NEW RYZEN SHITPOST THREAD

It would be lovely except for the people AMD tricked into buying a 4c4t Ryzen 3 2200G for 100$ last year where they are selling a 6c12t APU in it's place now, yeah fuckin right they don't like you guys that much to destroy last gen Ryzen 5's value

wat

Attached: 1528269341610.jpg (466x591, 23K)

Hello regards of this rumor isn't true
Intel product superior in processing and computer apps for runnings
AMD product are untrustworthy and no real use in regards of professionalism sir
Quality only to be found in Intel only

>4c4t Ryzen 3 2200G
wtf I thought it had SMT

The Ryzen 2400G has 4c8t. It also has 10 Vega units compared to seven in the 2200G. But for the $150 a 2400G sells for, most people would rather get an R5-2600 and dirt cheap GPU for non gaming.

I can see the 3700X becoming a thing for AM4 top end, but I don't see 16c32t on AM4. It would cut into their Threadripper line far too much. As it is I have an R7-1700X and would gladly pay $400 for the 12c24t variant. Hell I'd pay that much for a non house fire, 5GHz capable 8c16t. Needless to say, I'm pumped for 7nm. Who knows? Maybe Intel will get their 10nm under way and actually make some sort of halfway viable product.

No one got tricked, it’s just progress
It’s the same shit that happened in the 90s but unlike the 90s your not immediately obsolete

TR isn't just about core count, it's dem PCIe lanes and quad-channel RAM. 1st gen TR had 8 cores for people who wanted the other goodies but didn't need moar coars

>implying 12/24 can run on X370 and most X470 boards
(lol

16/32 for dual channel RAM...and it won't run on any X370 and X470 board

That's true I know. But people will be feeling quite burnt if they spent $1000 on a 16c32t 2950X, $350 X399 motherboard, etc.
Says who? Cheap shit motherboards will be put of luck no doubt, but who is to say top end boards won't just get a BIOS update like with X370 and 12nm Ryzen? And what does core count have to do with dual channel RAM? Dual channel is standard.

TR is also about getting some of the best of the best in terms is silicon, only Eypc has better silicon
If most desktop ryzens tops out at 5.1-5.2 than TRs should get 5.3-5.4

Eh, that's not that impressive for the money. I got a 2700 for $220 and I'm planning on using that chip for sevoural years to come.

>TRs should get 5.3-5.4
holy shit intelfags tears will be delicious

>Having more cores, using less power, at higher clocks on a node no one else has at their disposal sans mobile ARM chips that don't matter isn't impressive
TR has me tempted just for that.
>16c32t TR at 5.4GHz on air

Single thread performance still matters. Even though Ryzen improved their standing greatly, they still fall behind Intel. I would really like it if Zen 2 performs at least as well as Intel in clock for clock performance. If so, I can imagine my next upgrade being to Ryzen. If not, I'll probably wait for Ice Lake.

>ice lake
>2020
JUST WAIT

3600x Sweetspot for gayman
3700x/3850x for stunting on corelets and blowing cash

Which one Ya getting

Don't fourget that you'll probably have to buy a new moatherboard even though they promised they would work on existing am4's

I love this post.

Attached: 1548705329535.png (198x234, 18K)

>intelfags when TR4 bois upgrade their 12 and 16 core chips to 32 core 5GHz monstrosities on air cooling with less power draw than their 5GHz 9900ks

Attached: Salt.jpg (620x267, 43K)

The 3850X if it's real.

You're a crackhead Threadrippers pull more power than most high end gpus, even with a node shrink doubling the core count will most likely negate any power savings

AMDrones waited for 10 years for Ryzen

Intel fanboys can wait a couple of more years

3300 to dab on 8600 goys with a $500 build

Imagine being this delusional.

There is no reason for AMD to have to force new motherboards
With the way Ryzen is structured as an SoC most of the motherboard is just dumb interconnects

Wonder if it would scale good on dual channel or if it would be like a 32core tr vs 16 where 16 is the more logical choice to get the most out of the performance in most cases hmm

INTELAVIV BTFO

Attached: received_338481236710224.png (534x299, 89K)

1080ti averages around 346w
You guys should learn the hardware you are shilling

Attached: Snips.png (2058x417, 376K)

16c/32t works fine. The 32c64t 2990WX failed because of the windows scheduler. It literally doesn't know how to handle that many cores/threads since windows 10 is a consumer platform. Linux had better results. Which is why I'm betting on a 16c32t AM4 for as high as I'm willing to go because I won't gold.mt breath for MS to fix their OS.

>you'll probably have to buy a new moatherboard
What makes you think that?

AMD said that they're not going to release an APU this year. That list is so fake.

AMD said that they are not using the matisse design on apu's this year.

kek

Zen2 APU will be monolithic btw.

What frequency was the 8 core Zen 2 ES that beat the 9900K running at again?

4 GHz

mid 3's? I think.

>linustechtips
>presenting total system power consumption from the wall as just cpu draw
>overclocked chip using less power than stock
it's just an honest mistake right, it's not like linustechtips would ever lie to you

Attached: image.png.cc4bc6bca5b56e6001778847305102b1.png (1200x422, 214K)

Pls say this is shooped kek

based intel chad
amd incels btfo

Attached: 1551620816421.png (806x577, 61K)

eh. As much of a shill that user is, they have a point about power consumption for the top end TR 2990WX. Tom's Hardware review of it had the CPU pulling almost 500w when using AMD PBO. At stock it was only pulling 250w. This was with Prime95 AVX testing though too.

Fuck that's still only 16w a core worst case
At stock each core is almost getting down to laptop levels of power consumption
yeah it draws lots of power even compared to a GPU but look at what your getting in return
Besides it's much easier to make a GPU draw 350w than it is to make a TR draw 350w on anything that isn't an artificial load

>still lower than intel's 800W xeon garbage
based
7nm zen is going to mop the floor with intel abortions

youtube.com/watch?v=dt5CNi0aEpI
3:18 in, linking to timestamps on Jow Forums is ass
autists on the ltt fourms were rationalising it as 'it's just averaging' 'ryzen is averaged higher' 'if anything it shows how close they are', I didn't bother looking in the thread to see if there was an official reply, the other games have similar results

agreed that it's an absolute monster but it's 32c/64t and comparing it to an 18c/36t 7980xe like most reviewers have done is downright disingenuous, and the poster suggesting a 32c/64t intel would be any better in power consumption is nothing short of shilling

a 32c/64t intel cpu would destroy the universe as we know it

uma delicia and favelapilled
awesome pasta

uma delidcia*

What I find funny is the common reason the 2990WX and Intel 7980xe get compared is availability and pricing. The 2990WX punches FAR above its price and goes toe to toe with Intel Xeon Platinum 8180. Google says the chip is "supposed to" sell for ONLY $10,000, but realistic prices are around the $13,000 mark per CPU because of availability and low yields. This of course doesn't even consider EPYC.

Implying there is some magical reason it won't

The better reason for why they are compared is the fact they are both the top of the line for consumer/enthusiast parts, it just happens that the 7980XE is quite available

pls do the needful and delet this sir thank and have a nice day

Attached: OYVEY.jpg (1440x809, 194K)

Competition is good for everyone

>too good to be true and makes no business sense
>its just progress its not fake bullshit.
You fuckards act like you have never seen a die shrink before. 22nm to 14nm didnt change that much and i cant wait to see how btfo you all are when you see a die shrink isnt some kind of magic

>22nm to 14nm
That's because it's a half node. 14 to 7 is a full node shrink

22-14=8nm, half node shrink
14-7=7nm a full node shrink

Elaborate please

That's not how it works.
14/2 = 7
22/2 = 11
if greater than prev_node/2 = half-node
if less or equal than prev_node/2 = full-node

Do we seriously need to wait until july?

It's 3 months between packaging, binning and testing, and no QS floating yet, so yeah, maybe even a bit longer.

I mean fabbing, packaging and binning

its 12/2

12nm is a marketing name. It's still 14nm but with small optimizations

I would love to see a ryzen competitor to intel's core-m series

Any reason to upgrade from R7/1700?

Its 12nm cope more its a half die based on your retarded explanation

if you gayme
Ryzen V?

There is no competition

Intel CPUs went AWOL since August 2018 and stocks and prices haven't normalized up until this time, they may have released 9th gen CPUs but they're overpriced

AMD has a monopoly over desktop CPUs for at least 6+ months already

>die shrink
>poozen still maxing out at 4.7 @ 1.4v
Gun be gud

go back to /v/ you retarded nigger

When are these suppose to release? I’m about to build a new pc and was planning on getting a 2700. Should I just wait?

>14nm isnt a fake marketing meme
>12nm is a fake marketing meme
>7nm isnt a fake marketing meme
>half node and full node shrinks arent fake marketing memes
This is how fucking retarded you sound. Either listen to the manufactuer or dont, but you cant pick and choose to make up your retarded theories

Lay off the meth. Reminds of the old Hive posts

Eat shit, you clueless moron

>t. retard parroting dumb shit other amdrones post in /pcbg/ every day

Whatever makes you sleep at night, kike. Post some AMD wojaks, that might help with your anger

kek

Attached: lain-smug.png (811x710, 542K)

Attached: 1449571016568.gif (322x354, 2.83M)

Where is the class action lawsuit if 12nm is just a marketing lie? You know you cant lie about computer parts and not get drug through court. Is the rtx line a lie too because its 12nm? Are you saying all processes are a lie and 7nm is a lie too? You know best so please let me know smart guy

Why are they upping core count this fast? Why they wait like 2 years to up core count and tweak design the year in-between like Intel. It worked for Intel so I don't see how it shouldn't for AMD

You don't understand mommy su, so i will explain her to you.
She will get every list on the internet, EVERY list, and will assure that all the chips names, models and prices don't match any of em.
Even if you manage to somehow predict her plans perfectly, she will just add an 1 on the numbers in the CPUs just to be sure you will fail your prediction.

You're truly special.

Attached: logo.png (584x251, 21K)

Because AMD is starting an arms race that at best will alienate users of older dual core and quad core systems as application development shifts towards leveraging multi core work loads (Blame the possiblily of AMD Releasing a 6c/12 Ryzen 3) or at worst AMD gamers will have like 6 idle cores and productivity workers will save like 1-5 minutes

Because its the most efficient way to increase performance.
Every extra instruction per cycle you get in a CPU
double the size of the CPU, like, a core that can run 4 instructions in parallel is twice as big as an CPU that can run 3 instructions in parallel, that is again twice as big as one that can run 2 and so forth.
This happen because the CPU have to reorganize the code and rename registers and to all sorts of under the hood magic to give the program the illusion that it is being executed one instruction per time, with the expected non fucky results in the registers and flags etc..
So, when you just duplicate the CPU, you double the theoretical number of instructions per second while only doubling the CPU size, instead of making it like 16 times bigger.

But there are other options too. A popular one is to make more complex instructions, such as vectorized math, where you add this 256bit ALU in the CPU, and make it process eight 32bit numbers glued on a pack, thus making the ALU do the equivalent to 8 instructions in a single instruction, but this only makes DSP like shit faster, and won't help with more complex CPU shit.
Also there is the one intel tried, where the CPU runs "instruction clusters", where you have a super instruction that can do 5 things at once, as long they don't "collide" with each other, but good fucking luck writing a compiler that can handle this shit.

And finally, you can try increasing the clock rate, it was a sure fire way to increase performance until the moore's law did hit a brick wall, so now you can only get marginal gains from this side.

can t tell if this is a copy pasta or a pajet trying to type in english

>dual and quad core users are fucked because AyyyMD
Not true. If anything web devs and Web2.0/3.0/4.0 or whatever we have right now are to blame. Partly of the blame falls on Intel too for holding core count so long. If anything AMD are the sensible ones on that front but my point was a fiscal one. It doesn't seem sensible to me for AMD's aggressive core count increase. I haven't seen their fiscal year on year but it seems like shooting themselves in the foot.

It's not AMDs fault they are gonna alien low core count people, they are moving the industry forward, it's worth it but people will bitch.

I honestly think core counts will be around the same i think this list is leak bullshit, but the magazine likely confirms the names. Technically if the IPC gains they claim are true they could make the Ryzen 9 a 12c/24t part sell it for 500$ and still be on top, I think expecting a 16c desktop AMD chip this gen is at best a pipe dream as it would destroy the "value" they presented with the 2700x

Clueless with no arguement

In a perfect world, companies like Intel would not exist, but this is not a perfect world...

In a perfect world, there would be three viable choices for CPU.

>16 core
>5.1Ghz
Fake