Can we buy intel cpu now? Is it finally good?
Now you get what you can.
>>108436372Not buying intel after having suffered with that piece of shit raptor kek that kamikazes itself into corrupting itself in the middle of executing code after merely a week of vidya gaming
>>108436440Ryzen 9000 does the same thing though
>>108436440by the way these fucks lied about it not affecting laptops equipped with these awful processorsthey absolutely do degrade and still pollute bug trackers to this day, if you try to catch the corruption in a debugger or crash report you always get WTF impossible shit like the cpu reading an XMM float register as a pointer in a ucomiss instruction that doesn't use memory operands
I bought intel stock at $25
>>108436446proof? i have a 9800X3D without ANUS motherboard and the cpu has survived many times the lifespan of the average 14900KEK/14900KYS.
>>108436372how fat is this guy?
>>108436372remember guys, every 25 core ultra 5s sold means israel gets to defend itself from 2-3 muzzlim missiles
>>108436372my 9800x3d is better
>>108436372post benchmarks
>>108436372Intel finally listened to /g/.>2500k>Let me guess, you need more?Introducing the 250k
>>108436535I dont care about niggers killing themselves over sky daddy, im playing video games
>>108436722LOL NO
How much will the ram cost for such a system?
>>108436372>1 generation socketdealbreaker. Kiketel needs to learn from AMD and have scokets that last at least 3 generations
>>108436928just use 8200mhz cudimms to get this level of performance>no I won't show data for slower memory
>>108436931AMD would have a socket a year if they could get away with it. They couldn't with ryzen 1 as otherwise they would be dead in the water and because of that people now DEMAND multi-generation socket support, which AMD and board partners hate.
>>108436966>DEMANDThe main reason I'm buying AMD over intel is because I can update the CPU in the future while keeping the other components
it loses the value proposition if you have to use ddr5 with itat that point it's a dead socket and it's going to be competing with zen6 and its 12 core CCDs
i mean it's nice that we're getting competition in the CPU space but there's no way in hell i'm gonna build a PC right now with current RAM and storage prices
>>108436899>AMD tested on 6000 ram>Intel tested on 8200 and 7200 ramhmm
What i'm learning from reviews is that for vidya the winning move was to buy a 5800x3d 3 years ago.
Intel is back baby!
>>108437053>cinebench>caring about rampick one and only one
Okay but can I plug these new CPUs on my AM4
Uh..
>4nm on a socket that has 2 more generations left vs 3nm on a dead socketthis is a pathetic showing holy fuck, it should be way more efficientzen6 is going to smoke this
>>108436372>>108436899E-cores don't work on Windows 10 so it's dead on arrival>>108437095also it looks like a massive power hog
>>108437103>E-cores don't work on Windows 10 so it's dead on arrivalLuckily they do work on functional operating systems so that isn't a concern.
How much faster is it compared to the 14700f?
>>108437095>55% more power for 84% more performance le bad
>>108437367who cares about this shit? this isn't a laptop chip and i don't live in some off-grid house where i have to monitor how much solar power i have available.
>>108437386>who cares>>108437095
>>108437392i hate the tech reviewer fascination with this stat. gamers nexus spends way too much time on it.
>>108436502Not at all desu, I'm a fatso myself and I never thought he was fat looking.
>>108436372They need to release those increased L3 cache CPUs already like they promised so they can make a comeback.
This is good. Looks like Intel is finally putting some pressure on AMD again, who now also have the budget from their extremely successful last few generations of CPUs to start innovating again, and will have to do so if they don't want Intel to eat their lunch again.If only we could have a similar situation in the GPU space, and have reasonably priced RAM and storage again.
>>108436372If you bought a CPU in 2019 onwards there's literally no reason to upgrade ever again.
>>108437367AMD housefires seems to be a universal constant no matter how much time passes.
>>108437135>functional operating systems Lol cuckjeet marketing the intelaviv slop as well as microslop
>>108437367my 9700x is sitting at 16W now, intelaviv cpus still use at least 200W under load, I won't believe your jeet lies lel
What good is a processor... when you're unable to RAM?
>>108438608who said anything about w11?
>>108437103>E-cores don't work on Windows 10 so it's dead on arrivalThey literally do unless you're running some ancient version of windows 10 that's like 1909/1803 level shit. Just install the intel chipset drivers and roll, they just work far more effective on windows 11 for <.8ghz run modes
>>108437367as a euro i care about this, but whole system power is almost the same once you toggle all powersaving options is likely 60w vs 70w in idle. meanwhile the heat output is a factor in summer under load. i like the multicore performance of these, but dead platform, while am5 gets zen 6, and it needs faster ram that is unaffordable. i just grabbed a 6000 cl 38 kit at €350 because that's considered a good deal here, faster ram would be €100 more at least
>>108436966>which AMD and board partners hateDo they hate bankruptcy less? Adapt or die
>>108436890kikes ruined your video games faggot, stop this glowshit tel aviv needs to go
>>10843866532gb of ddr5 is sub 400 bucks. It's not like it's 4000.
>>108436372When was the last time there was a good Intel CPU?Difficulty mode: you must disable all features that make your CPU insecure.
>>108437419he's still fat. you're just fatter than him.
>>108439665265k
>>108439651>32gbWhat is this, the 1920s?
>>108439682i gots me a 100 gigabytes of ram
>>108439682Tell me exactly how you're gonna use that much? I've got 32 and never used more than 16, ever.
>>108437053AMD spec is 5600, intel spec is 7200
i'm buying an am5 mobo right now and i'll be buying zen 6 early 2027. hope ddr5 is cheaper by then
>>108439699What if I edit video of me dancing or host a home server.
>>108439744I'm sure you can use your older pc as a home server
>>108439724Why buy the motherboard now, there are always better stuff coming out, almost all am5 motherboards have coil whine Buy a psu and a case or something instead, as for ram, I think you're out of luck it'll go higher for sure
>>108437053That's because AMD CPUs don't even support 8200, LOL
>>108437103>E-cores don't work on Windows 10Yes they do.
>>108436372Lads, I have a 13900K since new.Never updated any bios stuff.Just set the TDP limits to 90w in bios when I got it.I render a video once a week and otherwise make small 3d models for car parts.Is that enough to avoid the crazy 1.7V spikes people were seeing?The max load temps are under 60C if that matters
>>108437062>>108436899Why are you acting like Cinebench is relevant for anything? Go get a 9950X3D if you want to crush Cinebench scores.
>>108437098Does it matter? You're not going to be upgrading because DDR5 isn't going to be affordable until we've had DDR6 for years.
>>108437053Why the fuck can't we test CPUs with the same RAM? If it supports faster memory then do a separate test for comparison.
>>108436372Imagine believing that a 270k is better than a 285k or a 9950x
>270k is better than 285kNice naming Intel
>>108441926cool it with those remarks
>>108436372>186 wattsvs>98.7watts of a 9800x3d
>>108442129BITCHIN 3D
>>108442129Yeah that will matter after a decade of 100% 24/7 use given the cost difference in those chips.
>>108442217I want to see a cost comparison with inflation reflected in that stat sheet. Besides most people upgrade every 4 years but i upgrade every 8 years.
>>108442217if this was true then everyone would have continued to buy intel my goy but the reality is that when you offer inferior perfomance for more than double its wattage nobody on their right mind will even think of buying it
>>108442217It doesn't matter since Eypc is owning the server market. AKA bread and butter for intel. Laptop market is going to be taken over by ARM eventually and Apple is already winning in that regard due to its reduced power consumption.
>>108442736>if this was true then everyone would have continued to buy intel my goy Everyone did - the whole "AMD is taking over intel" isn't backed by reality. It only comes from the DIY market which is a tiny portion of the whole x86 cpu space.
>>108442763yes they continue buying intel in your reality however in our reality the trend isnt going the way you think its going
>>108442763No, it comes from everyone in the industry. Cloudflare literally bought a whole load of new Epyc servers for their infrastructure. Why? Well, it's quite simple why, AMD is simply better for them.They are growing their data centre market share extremely rapidly too, it is very profitable for AMD and they keep increasing every day.>>108442771 is right, Intel may still have the majority share but it's a downward trend.
>>108442771>steam hardware survey>accurateni hao +30% in a single month
>>108436464What did you sell it at
>>108437135I didn't know the linux kernel supported e-cores
>>108439699I have 64 and Firefox uses up to 50gb of it
>>108438737I was still running win10 1604 until cyberpunk 2077 came out
>>108436372Never forget! 6 million lives were lost during hamaz attacks! Support the oppressed God's chosen people goy
>>108443203porn addict?
>>108436899honestly amd can go fuck itself for sitting on 8 cores for a decade
Are ARC linux drivers still shit?
>>108443721Except they haven't.
>>108436372My 265k runs fine
>>108441864Try setting a thermal limit instead and see how it performs.
>>108443143>40 million pool isnt accurate
>>108443852So you be saying 10 million chinese people all signed up to steam in a single month?
>>108443172$5
>>108443869its ok bro do not cope harder you might get aneurysm
>>108443869chinese new year
>>108443915So this happening repeatedly and valve correcting it doesn't count? Fair enough.
>>108441864There's a bios setting in nearly all bioses to cap VID requests to 1.40v. I have done that with my i9 13900k and zero issues for 6 months of daily use
>>108439651>tfw 32GB DDR4 was going for $50 just a year ago
>>108444460This, I remember paying like 90 bucks for mine back when DDR5 was just barely a thing. "Only" $400 for 32GB is crazy talk.
>>108442736Meanwhile Ryzen idles at 40w and Intel at 10 lmao
>>108436372Modern intel things>200$ cpu>cheapest motherboard is 250$
>>108445355are you an LLM with out of date pricing information?
>>108437367Idle power is really nice for server use if you live in a country with expensive electricity.I read about a guy who got his whole server down to 6W idle, as in a full PC build, not a Raspberry Pi or other SoC. Intel chips are good for that and their video transcoding if you do media on your server.You get into the weeds a bit with C state compatibility, motherboards, and which components are on the CPU rail and which are on chipset, but you can take it pretty far with good planning.
>>108436372Nice try but too late.
>>108445591problem there is that your whole server needs to be low power even in full load then, as PSU works worst at lowest end and highest end. So you want as small PSU as you can get that still manages to do load conditions with any kind of efficiency, otherwise it doesn't matter at all how low power parts using power are. So in short, intel CPU idling at X single digit watts in PC with 800W PSU doesn't mean fuck all as PSU is wasting power at idle there.
>>108436372DDR4 or fuck off.
>>108445778Yeah, it's definitely a full-server design you have to work on, and you can destroy all your gains easily with some simple-to-make mistakes and kind of arcane motherboard settings.PSU efficiency does get worse outside of the mid-range, but at Platinum rating the efficiency curve is manageable, flatter than you might expect. That 6W server I mentioned is measured at the wall, so it takes that inefficiency into account. If I recall correctly it was a 350W PSU which gave enough headroom for everything at TDP plus headroom for start up jumps.The iGPU in most Intel CPUs can help with total TDP too because you don't need a comparatively more power hungry dGPU, but you can still decode video at higher resolutions and more simultaneous streams with quick sync and it's lower power than an equivalent act on a dGPU too. Only really works if video encode and decode is what you're after, because it is specialised for that, but many server users are after that.
>>108443632No, YouTube memory leaks to punish ad block users
>>108447107>YouTube memory leaks to punish ad block usersis that a real thing? always wondered why having a lot of youtube video tabs open lagged firefox
>>108447107Youtube takes no more than 1GB of ram per tab on my Firefox (which is still ridiculous). Why do you people not close your tabs?
RealisticallyHow long will my 13600k be relevant
>>108437386>who cares about this shit?Me, but only because I stuffed picrel into a 4.5L ITX chassis and thermals are actually a concern
>>108437367>Boughted a 265k last year for my birthday>Deathly afraid of using it because intel power draw>It's at the bottom of the chartsCould've fooled me, I emerged @world once and it pinned to 105C almost instantly. Here I thought I needed to go out and fall for the watercooling meme.
>>108447748intel power draw this gen is quite high in extremely multithreaded scenarios (gaming is not one) because of the fuckton of cores. the same video as that source has the 250k/270k pulling 156/220 watts average in cinebench.but that's commensurate with its better performance, actual efficiency is about the same as most other chips.same way the 9800x3d uses more power in games but it's also a better performer so eh. intel's idle consumption is nice though.
AM4 Chad checking in. 5950X with 64GB of RAM
>>108447748>intel power drawwhen will you cat niggers learn that power draw is related to process node and factory tuning and not A or B brand
>>108448694Same. Still going strong and no intention of upgrading it any time soon.
>>108436372I will buy. number go up, simple as.
>>108441864Just update your BIOS you fucking retard.
well the price is actually justified now
>>108436446It does not, you retarded brand warrior.
>>108436372I don't like e cores and I don't want to buy a desktop CPU with them.
>>108448745And who decides that? Intel and AMD respectively. But architecture design alone can make a significant difference as well. Intel deserves the reputation of being hot and inefficient in the last 10 years:>Skimping on cores (more cores is more efficient)>Cheap tim under heatspreader while AMDs was soldered (5-10C hotter)>Focusing on high single core clock while AMD focused on IPC and cache>inferior intel process (remember 14nm++++) vs. industry leading TSMC process
>>108436372>still no large l3 cachewake me up then. something pretty easy to do to take back enthusiast marketshare and they still havent done it. just swap out the cuck cores for space.
>>108449543>>still no large l3 cacheWhat are the benefits of that?
>>108449543RAM chiplets don't grow on trees>>108449554Lower latency and increased FPS in CPU bound scenarios due to increased cache locality. It doesn't always matter but it matters a lot when it does matter.
>>108449137>cheap timTim Apple?
>>108436372Maybe you should go buy some useful cpu from Nvidia or AppleNopeand NOPE!
>>108447852All their desktop chips have been produced with shit processes starting from 10nm down to Intel 7. The saving grace with Arrow Lake was finally fixing one side of the equation with TSMC N3 nodes. However, the architecture was only about okay, with Lion Cove, the P core, being mediocre as fuck. Skymont, the E core, is great. Nova Lake needs to be better but 18A and only a minor E core revision and supposedly the last iteration of the P core doesn't strike me as confident. The only saving grace is that they'll have a last level caching solution that matches X3D and has better integration technology with Foveros 3D being better than AMD's TSMC used COWOS with it being 2.5 and 3D instead of full 3D. Will be interesting to see where things go from here, next gen will be very interesting.
>r5-5600X + 6700XT - desktop>r7-5700G + 3060 + 3090 - AI server>r5-2600 headless - NAScomfy
>>108436372intel was always good
>>108454018If by good you mean antitrust behaviour then stagnation followed by Spectre/Meltdown and finally silicone oxidation/degradation then you'd be correct
>>108447293>>108447107nvm just had a Youtube memory leak today. The memory use suddenly increased very rapidly and firefox froze.
>>108451482>r7-5700G + 3060 + 3090 - AI serverInteresting use of a 5700G there. I guess you're using it for video output and running the two GPUs headlessly?The 5700G is a great CPU. I have one in my OpenWRT x86 router I built because its 8 threads mean I can have a 1:1 mapping between network queues and the CPU.
I dont see the point. They have a new generation coming out at the end of the year on an entirely new socket.
>>108457296Its for the 2 people that bought Intel by mistake this generation and don't mind a nice uplift.
>>108437367i used to care about that. but even if your cpu only uses 3W, the entire system with a dgpu uses 100w idle at the socket. difference between 100 and 120w idle is not really worth a thing.
>>108457285>I guess you're using it for video output and running the two GPUs headlessly?That is correct, although I don't really use the video output (yet), the main reason is that it is the highest core count CPU I own and 4 cores are recommended per GPU as far as I've read.In the past I used the 5700G's video output for my own usage and having the RTX's headless. But figured it's best to have the AI server dedicated.
>>108436372Intel ME
>>108437053AMD isnt stable ar higher speedsThey all have shit memory controllersHoping it gets fixed in zen 6
>>108458980>amd pspwhat's your point?
>>108436966>people now DEMAND multi-generation socket supportPeople DEMAND multi-generation sockets from intel, good GPUs from AMD, reasonable GPU prices from Nvidia, appropriate GPU prices from AMD (not just "nvidia - $50"), and lots more. Generally this doesn't affect what actually gets made and sold.If AMD really wanted to refresh their sockets often they absolutely could.
>>108436372No
>>108448745Then why does intelaviv run so much hotter while shitting the bed in performance?
>>108436899>the 7800x3d is still being used on 2026 benchmarksapril 2023 was forever ago in internet time
>>108436372I'm buying their next-gen desktop cpu no matter what as amd platforms (am, sp, tr, embedded, it does not matter) have been dogshit since day forever and "we're a datacenter company now" so fuck amd.>hurr work on my machineno, it absolutely fucking does not, you just don't realize it because you're a fucking gamer, you only care about raw fps numbers in light tasks (yes, games are pretty light tasks btw) and you don't know shit about computers.everything starts falling apart once you add hardware and do long-running heavy stuff, it's also true on their epyc/tr counter-parts and I've had to diagnose nasty shit because of spurious non-sense that randomly happen on these $800 fucking boards with $10k cpu.am4/5 have latency issues, memory controller issues, bios issues, low-power state issues... Remember the usb fiasco from am4, the faulty cpu that segfault under load?certain epyc fucing crash after being on for a certain period of time, how is this acceptable for a "datacenter" company?what about fTPM stutters?WHEA error 18 on zen3 that crashes cpu on idle?FUCK. THEM. all of these things should happen in final products.cpu perfs are solid, the platforms are utter dogshit, I'm not falling for this jeet shit ever again, atleast intel aviv board partners (the same ones as amd ironicaly, just only one of the two care about quality, btw, not saying intel boards have 0 issues) have to release working fucking board unlike the forever number two sloppy rival.in the mean time I've tested the new whatever rapids cpu from intel and while each one needs its own dedicated nuclear power plant to run, it has been rock solid with no weird behavior especially for taks that run for months and io does not produce latency spikes.t.hpc engineer
>>108436899>amd fanboys tasting their own cinebench medicinetop kek
>>108462128All 6-10 core Skylake derivatives throws ring parity errors in heavy multithreaded programs so Intel ain't faultless either
>>108436440I have a 13900KF and it's still running well. My friend got a 9800x3d a year later and it just broke. I guess it's luck of the draw.
