naming scheme, that doesn't mean you have to say it. I'm just going to say "Ryzen 9 HX370" because it literally makes no difference.
Why are we comparing AMD's GPU to Intel's? Intel barely started their discreet GPU line, and AMD has been on the GPU market for decades. Intel will be irrelevant in the GPU space for at least another 5 years. Compare AMD to NVDIA instead.
It was awhile ago that i saw it so cant remember exactly what channel it was on. But some time ago i saw a video about why intel where removing SMT. in it they said the extra space taken up, and the increased architecture complexity of the cores to support SMT was not worth it when they could allocate that space to AI processing or additional cores. What i took from it was that they were planning to skip SMT in favour of replacing that space with additional compute power, for example a 12 core with SMT may allow a 16 full core instead with the same die size. I'm sure people can google around and find the video, may have been gamers nexus or something but kind of makes sense, so lets see what they release in the future. I've had AMD for awhile now (AMD and Apple, just got rid of my last intel system a few weeks ago) but only because they had the more suitable products for me at the time. I'm not going to hold having a faulty product against intel as it can happen to any company but how they deal with it in the long run will impact how likely i am to buy intel in the future.
The first time for THAT was a couple decades or so ago, the Thunderbird-series vs the then-current Pentium III. Thunderbird was slightly MORE efficient IPC, clocked up to 1.4 Ghz where Pentium III a couple years later EVENTUALLY maxed out at 1.33, and other such issues. That is why Intel went to a completely different arch for the Pentium IV, that was lower IPC but could clock a TON higher and eventually passed the Athlon series on actual performance.
As I have been saying, it is the motherboards that is doing this. The MB default is NOT the processor default. They are all overclocking the processors out of the box. Puget have been setting more conservative settings and when they did thst, failure rates dropped, even lower than AMD systems in fact 😂
I am pro small cores, but just a handful of the them for low power situations in a laptop. Also, it might make sense in server situations where you want to pack as many as possible under power and space constraints, but, for normal PC or high performance environments, there is really no need for them, especially this lobsided 2x8 PxE configuration. New laptops feel much slower than older ones. We are going backwards!
Sees something in an Amazon product description. And immediately claims that it is a confirmation... Your RUclipsr really tries to create a headline out of every little piece of information! Get a real job!
So sad to see what is going with with Intel lately in the media.... but good to see Puget reporting good things. Hopefully Intel makes it through this "ok".
I am not sure if we could trust the data from Puget systems cause Wendell has found that Datacenters using 13th and 14th Gen CPU's were also failing and those are using motherboards from Supermicro if there is one advantage that Datacenter focused motherboards has over Retail is that they are focused on Stability so they are not pushing Processors as hard nor are they pushing the voltages too high either so if those are experiencing failures then retails is even more screwed.
Of course AMD's iGPU, based on their refreshed RDNA 3 architecture - an architecture originally designed to compete with Nvidia's 40-series cards - will outperform Intel's iGPU based on fucking Alchemist - an architecture that barely competes with Nvidia's RTX 3060 on it's highest tier card. Intel being this "close" to AMD in performance on their iGPU is more of a win for Intel than it is for AMD
My last new intel product was Celeron 333Mhz and Nvidia Geforce 7000 series GPU. If i would buy Intel or Nvidia it'll be as used ones. Atm im at i7 980x, Gtx 980 bought used.
@@avoxxtea 1,000 to 1,200 ?????? Someone else who comments and doesn't know what they're talking about You probably won't even be able to buy a 5080 for that
Ambulance chasing law firms looking to make a quick buck, yeah, they are suing on behalf of people that got bad CPUs, but like most class action law suits, the law firm gets most of the money.
-If we combine 13th and 14th gen, they're about the same as AMD Zen 4- AMD Zen 4 has higher defect rates ... on Puget systems. Which override the BIOS settings to more sane defaults. In other words it is again a testament that 13th and 14th gen aren't that bad, if configured correctly. But nobody knew that unlike gen 1st to 12th, that now you HAVE to manually configure the protections and voltages and voltage limits and load line calibrations and so on, in order to have a safe, stable CPU (which might also have good performance and much saner power draws - like 250W in full load - and temperatures). It is both the motherboard manufacturers and Intel's fault here, but given the history, it's basically mostly Intel's fault here for giving such vague guidelines.
I can't wait for 5090, I'm tired of driver timeouts with my xtx, I probably won't buy another amd gpu again. My 6950xt was pretty good but my 7900xtx has nothing but problems
What brand of xtx did you get? my XFX 7900xtx has been solid since it launched. You may have just gotten a bad card or maybe didn't wipe the old drivers from a previous GPU. I never get any driver timeout errors.
@@matilija and you don't hear more people talking about their intel card being bad too 😆 There is a reason the adoption of 7000 series AMD GPU is not as good as 6000 series Because of how Solid 6000 series is and probably better aged driver support
Are they seeling more i5's than i9's? remember that higher the power, more issues it has.WHY COLLUMN IS MERGED FOR ALL CPU'S? if they sell LP laptops then it can hide all of it. 10000 laptops 1137g711 whatever ones that have 1% failure rate AND 500 i9's that have 500 returns, would still show 1.05% failure rate. remeber current claims are that i9's fail at about 50% per year rate, i7 & i5 at between 10 and 30% and it mostly K cpu's. something like i7-13700T afaik, have 0 failure rate, from this problem. its BC that cpu is so low power, it would take years to develop that same degradation level. ESPECIALLY shop caught, is unrelated to the problem. 1% damage is most likely shipping damage, stuff that got dropped. We can remove that from all of them.. most of those is either manufacturing defect or shipping damage, and it just shows they are selling way more CHEAP board of AMD, than intel's ones. (especially cheaper ones have worse boxes, so they break more often) 7:40 this is clearly that they just dont get how that damage happens. it needs TIME. CPU DEGRADES OVER TIME. so now, 2 month from sale, they see 0.5%, then 1Y in they will see 15%, then 2Y in they will see 35%. is their warranty covering that long? this graph clearly shows they either dont understand the problem, OR they are trying to hide it.
Intel's decision to develop and use their own GPU proved to be their own undooing. They would have been in a much better position if they focused on CPU and let the third-party laptop manufacturers use Nvidia GeForce. Now Intel is 3rd place at everything. Apple, too, should have continued using X86 instead of developing their ARM processors for laptop. Many of my friends boughtMac Proo because it had fast X86 processo and GeForce graphics card. You could play Windows games on it. With ARM, games now have to be emulated. Als, ARM doesn't come with a fast GPU. Sad, but that's how companies rise and fall, all because of some poor decisions by thei executive. Tim Cook doesnt have the vision like Steve Jobs. Neither does Pat Gelsinger.
1:44 the ryzen 9 7950x3d has Problems with utelysing the x3d Cache. If you Do not game It trys to Use the x3d Cache. The Performance could be hold back because of this.
If I can give a little bit of constructive criticism about your moderation: Could you vary your use of the word 'but' a bit more? You use that word a little excessively sometimes. I mean, maybe you consider that a signature style of yours, but as we say here in Germany, 'that has a beard already' - which means it has lost its appeal, is getting old. In the end it's your decision on how you do things, as it's your channel. I just wanted to share my perception as a viewer - and most creators usually want some feedback.
Was about to unsubscribe for yet another click bait 5090 video and realised I’d already unsubscribed! It hardly gets a mention at the end of the vid and it’s not even worth mentioning! Sorry for being negative but that’s enough for me!
Intel's c-suite management level has to make change immediately, they need a phenomenon leader like Lisa. The overall cpu architecture need a redesign, what intel experiencing is still better than AMD's dark age. Intel needs its own Ryzen, if intel keeps patching current architecture just to save money on development............we might gonna lose another blue giant like power pc.....
Mind blown. RTX 5090 will be called RTX 5090. You heard it here first.
AMD doesn't need do destroy anything, Intel already destroys itself
@@szlaven8254false, also you shouldn't be coping this hard if you're not sponsored by Intel yk
@@szlaven8254sir are you aware of the multiple bsod and chip destruction by faulty 14700k. Once the chip is destroyed its lost. They are not even honouring their warranty
@@szlaven8254 its not alway about performance, in this case, it has to do more with how long it will last. I already seen issue with the 13 and 14th gen intel cores even in business landscapes that dont even care about performance. They simiply just stop working or started having other issues. Even had a few just pull more power to everything causing the whole system to over heat until some of the microcode stuff got release to fix some of that lol Like Intel side of things is just a mess atm. The main reason we got Intel was so we didnt have issues or problems like we saw with AMD stuff. On the flip side now, we're seeing more stable designs with AMD over Intel. Intel would still be in the lead even if it lost the performance crown if they could just make sure everything is stable enough.
@szlaven8254 i don't understand the coping, you can go watch from actual benchmarks like gamer nexus, the 7800X3D is 4% better on average than the winter heater 14900k with 200w
LOL facts!!
Ryzen 4070 better
4070 ? we dont have in ryzen ,,, you mean 7800x3D is best of best
@@raminpro9765 no ryzen 7900xt is the best
@@predator9909 I prefer the ryzen 3090-Ti
@@predator9909No
@@raminpro9765 I guess you haven't checked the news and the market lately
There's going to be a 5090? No way... lol
U mean ryzen 5090
I switched to AMD when they introduced Ryzen cpus. I haven't purchased an Intel product since
Bro ... Same. My last Intel CPU was the 6600k. It was great but I was so tired of Intel making me buy a whole new board for like 5% gain. Sold my 6600k and got a 1700... Never looked back.
Same here. No more Intel, no more Nvidia.
Me too. My last Intel CPU was a Pentium II.
Same, 6700K was my last cpu, I’m currently using 5900X
Same for me, good value, and I can keep my mobo longer and drop in newer CPUs. The longevity is insane compared to Intel.
I don't miss Intel.
Ht is bad for Intel because how they use and abuse power among other things
5090 confirmed 👀🤦♂️ The click bait from this channel recently has been tiring
fr
Click bait and listen about news you already know lmao
The thing is Intel and AMD do SMT differently, although I don't know how they differ. And when Intel came out with it first it was called Hyper-Threading, but then AMD came out with their SMT it was called something else (although it still had the initials H.T.) but we are going way back to the Pentium 4 days.
ALSO, when I was reading up on WHY Intel decided to ditch SMT is wasn't about performance - they KNOW that SMT increases performance for sure - it's about the amount of REAL ESTATE adding SMT to the chip takes up vs. the performance gains it brings. Intel figured (just as an example, I don't recall the numbers they used) for that 18% average gain from adding SMT, it took up something like 60% the size of one of their little core. and so using that space for SMT on all their big core to add a few more little core would bring more performance.
This would NOT work for AMD tho, since they don't have "little" cores, they have 'DENSE' core... which are essentially THE SAME as a normal core, minus 100 to 200 MHz (I think) AND they are multi-threaded... So it would be pointless for them. :)
Intel copyrighted the name "Hyperthreading" - AMD eventually matched the technique, but had to use a different name for it.
But what are they even doing with the extra space? The performance uplift still doesn't show
How do you get extra space when cutting hyper threading ? You get extra space cutting cache size that is why efficient cores have small caches compared to performance cores.
@@melheno It does take SOME die space to impliment hyperthreading.
@@bricefleckenstein9666 SMT was patented by AMD in 1995. The initial HT stand for Hyper Transport - the bus technology which came for K8 to Excavator.
I can't wait for the 2.5k msrp (scalped for 3k) 5090.
😭
That was only during covid. People are broke broke the economy is hurting. They can try to scalp but they won't have any buyers and inventory will quickly grow. This ain't 2020.
I think you mean scalperprice of 4k, not 3
In Croatia 4090 2100-2500€ anyways.
cant wait for the 5070 ti to have 12gb of vram!
But Puget said it had to subpoena - lol - Intel for the default setting, where as the MB makers were following the vague guidelines set by Intel, ie, Intel is still at fault if they say it CAN do it, but latter find out it CAN'T actually do it long term.
You can't "subpeona" a company for that sort of information. 😂
What they did was ask and then did as Intel said, whereas the MB vendors are coming up with their own settings to maximize bench,arks in reviews.
Puget also said AMD Ryzen cpus actually have a higher failure rate than 13 and 14th gen.
Hyper threading does improve performance but slightly and increases temps considerably. Removing hyper threading is correct approach for Intel to provide headroom to the main thread not the fake thread.
Only problem with your analysis: the Windows and Linux schedulers distribute the work on real cores first and then starts using the "fake" HT cores. So as long as you are not going full MC, activating or deactivating HT has no impact on temps.
And it seems that the HT approach Intel uses is much weaker than the AMD approach.
Bro reported about nothing in regards to the 5090. Congrats, someone put 5090 in the description, like who cares bro. It told us nothing, and then ur argument for why cooler master ought to know like wtf. I care that’s why I’m mad. You need to improve your reporting skills pls
Not just that man but how can this coolermaster story "prove" it when multiple other stories that mention the 5090 don't? I agree man. It's also based on his opinion because he even says it in the story where he really doesn't think so about the typo. Still no OBJECTIVE proof. It's all subjective.
@@danielhulan3058 guy is just a clickbait monster, him and graphically challenged really never report anything of merit and have weak supporting arguments etc. idk why I keep watching these clowns.
Bro this channel has been putting 5090 in the thumbnail since like a month after the 4090 dropped and literally never says anything relevant about it. For a while i thought this was just a gossip channel that never reported anything real, just total guesses and assumptions with clickbait. But they’ve gotten a little better
damn Intel can’t even do anything to AMD and if Intel goes bankrupt then we might face serious issues.
intel isn't going bankrupt, no one in the tech industry would be allowed to buy it because of anti-trust so worst case the US government bails it out as a national security interest.
They never go bankrupt. Believe me! 🙂 They still have their tricks at the Notebook and Server segment.
@@Theworthsearchermmmmm... Hopefully
@@TheworthsearcherI think they're going bankrupt either. But their internals are actually weaker on servers than they are on the consumer market. They are barely making a profit, for the huge gross, in servers. That shocked me!
We will reach a point where we have an Nvidia cpu and an intel gpu in our pcs.
I hope the 5090 comes with a better power connector 😊
I'm pretty sure they fixed it a while ago
@@justmatt2655 the fix still burns out.
It hasn’t been fixed and still has the same exact issues
that 16 pin GPU connector that nvidia uses is way too undersized for use at 600W and beyond.. They should have used the regular ones that all other GPU uses.... the 8+8 power connectors... these have bigger terminals that can handle higher power without melting the connectors. just imagine, putting 600W of power on 16 connectors the size of the 1mmx1mm terminals (almost the same ones used on the motherboard panel connectors.)
@@jonasmartinez1739 that connector is only rated for 600W anyway. And tbh it should be rated lower. It's probably a little too unstable for the 4090 but anything below like 400W should be more than fine
How does the Ryzen 4070 compare?
I was wondering the same thing 😂
I am confused I missed something is the ryzen 3080 4070 and 7090 thing like a joke going around
@@avoxxtea Ryzen 4070 is the best GPU on the market
@@avoxxtea ZTT (Zach's Tech Turf) Shorts is the answer you seek
@@ChairmanMeow1 4070 ? we dont have 4070 in ryzen
I think one reason why Intel is trying to drop hyperthreading is due to various security flaws involved in Intel's implementation of hyperthreading. You can't have those security flaws if you don't have hyperthreading. I think Intel really needs to redesign their architecture, and dropping hyperthreading to restructure their cores is a good first step. Once they have the single threaded cores figured out, they can work on adding hyperthreading/SMT back into the cores or provide other enhancements.
AMD is about to finish off Intel with the 9000 series 👀
Intel: Let's get rid of SMT
Also Intel: You must now only use ONE DIMM slot for our motherboards.
and do not run unreal engine games, just in case
Code has nothing to do with 13 and 14th gen Intel CPU's ! It all has to do with RUST & OXIDATION !!!
We need this new APU for desktops, I have a few min ITX PC to build and would love it.
Ryzen 8000 series is already a thing though
the igpu in the 8700G is like gtx 1660 level
sounds like they need to cut cpu prices in half and we all take a gamble.
Its so lame. They mentioned 5090 because its obviously coming out in the future. Its like saying GTA 6 will run on ps 5 and PS 6 and 7. We know its going to happen. Such click bait shit. Unsubscribe
what if my cpu hasn't showed any signs of instability or crashing its a i7-14700k I had it for 5 months. Am I safe?
Just stop using it until the microcode patch comes
lol, i would return that pos asap
On another note: The whole thing about big and little cores, or 'performance and efficiency' cores as Intel wants to call them... It's a dumb idea that seemed like a good idea, that needs to go away.
Why it's a dumb idea: You are effectively causing a new bottleneck in the chain of processes going on when you are doing anything. The computer is having to decide which tasks need to go to which cores based on their complexity and other factors. Where as before it was more or less "If core available, take work and do it", it is now "If appropriate core is available, take appropriate work and do it".
That's a pretty huge nuance of a problem, because if the cores that are capable of the work are not available... where does the work go?
Sure enough, Intel ran into this problem basically right away with having to make a different version of the task scheduler code that handles this. So now it's fixed, sure... but as I understand this caused some new problems for cpu's that still do things 'the old way'. Which of course hurt AMD at the time, and so a new fix was released to fix the old fix. Unless I experienced some sort of other universe temporarily before writing this in this one... this is a problem that will keep on popping its head up any time they try to make a 'new' cpu with a different variation on how they handle their performance and efficiency cores.
Honestly, I wouldn't be surprised if Intel's current problems with 13th and 14th gen also somehow stem back to this, just not realized yet. How? Well, different cores need different voltages depending on their capabilities, right? And most of the problem revolves around voltages spiking, and causing the minimum load to increase.
I'm not sure, but I don't believe anyone has bothered to see which cores are failing first. Just that they are failing. I wonder if it's the efficiency cores that go first. They are after all effectively laptop cpu cores being given the lesser of desktop cpu core tasks. If the performance cores are spiking the voltage in a way that raises the minimum that affects all of them... Well... I wouldn't be surprised if it's ruining the efficiency cores first.
And that's aside from the copper oxidization issue.
The only reason it seemed like a good idea at the time, was because from a power efficiency point of view, it doesn't make sense to run power hungry cores running tasks that don't require a lot of power in the first place. But due to Intel having to push the power meter to keep up with AMD, this very thing might be part of what is biting them in the ass so to speak.
Then again, this all could be wrong. Perhaps I am not remembering some aspect correctly. Perhaps I am wrong about the cores.
But it wouldn't be the first time I've been pretty close to the target when it seemed like I was way off.
I bought 10th gen, instead of 12th or 13th, because I suspected problems from voltage increases, and all of this was part of the logic going on in my head at the time. 10th gen was slightly better than 11th gen in some ways, but the motherboard I was looking at could take 11th gen Xeon later when I got the Arc card for AV1 encode. This would, as I figured at the time, allow me to side step all the potential issues Intel was about to face from my point of view.
And I figure they are still in for a wild ride yet. I probably won't be buying anything brand new from Intel for at least 3 or 4 more generations, possibly as far as til 20th gen; if they aren't bankrupted at that point. Even if they get this voltage issue under control, they have found out the hard way they need to double down on their quality control too. The fact this was able to happen at all to them, means that they have some seriously deep rot going on in the company that needs to be rooted out.
That takes time. Thankfully, some of that process has already started to occur at the time of writing this. One of the people in charge of the wafers that oxidized, has resigned. I forget his name, but I doubt he's getting that sort of job again in the future by any reputable company.
Also, Intel will need time to figure out it actually needs to stop trying to compete with AMD, and instead focus on the things that people prefer to use them for the most. That's not gaming... that's servers. The folk running servers need those parts to be super reliable. They cannot have any issues. If they do, they will be replaced if erroneous enough. Like we are seeing with some game devs now with their servers. And Intel has new players trying to enter this market as well.
Intel needs to focus on servers, not gaming rigs. Not laptops either. Those are secondary. They need to focus on servers, and to a lesser extent, workstations; as workstations are the second form of computing that needs to be absolutely infallible as well.
The result of this will be like in the old days to some extent. Intel will release new server hardware, and in due time, it trickles down to the consumer market for desktop use; perhaps with a few tweaks to make it better for gaming.
This always got them a respectable turnout, while securing their profit margins with the server and workstation clientele.
But to do that, means going back to the old ways of cores as well. They'll have to drop the big/little setup. For a server or workstation, this doesn't make sense to have. All tasks are important tasks for these units. Sure, saving on power usage is nice still. But not part of the equation in these units cases. They are meant to be hungry. We must fill them.
And to their credit, it does look like they are sort of taking this approach again. But they are trying to keep their big/little setup in some way. Understandable, since it means losing lots of invested time and money and labour by dropping it. A shame, really. But they dropped Optane, and they should have kept that one. Should have kept that one, and dropped big/little instead.
Big/little makes sense for mobile devices. They don't do super powerful tasks usually. Desktop... this applies less, but can still apply in some scenarios. Anything workstation or server related though... and that's where Intel's bread and butter came from for a long time... well, that's not a good idea to mess with. You don't mess with your bread and butter.
And they messed with their bread and butter. If they don't reverse course soon, or find a safe path through these treacherous waters; they will go bankrupt due to it. Their profit markets aren't the consumers that normally need this sort of CPU.
Some people disabled all e-cores as mitigation, we'll see what happens with crippled chip durability.
e-cores are a stupid gimmick anyway. Ryzen CPUs don't have that BS yet they still use less power than an equivalent Intel chip (and they don't degrade).
@@Psythik True, intel still stuck on high nodes just pumping more voltage.
Hi, Is there RDNA 3.5 available for desktop, after 8000 series when will AMD launch next gen APU for desktop?
My God! 40fps @17W . Is AMD sleeping? This is what we get after so many years of "improvements"? I have to give it to Valve! They did a great thing with their APU!
So the argument is... Intel motherboard vendors have been putting dangerous voltages, going over Intel recommended specifications for a long time over multiple generations, and now that they have gone too far that's now completely Intel's fault? Even though Intel released correct voltage specs from day 1, they're still responsible for third party manufacturers dodgy products?
Idc if it uses less energy, I just want my cpu not to crash every 30mins under high load.
Confirmed my ass what a clickbait title. A PSU listing that it supports the 5090 isnt confirmation we all know its gonna launch i just want to know when.
typical on this channel we re used to it xd
october as always?
This is kinda dumb. Ryzen is designed to work best with hyper-threading. Intel is not. If they gave an Intel chip that does have hyper-threading enabled the same test they will find the same results
Oh god... disabling HT and the e-Cores does not do any harm to gaming performance as of right now. Disabling HT and adding 8 e-Cores has way more multi-threading performance than 8 p-Cores with HT on. So what Intel does is:
1) retaining single core performance crown with ht disabled p-cores, that need less voltage and less amps therefore run cooler
2) reitaining multi core performance crown with added e-cores, because 8+20 is better than 16+HT.
What is not stated in this video, but is already leaked info:
the upcoming e-cores are going to be on par performance wise with current p-cores.
to be fair, all amd has to do is not rust away and dont suck 400 watts to play a game and will beat the crap out of intel
I'm an adrenaline seeker, i run a motorcicle at high speed without a helmet, i rock climb with no gear, i purchased a 14900K
Be sure to use it at the default settings!
and run it at unlimited power limits
You must understand Hyperthreding. It was a trick implemented on Pentium 4 to gain some processor speed. "Virtual" core provided by hypertreading is ~20% speed of physical core. So, 12/24 core means 12 x 100% + 12 x 20% performance = 14.4 physical cores. Then, 16 physical will be faster and simpler than 12/24.
6:50 the issue is that the VID table is all wrong though, this isn't something they can change, isn't it baked into the micro code, and added to that post vdroop the CPU requests too much and the voltage spike is insane 1.5-1.6v+
I think Intel is killing Intel at this point
Testing SMT on a Ryzen to compare it with an Intel, which uses a markedly different design and is an expert in single thread performance but historically was lagging behind AMD with multithreading, is a bit of a stretch.
Also, what it effectively did and still does is use potentially unused resources to run a secondary task, having an indirect optimization effect compensating for applications not making full use of a core. So doing away with SMT makes sense only if you also reduce the number of processes a CPU pipeline can handle in parallel therein reducing the potential of SMT, such as with compact efficiency oriented CPU designs.
If Intel wants to drop production costs or simply outdo the competition in core count possible, i don't know. I do know though that virtual / shared cores have little chance compared to the real deal.
The 5090 isn't exactly "coming soon". They will announce it in what, January 2025; it will just be a paper launch with only privileged content creators and reviewers having access to it at first. Then give or take 2 weeks later the only way to buy that thing will be from resellers on eBay at about 200% MSRP, or more. It won't be regularly available on Newegg or the likes until maybe March 2025.
clickbait title
I'm excited going to the beach and relaxing in another country enjoying the sights and great things to do along with the culinary dishes.
I still love to edit and make videos, but I think most of the "work" I will be doing is camera work.
No deadlines, so I can get buy with some server GPU or whatever is out there right below the cutting edge.
The 5$$$ series is yesterday's news. It's time to talk about NVIDIA's 16,000 series GPU, which can power your games, be removed from the computer, and transform into an AI friend and companion! It even knows how to file your taxes!
Dont know if this has anything to do with it or not but I've been running a 13900K for over a year now and have not had these problems. I have the aftermarket cpu bracket which helps keep temps down far better than the OEM bracket , along with an ASUS ROG STRIX LC II 360 AIO which hardware kanucks showed with testing that it was one of the best ones, a ton of fans as well in my case BUT Im not running z690 board which is probably the biggest culprit of the crashing issue. Instead Im on a rog b760i motherboard and Im still running the bios released from June of 2023 IIRC . Ive played countless hours beating Cyberpunk at max settings several times and never had a crash to give you some usage perspective. Im sure intel is partially to blame but my advice for anyone looking to run this cpu or a 14900k is to stay away from the z690 boards, run it with the oem cpu limit settings, do the aftermarket cpu bracket, a good liquid AIO and have plenty of airflow. Yeah the issue is over volting however my experience so far has been great so I think the combination of not running a board engineered to use as much power as possible and keeping the cpu as cool as possible helps tremendously with stability. On a side note : I'll either wait for their true next gen CPUs or go for Nvidia's if its out by then but until then Im not updating anything my voltage is already well within Intel's acceptable limit and again I believe its because of the board and the oem settings .
The power advantage can only exists if the chip removes the physical components for SMT
Removing SMT is really stupid, but hey. Intel is aiming for that 1€ stock price, so I approve their move.
I'm most excited about the Strix Halo APU. Hopefully, the future of mainstream gaming is powerful but affordable iGPUs.
This is not good. If there is no competition prices will rise.
They got a lot to work out unfortunately. I started looking into some financial stuff about Intel and their stock downturn wasn't even on this issue. It's about falling profits, loss of sales in several sectors, huge expenses building out their Fabs, and stopping dividend payments. They're one of the few in the chip Industry to *not* have growth over the last few years. Most investors aren't even factoring this 13 and 14 gen issue in yet!
But I don't think they're doomed. They'll likely do what AMD did and sell the Fab. Besides the US is currently highly concerned about this area so the government isn't gonna let them fail. But they need to skin down and focus on good designs. Hopefully in a few years we'll have great competition from both(Intel and AMD) and neither will be in a dominant position.
Intel is 100x bigger than AMD. They have a lot of time. Even if Intel was behind AMD for ten years, AMD still might not reach the same market share as Intel. Plus, ARM is trying to take some of the pie now too.
@@Trifler500 nope. amd is worth 218 billion while intel is worth 112.
@@bal7ha2ar AMD has much lower market share, and always has.
@@Trifler500 that is true but its changing and changing fast.
Puget Systems basically took a lot of AMD cpus, overclocked them to the point of destruction for Intel then blamed AMD as making bad chips..
Thank God I'm still rocking my 9900 k haha
Clickbait, of course the 5090 will be called the 5090... smh.
SMT adding performance is nothing new. We all (should) have already known that SMT being enabled on AMD but not available on some Intel CPU's due to the vulnerabilities they suffer just a few more of than AMD. Reason being that back when that was all news for us all, the main topic was that the only 'fix' for the security issue was to turn off SMT or any form of hyperthreading and predictive options in the bios. And this was hated by basically all back then, because it meant taking a whopping hit to the performance. All to avoid a security flaw.
So of course if AMD is still using SMT and other predictives, they are going to have more performance than Intel's cpu's which do not have it available.
Honestly, it's one of the most 'well F'ing DUH!" kind of moments in tech right now.
The only real neat thing about it, is the lack of power consumption increase. That power consumption barely budging is really nice, because otherwise the tech does normally increase power consumption a bit more than that. Not a lot, but more. Enough that turning it off is a legitimate way to reduce temps on a hot day where the extra threads are just not required. (Like playing an older game for example that is gonna be just fine with 8 full cores, cause it was programmed when 2 cores was the norm.)
To be honest.. I'm not excited about next gen AT ALL. I'm burnt out on trying to keep up with $2000 graphic cards. I'm at a 7900xtx liquid devil and perfectly happy with it and will be sticking with it
I've just decided now that even if AMD put AI into their notebook CPU naming scheme, that doesn't mean you have to say it. I'm just going to say "Ryzen 9 HX370" because it literally makes no difference.
after the 285 will be the I 86 also known as the 80286
Just got the 4090 and ill shop around when its 8090 time.😂😂
Why are we comparing AMD's GPU to Intel's? Intel barely started their discreet GPU line, and AMD has been on the GPU market for decades. Intel will be irrelevant in the GPU space for at least another 5 years. Compare AMD to NVDIA instead.
If specs would be like this, even rtx 4000 owners will upgrade!
RTX TITAN AI 32GB GDDR7 1799$
RTX 5090 28GB GDDR7 1399$
RTX 5080 24GB GDDR7 1099$ = RTX 4090
RTX 5070T 20GB GDDR7 799$ = RTX 4080S
RTX 5070 16GB GDDR7 599$ = RTX 4070TS
RTX 5060 12GB GDDR7 399$ = RTX 4070
RTX 5050 8GB GDDR7 299$ = RTX 4060T
Intel's nonsense is going to make 2025 a cake walk for AMD. If they introduce Zen 5c into desktop Ryzen, it's game over.
It was awhile ago that i saw it so cant remember exactly what channel it was on. But some time ago i saw a video about why intel where removing SMT. in it they said the extra space taken up, and the increased architecture complexity of the cores to support SMT was not worth it when they could allocate that space to AI processing or additional cores.
What i took from it was that they were planning to skip SMT in favour of replacing that space with additional compute power, for example a 12 core with SMT may allow a 16 full core instead with the same die size.
I'm sure people can google around and find the video, may have been gamers nexus or something but kind of makes sense, so lets see what they release in the future.
I've had AMD for awhile now (AMD and Apple, just got rid of my last intel system a few weeks ago) but only because they had the more suitable products for me at the time. I'm not going to hold having a faulty product against intel as it can happen to any company but how they deal with it in the long run will impact how likely i am to buy intel in the future.
"Confirmed" in Gamer Meld speak means "somebody wrote it in an Amazon product page" 😅🤣
How does caractéristiques of PSU can be a confirmation of 5090?
Pretty sure a lot of companies get information about upcoming products so they can have compatible hardware ready when these products are released.
because the hardware and software companies get to know before the public so they can be ready
5090 was confirmed like a year ago 🤦🏽♂️
it's not just the gpu itself it's the associated components to deploy together make it awesome for pc diy
Given preformance and recent intel problems, we will all remember this time, the point in history where AMD officially overtook Intel in the CPU race.
The first time for THAT was a couple decades or so ago, the Thunderbird-series vs the then-current Pentium III.
Thunderbird was slightly MORE efficient IPC, clocked up to 1.4 Ghz where Pentium III a couple years later EVENTUALLY maxed out at 1.33, and other such issues.
That is why Intel went to a completely different arch for the Pentium IV, that was lower IPC but could clock a TON higher and eventually passed the Athlon series on actual performance.
this guy always sounds like he just took a deep breath before talking
NEWS JUST IN : CPU THATS RELEASED 1 YEAR AFTER LAST CPU FROM INTEL IS BETTER.
Yeah no shit.
Unfortunately, the Ryzen AI 9 HX is only for laptops.
As I have been saying, it is the motherboards that is doing this. The MB default is NOT the processor default. They are all overclocking the processors out of the box. Puget have been setting more conservative settings and when they did thst, failure rates dropped, even lower than AMD systems in fact 😂
Hyperthreading is more work than it is worth.
I am pro small cores, but just a handful of the them for low power situations in a laptop. Also, it might make sense in server situations where you want to pack as many as possible under power and space constraints, but, for normal PC or high performance environments, there is really no need for them, especially this lobsided 2x8 PxE configuration. New laptops feel much slower than older ones. We are going backwards!
Sees something in an Amazon product description. And immediately claims that it is a confirmation... Your RUclipsr really tries to create a headline out of every little piece of information! Get a real job!
So sad to see what is going with with Intel lately in the media.... but good to see Puget reporting good things. Hopefully Intel makes it through this "ok".
lately? this has been going on since at least 2016, these are just the results
Sick. Could you do a comparison with Ryzen 4070?
I am not sure if we could trust the data from Puget systems cause Wendell has found that Datacenters using 13th and 14th Gen CPU's were also failing and those are using motherboards from Supermicro if there is one advantage that Datacenter focused motherboards has over Retail is that they are focused on Stability so they are not pushing Processors as hard nor are they pushing the voltages too high either so if those are experiencing failures then retails is even more screwed.
Intel isnt just abandoning smt, i've heard that they're using a new technology that optimize the scheduling of each core
Something about Rentable Units, if I remember the term correctly. I still don't quite know what that means as I haven't looked into it yet.
Remains to be seen how good this works.
Of course AMD's iGPU, based on their refreshed RDNA 3 architecture - an architecture originally designed to compete with Nvidia's 40-series cards - will outperform Intel's iGPU based on fucking Alchemist - an architecture that barely competes with Nvidia's RTX 3060 on it's highest tier card. Intel being this "close" to AMD in performance on their iGPU is more of a win for Intel than it is for AMD
My last new intel product was Celeron 333Mhz and Nvidia Geforce 7000 series GPU. If i would buy Intel or Nvidia it'll be as used ones. Atm im at i7 980x, Gtx 980 bought used.
I do enjoy this new wave of low power high performance CPUs, make them cheaper!!
I can’t wait for the 5090!!!
Why tho only you can afford it.
@@avoxxtea enlighten us mr time traveler. How much does it cost??
@@Haktic probably 1,000 to 1,200 like the last 2
@@avoxxtea 1,000 to 1,200 ?????? Someone else who comments and doesn't know what they're talking about You probably won't even be able to buy a 5080 for that
@@P5YKOTIK1870 youre slow af LOL
Pointless making AMD to intel comparisons. Only people who will buy intel are those that have no clue about intels shuddy QC.
Where are the lawsuits for RTX 4080 & 4090 burning connector? 🤣🤣🤣🤣
they dont sell enough of those and the few people who bougth them are busy doing rma to make lawsuits
Ambulance chasing law firms looking to make a quick buck, yeah, they are suing on behalf of people that got bad CPUs, but like most class action law suits, the law firm gets most of the money.
Battle mage will be killed before it cam be released
cant wait for new nvidia gpu hope 5090 will not cost too much, lets hope intel will recover amd fanboys make me rly hate amd.
Ryzen 4070 better then 5090
Yes but it needs PCIE-AI
Intel Error Lake?😀
How about a bit on the Logitech subscription mouse?
What? So amd has more failure rate than 14gen? Meanwhile yesterday I ordered an am5 motherboard, don't know what to believe anymore
-If we combine 13th and 14th gen, they're about the same as AMD Zen 4- AMD Zen 4 has higher defect rates ... on Puget systems. Which override the BIOS settings to more sane defaults.
In other words it is again a testament that 13th and 14th gen aren't that bad, if configured correctly. But nobody knew that unlike gen 1st to 12th, that now you HAVE to manually configure the protections and voltages and voltage limits and load line calibrations and so on, in order to have a safe, stable CPU (which might also have good performance and much saner power draws - like 250W in full load - and temperatures). It is both the motherboard manufacturers and Intel's fault here, but given the history, it's basically mostly Intel's fault here for giving such vague guidelines.
Dont trust puget they are the only website talking about am5 'failures' this userbenchmark level of diagnosis.
@Winnetou17 Those are percentages. The 13900K and 14900K are both 2%. If you combine them it is still 2%.
@@Tugela60 You're right!
I can't wait for 5090, I'm tired of driver timeouts with my xtx, I probably won't buy another amd gpu again. My 6950xt was pretty good but my 7900xtx has nothing but problems
Sounds like you have a bad card, I haven't heard anyone else complaining about drivers on the 7xxx series AMD gpu's.
What brand of xtx did you get? my XFX 7900xtx has been solid since it launched. You may have just gotten a bad card or maybe didn't wipe the old drivers from a previous GPU. I never get any driver timeout errors.
@@matilija and you don't hear more people talking about their intel card being bad too 😆
There is a reason the adoption of 7000 series AMD GPU is not as good as 6000 series
Because of how Solid 6000 series is and probably better aged driver support
waiting for battlemage performance and amd 8000 ,Nvidia is dead for me .
Are they seeling more i5's than i9's? remember that higher the power, more issues it has.WHY COLLUMN IS MERGED FOR ALL CPU'S?
if they sell LP laptops then it can hide all of it. 10000 laptops 1137g711 whatever ones that have 1% failure rate AND 500 i9's that have 500 returns, would still show 1.05% failure rate.
remeber current claims are that i9's fail at about 50% per year rate, i7 & i5 at between 10 and 30% and it mostly K cpu's.
something like i7-13700T afaik, have 0 failure rate, from this problem. its BC that cpu is so low power, it would take years to develop that same degradation level.
ESPECIALLY shop caught, is unrelated to the problem. 1% damage is most likely shipping damage, stuff that got dropped. We can remove that from all of them.. most of those is either manufacturing defect or shipping damage, and it just shows they are selling way more CHEAP board of AMD, than intel's ones. (especially cheaper ones have worse boxes, so they break more often)
7:40 this is clearly that they just dont get how that damage happens.
it needs TIME. CPU DEGRADES OVER TIME. so now, 2 month from sale, they see 0.5%, then 1Y in they will see 15%, then 2Y in they will see 35%.
is their warranty covering that long?
this graph clearly shows they either dont understand the problem, OR they are trying to hide it.
So while that does it for today which GPU are you most excited about with next gen.... HOWBOUT A GPU THAT DOESN'T COST AS MUCH AS A NICE CAR!!!!!!
Intel's decision to develop and use their own GPU proved to be their own undooing. They would have been in a much better position if they focused on CPU and let the third-party laptop manufacturers use Nvidia GeForce. Now Intel is 3rd place at everything. Apple, too, should have continued using X86 instead of developing their ARM processors for laptop. Many of my friends boughtMac Proo because it had fast X86 processo and GeForce graphics card. You could play Windows games on it. With ARM, games now have to be emulated. Als, ARM doesn't come with a fast GPU. Sad, but that's how companies rise and fall, all because of some poor decisions by thei executive. Tim Cook doesnt have the vision like Steve Jobs. Neither does Pat Gelsinger.
1:44 the ryzen 9 7950x3d has Problems with utelysing the x3d Cache. If you Do not game It trys to Use the x3d Cache. The Performance could be hold back because of this.
how many times are you gonna say 5090 confirmed lol
this is a slow news week, so
I'm still buying Ryzen 4070
5090 or titan ai
Amd is beating a dead horse //Intel 14900KS owner/ sucker
If I can give a little bit of constructive criticism about your moderation:
Could you vary your use of the word 'but' a bit more? You use that word a little excessively sometimes.
I mean, maybe you consider that a signature style of yours, but as we say here in Germany, 'that has a beard already' - which means it has lost its appeal, is getting old.
In the end it's your decision on how you do things, as it's your channel. I just wanted to share my perception as a viewer - and most creators usually want some feedback.
Was about to unsubscribe for yet another click bait 5090 video and realised I’d already unsubscribed! It hardly gets a mention at the end of the vid and it’s not even worth mentioning! Sorry for being negative but that’s enough for me!
Intel's c-suite management level has to make change immediately, they need a phenomenon leader like Lisa. The overall cpu architecture need a redesign, what intel experiencing is still better than AMD's dark age. Intel needs its own Ryzen, if intel keeps patching current architecture just to save money on development............we might gonna lose another blue giant like power pc.....