Comfy AMD Thread

>Ryzen processors using Zen+ will see an approximate performance boost of about 10%

Ryzen 1700 Single Thread Rating: 1774
Ryzen 1800x Single Thread Rating: 1965
Ryzen 2700 Single Thread Rating (approx): 1951


Throw in improved memory latency, memory compatibility and IPC, that effectively makes the upcoming 2700 an 1800x according to PassMark. Have there been any other giant leaps in performance like this in CPUs recently? Also, comfy AMD thread (Zen appreciation, etc).

Attached: ryzen-galaxy-s7-edge-rom.jpg (1139x759, 67K)

Other urls found in this thread:

cpu.userbenchmark.com/Compare/AMD-Ryzen-5-1500X-vs-AMD-Ryzen-3-1300X/3921vs3930
cpu.userbenchmark.com/
youtube.com/watch?v=sjzzdkUvKhk
twitter.com/SFWRedditVideos

>giant leaps in performance like this

Attached: 1521501110562.png (800x729, 48K)

>Zen appreciation

Attached: 1521583324098.png (768x752, 88K)

>AMD thread
Who else is anticipating Navi? Will it be our saviour against the Nvidia Jews with their proprietary software and GPP? Will it be the second coming of Vega disappoint?

Same

>giant leaps in performance

More like they derpt the pipelining of zen1 and needed zen1.1 to fix it.

>giant leap in performance
i miss the old times. my 4770 is still going strong ill definitly wait out for zen 2 or whatever intel held back

Attached: 1496407655703.png (1070x601, 495K)

>It's okay when WE fellate a multinational corporation!

Attached: 1521667796646.png (394x450, 17K)

>pissmark
just use geekbench while you're at it then

whoa AMD BTFOs intel with 10% boost

Still better than the -1% improvements Intlel has us accustomed to.

I would say Bulldozer to the Zen architecture was a leap. Not Zen to Zen+ (10% performance jump isn't that major FYI)

It can probably be seen as major in closing the gap with Intel. Give it a couple years.

How much more optimization can AMD get out of zen?

Attached: 1437392201671.png (741x649, 23K)

well that IS about 10% dumbass

srs question here AMDbros

what's the diff between the 1300x and the 1500x, besides the threads? a 60 dollar difference for literally the same performance and cores makes no sense to me.

cpu.userbenchmark.com/Compare/AMD-Ryzen-5-1500X-vs-AMD-Ryzen-3-1300X/3921vs3930

Attached: 1521670564587.gif (645x773, 41K)

Aka CPU shitposting thread.

So not really worth for anyone on Zen. Gotta wait for Zen 2.

less L3 cache
also stop using that bullshit site for comparing cpus, look for benchmarks made in actual reviews

>actual reviews

Attached: 1521616584623.png (642x722, 371K)

I believe a gorillion basement dwellers more than I believe a single paid reviewer

BAD GOY

Attached: 1500647453955.jpg (250x250, 9K)

>don't want to believe paid reviewers
>but believes in paid "benchmarks"

>paid "benchmarks"
Explain yourself

Kinda disappointing for 12nm

>Have there been any other giant leaps in performance like this in CPUs recently?
Considering that the one other relevant CPU manufacturer out there actually had a regression in performance, it isn't that impressive for AMD

10% is more than what Intel has been doing. Shit, Intel is LOSING around 1-2% of performance on their 10nm.

12nm is a meme node, it's more like 14nm+

>you can only buy processors made by scheming kikes or incompetent poo in the loos
WE TRULY LIVE IN WORST TIMELINE REEEEEEE

Attached: ba2bdda2fc37c3a3975e7c1c79465aa918191cf54553876c21c62ce23e972a7a.gif (500x351, 84K)

what kind of reasoning is this, you could argue as well that intels cpus were more efficient already

Did you guys hear, that the new patches for the "fatal flaws" have no performance impact what so ever? CTS is a joke.

If it turns out to be paid by Intel, why are they so bad at their attempt on hurting AMD?

Attached: AMD_Attacks_CTSLabs.webm (480x360, 2.01M)

yeah, i'm on my even older Phenom II x4 965, it still holds up reasonably. However the new Ryzen 2000 might make me consider it. The interesting thing is not the Mhz, but XFR2 and the Cache and Memory speed optimizations.

>paid banchmarks

cpu.userbenchmark.com/

nigga are you fucking blind? the software download is RIGHT THERE, this website just gathers actual benchmarks made by losers just like you and me, no money involved.

the kikes bought king pajeet though

How is this paid? passmark is not paid

obviusly disappointing.
it's still fucking gcn and amd/rtg is still poor. hint: you need a lot of money to develope a good architecture, and that is something nvidia has and amd/rtg has not. honestly it would be the best if intel just buys rtg, they would throw in some cash and release something a few years later that would pretty much rape nvidia. or any other company with enough cash to invenst really, rtg just needs money. when rtg stays with amd they will need to fight for every penny with the "goliath" intel, leaving them with basicly no real budget to create a good architecture(basicly what happened the last 6 years)

intel doesn't hold back, their architecture is already maxed out, their process is already pretty good (10nm is a rather small step for them), so all they can do is add more cores which does not scale that good with their architecture anyway. the main question is: will zen2 on 7nm be enough to compete with intels single core performance or will these cores still be weaker. it's all about one massive step amd needs to make. If they fail to deliver intel will ass rape them once they release their new architecture in 2020/2021.

This.

>t. Sup Forumslet

I guess Zen+ may be not faster, it allows better clocks and yield. Just like Threadripper can more easily archieve 4.2 to 4.3 GHz below 1.4 V.

Am I /comfy/?

Attached: speccyhtpc.jpg (1011x798, 92K)

I wonder if the exploits can be used to disable PSP.

What exactly is Zen+? I'm still a bit confused.

What's the difference between Zen, Zen+, and Zen 2?

Is a 1600X a Zen+?
Is it worth getting one, or waiting for the 2600X next month?

Sorry for stupid questions, been using Intel for my last couple builds and am wanting to switch to AMD now.

Attached: 1293089416447.jpg (246x245, 29K)

For those who are interested, I'll dump my researches.
>3500 MHz
>2666 MHz Dual Channel CL16-17-17-36
> 8 GB

Attached: 1cine.jpg (757x1966, 103K)

Attached: cine.jpg (758x2014, 103K)

Attached: 1cpuz.jpg (611x606, 65K)

Attached: cpuz.jpg (611x606, 65K)

Attached: 13dm.jpg (2411x503, 109K)

>Is a 1600X a Zen+?
No, the 2600X is the Zen+

>in layman's terms
Zen is all of the original AMD Ryzen processors like the 1600X, 1700 etc, and Zen+ is all of the processors that are being released shortly.

Attached: 3dm.jpg (2692x581, 130K)

ohhh ok, so Zen+ =Zen 2

>gtx 1080
>gtx 960

Attached: 1296895603864.jpg (640x640, 49K)

more precisely zen+ = Ryzen 2xxx

Its obviously not the same PC, but just look at the CPU scores. The rest doesn't matter.

I did these tests with hexacore and HT/SMT aga in.

Attached: 1cine6.jpg (767x1971, 107K)

zen+ is the zen architecture but on 12nm with some small improvements in turbo boost and memory management.
zen 2 will be the next generation core, with both architectural changes and smaller node.

cool, thanks for the clarification

It is noticeable, the 1700 scores better multicore wise, even at the same clock and RAM speeds.

Attached: cine6.jpg (746x1988, 107K)

Will the processors follow a Ryzen 3xxx naming scheme?

The thing that has me hopeful is the rumour the Navi GPUs may be structured like their Ryzen CPUs with an infinity fabric linking multiple dies. Dunno how it'll translate to graphics processing, or if there's any substance to this rumour.
>no good Radeon options in the last 6 years
Still miss my 5970

I don't give a fuck about Passmark and Geekbench scores, give me real world use case benches and then we can talk. For all we know the tweaks made might give you 20%, or 1% extra performance.

Convince me to upgrade from my i7 6700k
i do gaming and video rendering, what are the REAL advantages from getting Ryzen?

youtube.com/watch?v=sjzzdkUvKhk

In real-world, and has the bandwidth where Intel is gimped on context switching.

for each ryzen sold, 1 million jews are gassed

Intel 10nm is utterly fucked and is a performance downgrade in all areas except power usage at low frequencies.
GloFlo 7nm is based off IBM tech and targets 5ghz nominal operating frequencies.
Ryzen cores already compete 1:1, barring edge cases, they are just frequency limited as of now.
Intel's new architecture was delayed to 2022, and the technology behind it has never been proven to scale. It is such a radical change that compiler paradigms will need to be drastically altered to take advantage of its benefits. Sup Forums has already gone over the details that have come out sofar and unless Intel pulls a rabbit out of their ass then this will end up Netburst 2: Electric Boogaloo.

Attached: megatasking_intensifies.jpg (493x493, 59K)

>Intel's new architecture was delayed to 2022
Wat? Can I have some sauce with that? If true Intel's even more fucked.

Sapphire Rapids got delayed to 2021, that pushes back +1

>2 overclocked i7 in a card package
>core plenty'um 9000

I will buy 6

Affordale strix gpu's fucking when

Been waiting for almost a year at this point

Yes, they are similar to Haswell, but with even lower clocks. The only thing AMD has to do is decreasing latency and allowing more core clock.

What the hell is up with your multicore performance? I get 4560-4630 ish every time I benchmark, not to mention my single core performance is 430.

No. With 1xxx -> 2xxx the number got twice as big, but going from 2xxx->3xxx is much smaller relative change and that's why the next one will be at least 4xxx

>buying new cpus at this point

Attached: 1514025297931.jpg (620x470, 53K)

fake news

I currently have an 8700k and my friend's 1600 rapes my PC. Can't wait to get my hands on a Zen.

Attached: Ibringyoubaitchanhereyougoguys_66c7c6_6156191.jpg (900x1357, 118K)

>Ryze thumbnail
>Intelavivs crawl out to shitpost in this thread
free bumps i guess

>10% performance jump isn't that major FYI
first few ryzen microcode updates brought ~5-10% performance increase across the board. more than (((Intels))) jump from generation to generation.

>CTS is a joke.
OY VEY

Attached: 1521134686517.jpg (1170x1974, 287K)

>being this retarded
neck yourself

Attached: blender.png (480x288, 5K)

t. 2012

I don't see 1600 or 8700k in this chart.

If that were the case you should at least be able to get better than SLI performance out of chips linked by interconnects, and unlike a lot of desktop software, ALL 3D renderer are capable of using basically as many CUs as you can throw at them.

Honestly, it's less likely that they were paid by Intel than trying to manipulate stock prices.

12nm has almost completely closed the single core gap, with the exception of the tippy top of overclocking.

>what kind of reasoning is this
It's because intel has been refreshing the same architecture for the past decade now and have reached it's limitations. To keep competitive they have to change something. Monolithic dies are just too expensive to manufacture.

'Moore's Law' is dead

Moore's law is not dead. The number of transistors we can fit on a die of a given size still approximates mores law, but the effect that this has on real world performance has diminished.

>Who else is anticipating Navi? Will it be our saviour against the Nvidia Jews with their proprietary software and GPP? Will it be the second coming of Vega disappoint?
If AMD manages to pull out multichip trick with GPUs it will be perfect storm for Nvidia.
It's AMD whose chips are in PS and Xbox, Nvidia failed at that so console to PC ports which were optimized for AMD in first place.

AMD also is leading own Gameworks alternative and they do have money for that now thanks to Ryzen and console success.

I am not fanboy and I do not know whether Navi will be on top of everything but it will sure as hell be competitive even if not ideal at everything. Remember, Vega is competitive already, minus power consumption, top of the line performance and Gameworks titles, it can only get better.

amd not going to release any gpus in 2019?

I just want the Raven Ridge thinkpads already.

>Remember, Vega is competitive already, minus power consumption,

Vega failed because it has the same per-clock geometry throughput as Polaris and Fury/Fiji (4 triangle fragments/clock), since AMD dropped the ball on getting the primitive shader/discard functionality integrated into existing graphics APIs.

Cranking the voltage/clocks up and leaning on pixel shading power was their last recourse to semi-salvage the product.

They should have gone to 8 tri/clock or 2 tri/SE until the new discard stuff was actually proven.

When should I replace my 2500k with a ryzen?

Zen 2 will be a good replacement if you can wait a year or so. Zen+, coming out in April, would probably be good too. Try the Ryzen 2600.

where is my12nm vega64 @ 1800mhz?

I can wait, this thing still just werks surprisingly
Wish I had an amd card for that lunix support

Attached: 1478143898940.jpg (360x360, 21K)

I'm guessing it would make more sense to wait for Ryzen 2 to replace my R7-1700X?

Yeah, Zen+ won't be that much of an upgrade. It's really good for people like me still on a Phenom II 1090t, or any older chip for that matter.