[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
>> No.76142625

zen 3 launch september 1st 2020. check em

>> No.76142637

AM5 DDR5 5nm May 5th at 5:55:55

>> No.76142638
File: 120 KB, 268x212, felipe.png [View same] [iqdb] [saucenao] [google] [report]

>G-g-guys our next processor generation will be good! Promise

Literally intel the past like 5 gens

Hard doubt

>> No.76142641

I mean, they already are ahead, they'll just be even further ahead.

>> No.76142677

>Meanwhile Intel's 7nm (equivalent to TSMC 5nm) will be ready in 2021
Good joke

>> No.76142767

>AMD’s 5nm delayed to 2021
>Intel’s 7nm ahead of schedule ready for 2021

Haha AMD B T F O

>> No.76143000

No, they've been doing that for about 9 years.

>> No.76143004

Whether the 5nm rumor is true fully depends on how compatible the 5nm is to 7nm since generally you can't just quickly port from one node to another and have it work.

>> No.76143063

Geez the rumor mill has been all over the place this week. So is this going to be for Renoir? If so then Vermeer should still be released for the desktop in Q3 on 7nm. Either way is good news, either we get a new release faster which could prompt price cuts by Intel for the holidays or only a brief delay for better performing parts.

>> No.76143073

>Meanwhile Intel's 7nm (equivalent to TSMC 5nm) will be ready in 2021
But they still don't have even 10nm.

>> No.76143269

They should just go back to the Netburst idea and make a single core that runs at 10+GHz but has 4 or 8 way SMT to run multiple threads as needed

>> No.76143280

>theyve been resting on their laurels for 18 months?
well yes that exactly what they did since skylel dropped and never get their asses up.
where the fuck have YOU been, anon?

>> No.76143436

Yeah, but Intel isn't even to 10nm yet. Plus I think AMD staggering there releases will help sell all of the old stock.

>> No.76143438

Mobile is THE big market. Large volume. Servers are "just" large margins.
But since you need that many chips for mobile it's quite unlikely that amd will start with that.

>> No.76144716

>But since you need that many chips for mobile it's quite unlikely that amd will start with that.
i mean that depends on the branding. throw out something like HS variants out, so super high end laptop market first, high price - low volume as a pipe cleaner for TSMCs new node.
that will also give AMD some mindshare and presents in this price bracket. hopefully with 'good enough' rDNA2 gpus right next to it if Intel pays OEMs again to gimp on graphics

>> No.76144802


>> No.76144965
File: 91 KB, 818x693, 1541892583390.jpg [View same] [iqdb] [saucenao] [google] [report]

well, hypothetically speaking: Huawei is on their way out, thanks to the US, and TSMC is probably losing a lot of money if they dont have anyone to jump on it. bet they would very much welcome AMD to take the place and gladly work with them to have a working chip once 5nm+ commence high volume production in ~Q4.

but yeah, thats literally armchair analyst take with a lot hope.
its way more likely, as you said, that AMD have Zen3 taped out and working in their labs, ready to be released.
still, i cant let this rumor die. the timelines are matching just too fucking good for this to happen. keep an eye out for any Huawei/HiSilicon news anon

also see>>76144716 it wouldnt surprise me when AMD is going for next mobile chip ASAp

>> No.76145054

I don't think they'll use 7nm for desktop processors on 2021.

>> No.76145124
File: 1.07 MB, 757x2227, file.png [View same] [iqdb] [saucenao] [google] [report]

>Huawei is absolutely huge, I highly doubt that the US sanctions had a big enough impact on how many wafers they order from TSMC
anon, didnt you read the news? i carefully phrased it that way because obviously i could be very wrong and huawei get their chips but i have reasonable doubt now that TSMC strenghend ties with the US and literally announced they will build fabs on american soil.
>inb4 muh verge is no sauce

>It would be cool to get 5nm+ next year, but it makes no business sense to essentially throw away a product line that offers >15% performance improvement over your last gen
i mean sure but as longer i think about it the more likely would the move to 5nm for MOBILE, a insanely huge market as other anons pointed out, be. why would Lisa Su straight up delect that rumor and talk about mobile?
>She stated that new nodes usually start with mobile processors as these are easier to manufacture
very suspect
furthermore, console APU's could very well be build on 7nm (with EUV) as would be rDNA2. meanwhile Renoir is running strong on a matured nonEUV 7nm node. Matisse refresh coming on that same matured node, so why would AMD launch a Zen2 refresh when Zen3 apparently was scheduled for Q3/maybe Q4?

like i said, i'm probably wrong and just connect dots with lots of wishful thinking

>> No.76145131

>Intel's 7nm (equivalent to TSMC 5nm) will be ready in 2021
What makes you think that?
It was supposed to be ready in 2016.

>> No.76145164

>Interestingly AMD seemed to have noticed who actually bought Radeon VIIs
Supposedly (i.e unconfirmed hearsay at the time) AMD was surprised by the demand for the VII even though they were only ever going to keep it as a low volume product.

>> No.76145252
File: 1.25 MB, 1002x1480, JUST_krzanich.png [View same] [iqdb] [saucenao] [google] [report]


>> No.76145554

>they leave a one year gap between Zen 2 XT chips and Zen 3 5nm+ chips when they could just release Zen 3 this year?
I think the logic would be to take a small delay to get a clear, notable lead against Intel (the sort of lead Intel cannot match no matter how high they clock their cpus) than releasing "on schedule" and securing a small lead. If (again, this is hypothetical) 5nm Zen 3 let AMD have a 4800x that beats the 10900k in gaming and annihilates it for productivity that is a massive, massive win and will go a long way to convincing people that AMD is number 1 - which is what they want.

>> No.76145588

AMD might go for a Zen3+ on 5nm so they can launch Zen 4 and the new socked when DDR5 is being sold to consumers. Because lauching Zen 4 with both DDR4 and DDR5 support is going to result in a lowered margin and headaches with bios and motherboards.

>> No.76146184

It's like you have never seen any Zen launches.
Zen 3 launches at late 2020, server parts first. Desktop Ryzen can launch at 2021, nothing surprising about that.

>> No.76147084
File: 1.21 MB, 500x500, 1518905604635.gif [View same] [iqdb] [saucenao] [google] [report]

yeah good luck surviving with those breadcrumbs

>> No.76147196
File: 55 KB, 500x669, 1506436461200.jpg [View same] [iqdb] [saucenao] [google] [report]

tech illiterate here
What's the significance of Taiwan's 5nm node

>> No.76147835

For both Zen and Zen2, desktop launched first.

>> No.76147883

So what jewish tricks will Intel pull in 2021 to keep the gaming crown on a shittier fab than AMD offers?
250W TDP midrange chips?

>> No.76148160

>oh well it was a good ride for AMD

you people said the same thing for the comet lake

how did it turned up ?

>> No.76148251
File: 32 KB, 491x404, 1590775979136.jpg [View same] [iqdb] [saucenao] [google] [report]

>So what jewish tricks will Intel pull in 2021
copypaste big.LITTLE for desktop

>> No.76151382

rip amd

>> No.76151608

So Intel wins again. I laugh at those pathetic AMDrones. Every single benchmark shows that Intel is unbeatable. 10900k is absolute king and literally NOTHING AMD offers can beat it. NOTHING. And they will never have anything capable of surpasing intel CPUs.
>B-but muh cores.

>> No.76151935

Are they still going to work on B450/X470 AM4 chipsets?

>> No.76151943

0 Watts

>> No.76153863

>TSMC is dedicating 100% capacity to Huawei for the next 120 days

>> No.76154036

because the claimed functional charateristics are similar, though whether it will ship like that is another story , as seen with 10nm, where power consumption skyrocketed at higher frequencies, yields poor, and density targets not being hit.

>> No.76154040

AMD is shrinking nodes too aggressively for CPU which doesn't seem to benefit from increased density (thermal constraint) or small package size (at least on desktops)
They are burning money for a payoff that isn't coming. Even laptop CPUs are no longer scaling with smaller node.

>> No.76154071

The us doesn't make anything anyway, so is that really surprising?

I nearly spit out my drink, when i realised that you can't even buy a motherboard that is made in north america, and there are NO american motherboard companies left other than supermicro and they outsource to taiwan.

>> No.76154303

congratulations you saved $3.36 this month!
>the CPU was in use 24/7

>> No.76154335

It's more important for mobile and server architectures. With mobile architectures, low TDP is important because laptop cooling solutions are designed for a specific TDP, and with servers density as well as lower TCO are important.

>> No.76154409

$3.36 is worth a lot less to a company than it is to an individual.
Node shrinks benefit laptops and mobile devices the most but I'm not seeing any AMD powered smartphones. Laptops without ventilation (i.e. a bad idea) are just about all I can picture having a need for a 5nm CPU vs 7nm.

>> No.76154434

5nm will enable more performance out of the same TDP. And not only did you completely ignore the part about density, but when you're talking about a rack full of 250W chips being replaced by a half rack of 200W chips that's a significant savings, both in initial investment and in TCO.

>> No.76154518

Probably a 55-core chip featuring SMT 5, PCIe 5.0, 5Gb/s ethernet, quintuple-channel RAM, and support for 5G telecommunication.

>> No.76154626

Bro you think companies are going into the CPU market empty handed or jumping at opportunity to make massive purchases of the hottest new chips cuz because they hit the market?
They already have infrastructure that allows them to wait indefinitely or until Intel responds. They aren't peeing their pants over the latest Xbox title and doing extra chores so they can get it ASAP.

>> No.76154636

TSMC is manufacturing for huawei 120 days of stockpiling, 700 million usd order

>> No.76154788

Do you like cars?

>> No.76154892
File: 61 KB, 521x339, im-fine.png [View same] [iqdb] [saucenao] [google] [report]

>10 cores consuming a lot more power than a 32 core AMD cpu

>> No.76154908
File: 57 KB, 636x576, Speccy2.png [View same] [iqdb] [saucenao] [google] [report]

Mein nigga

>> No.76154937

bro I love how my MSI BIOS is randomly in 1/2 in Chinese
nice setup I see you have AMD-V enabled, that's pretty cache $$$
Keep dreaming of those semis nerd

>> No.76154960

even mp3s
5nm node does it all

>> No.76154995
File: 451 B, 45x18, file.png [View same] [iqdb] [saucenao] [google] [report]

so many

>> No.76155020

OK retard

>> No.76155036

are you the "muh servers" guy still? persistent

>> No.76155043

Prove it nigger

>> No.76155078

Is that a Trueno?

>> No.76155086

*eurobeat* it's a Ryzen 4600XT the ae-86 of processors!

>> No.76155177

OK retard

>> No.76155180

OK anon

>> No.76155227

You just need a constant supply of liquid nitrogen to keep it stable.

>> No.76156087


where were you when intel is kill?

>> No.76156101

ITT telling people that mobile chips on 5nm drop first from AMD

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.