Wait for Zen 3

that will probably be ryzen 4000

ryzen 1000 had awful issues at launch too but eventually stabilized, but ryzen 2000 was pretty solid overall save for some memory issues.

>Zen3 will be 7nm EUV
>EUV inclusion is basically just contacts/vias on a couple critical layers
>does come with improved etch resolution though so clock binning should be more consistent, known good candidates having less variation
>5nm FinFET EUV has drastically more EUV inclusion
>insane increase in xtor density, 30%! power reduction, up to 10-20% perf uplift
>Zen4 will either be 2nd gen 7nm EUV or will be 5nm EUV, still up in the air
>Zen5 at least would be taped out on 5nm

Constant arch and process node upgrades for the near term future. If AMD can manage at least a 10% IPC uplift with each iteration of the Zen core it'll be pretty incredible.

This

I went from haswell to zen 2, I think Ill wait for zen 5 at the soonest for my next upgrade, but more realistically im gonna be on this 3600 for the next 6 years

There's no point in waiting for PC hardware these days. It's not like the 90s where CPU power would double every 18 months.
Buy whatever's out now, and you'll be good for the next 4 years.

STFU you dumb ass nigger bitch me and my haxxor army are literally DDOSing all of your family members fb accounts rn when I'm finished your parents will lose their jobs and you will have to get out of their basement and suck smelly jew cocks for money to provide food for them. Shouldn't have started a meme war with me in the first place.

Attached: maxresdefault (1).jpg (1280x720, 187K)

Post yfw after 5nm FinFET we're making the jump to GAA transistors.
Pic related. Its my face.

Attached: cntwrapgatetransistors.jpg (500x398, 30K)

>2021
>intel is just finally releasing 10nm high perf desktop chips
>probably paper launch initially, supply will be extremely limited like OG desktop Broadwell
>AMD will probably be releasing a 5nm chip
Damn. Brutal MOG.

ye, lets wait

Attached: 1209180092710.jpg (480x480, 20K)

Is this bait? Cause that's exactly the opposite of the 90s. You had to upgrade every year and a half unless you wanted to get surpassed. If you sat on your hands and waited odds are you werent even gonna be close to able to run new software.

These days a 10 year old 2500k is still a competetive cpu (if overclocked), unless you need to upgrade you should always be wating for something substantial to come around. Which took a whole 8 years for my 2600k to not be worthily surpassed by a 3700x

how viable is it for large scale?
I dont give a fuck about lab samples or some experimental IBM nazi magic CPU that is never going to see the light of day

Damn, I think this was supposed to be a joke but it's cringey as fuck

Attached: 1561957769493.gif (400x250, 1.98M)

Well both TSMC and Samsung are using GAAs at 3nm, so pretty damn viable.

extremetech.com/computing/291507-samsung-unveils-3nm-gate-all-around-design-tools
community.cadence.com/cadence_blogs_8/b/breakfast-bytes/posts/tsmc1
community.cadence.com/cadence_blogs_8/b/breakfast-bytes/posts/iedm-2018-pv
anandtech.com/show/14666/tsmc-3nm-euv-development-progress-going-well-early-customers-engaged
anandtech.com/show/14333/samsung-announces-3nm-gaa-mbcfet-pdk-version-01

We in the GAA future now

GAAs will allow power to fall enough to outpace increasing thermal density as xtor density increases.

Attached: rcj_5nm-Nanosheets_1_1496432257.png (1508x1008, 1.39M)

Aight, goat
But I really need to upgrade my 1700 witch has that hw defect, which crashes my Linux machine when leaving it idle for some time
So, I will upgrade to zen+ now and wait for zen5 to upgrade to zen3, so I can wait for zen6
Is this good plan?

Attached: thinkpepe.jpg (1000x800, 85K)

Attached: 1466148204302.jpg (600x760, 69K)

Waiting for 5nm APUs with 20+ CUs