Nadolazeći AMD socketi, ploče i procesori

[Momir]

Meni se taktvovi bas cine realni, I/O dio ce vjerojatno raditi na taktu memorije, a jezgre bez problema na vecim taktovima.
Možda se same jezgre ne razlikuju previše od prijašnjih, ali su oslobođene tog mem konrolera i još kojekakvih sitnica.
Uostalom, mislim da je na trenutnim epyc/threadripper takt infinity fabrica vezan uz brzinu memorije.

[Sn4k3]

Meni isto taktovi ne djeluju totalno prenapuhano, iako se realno ne nadam nicemu preko 4.5-4.6Ghz. Logicno je da izolirane jezgre ciplet dizajna lakse dostignu vecu brzinu u odnosu na monolitni dizajn di su ugurani i razno-razni kontroleri... Afaik, na svim zen/zen+ cpu-ima je infinity fabric vezan uz brzinu rama u omjeru 1:1. Dok se za zen2 prica da ce infinity fabric raditi na duploj brzini u odnosu na ram oliti u omjeru 2:1 (nesluzbeno).

[DuHc]

Autor The Exiled

65W Ryzen 5 3600G 8C/16T@3.2GHz + iGPU i to je to kaj se mene tiče.

I meni je zapeo za oko kad je najavljen, savršen za neki mini-ITX build...Čini se zasad ništa od njega i ostalih APU-a?

[Smartic]

Lako moguće da dođu u sličnom terminu kao i 1st Gen Ryzen APUi (krajem zime), a sad će u Q3 izbaciti Zen+ refresh(3200G i 3400G) da pokriju rupu.

[stef]

Svi APU procesori će biti monolitni Zen+ čipovi. Zen 2 dolazi u chiplet verzijama.

[Manuel Calavera]

Future AMD Ryzen CPUs Built on TSMC’s 5nm Process Node Can Be Up To 80% Denser Than 7nm Ryzen 3000

Samsung B-Die Memory Production Ceased, Replaced by A-Die

[dewil23]

https://www.asrock.com/news/index.asp?iD=4238

[MadBoY]

A320 Chipset*
*Supports AMD Next-Gen Ryzen series APU only.

Ok, znaci definitivno nista od Zen 2 podrske za A320, i ujedno indirektna potvrda da ce Ryzen 3000 APU biti samo Zen+ refresh.

[The Exiled]

US DoE announces Frontier Supercomputer: Cray and AMD to Build 1.5 Exaflop Machine

Frontier is planned for delivery in 2021, and when it’s activated it will become the second and most powerful of the US DOE’s two planned 2021 exascale systems, with performance expected to reach 1.5 exaFLOPS. The ambitious system won’t come cheaply, however; with a price tag of over 500 million dollars for the system alone – and another 100 million dollars for R&D – Frontier is among the most expensive supercomputers ever ordered by the US Department of Energy. Frontier is an open science system, meaning that it’s available to academic researchers to run simulations and experiments on. Accordingly, the lab is expecting the supercomputer to be used for a wide range of projects across numerous disciplines, including not only traditional modeling and simulation tasks, but also more data-driven techniques for artificial intelligence and data analytics.

Designed as a next-generation CPU + accelerator system, with a mix of CPUs and GPUs doing the heavy compute work, AMD will be supplying both the CPUs and GPUs for Frontier. And as the principle processor provider, AMD will also be taking on a lot of the responsibility for developing the software stack as well, with the company working with Cray to develop an enhanced version of their ROCm environment to best extract performance from the massive cluster of CPUs and GPUs. On the CPU side of matters, AMD will be supplying a customized next-generation EPYC CPU. AMD has confirmed that it’s going to be using a future generation of their Zen CPU cores, and given the timing of the project, we’re almost certainly looking at a Zen 3 or Zen 4 design here. Just how custom AMD’s CPU is remains to be seen, but their announcement has revealed that Frontier’s CPUs will include new instructions for the optimization of AI and supercomputing workloads.

Meanwhile on the GPU side of matters, AMD and Cray are holding their cards a little closer. Rather than naming any architecture or architectural generation, AMD is only saying that the GPUs are “based on the Radeon Instinct family” and have “yet to be announced.” AMD’s current public roadmap goes out to “Next Gen” in 2020, and with GPU development cycles averaging 2 years, this may be the architecture we see. But with the particular needs for a supercomputer, AMD may have something slightly more bespoke. What the company is confirming for now is that they aren’t holding back on features. The HPC-focused GPU is being designed with Frontier in mind and will incorporate mixed precision compute support. Feeding the beast will be HBM memory, and AMD will be tapping a version of Infinity Fabric to connect the CPUs and GPUs.

Izvor: AnandTech i ServeTheHome

[kopija]

Skynet: powered by AMD
sasvim neočekivano...

[darkydark]

Jedina dobra stvar je da će se taj stroj pregrijavat i sporo radit pa smo sigurni jos neko vrijeme... Oh wait AMD će biti unutra...

Sent from my FIG-LX1 using Tapatalk

[Manuel Calavera]

https://twitter.com/TUM_APISAK/statu...-4-2-ghz-es%2F


AMD CPU Market Share Rises In All Segments Except Servers in Q1 2019 – EPYC Rome Expected To Continue Server Momentum in Second Half of 2019

[Brko]

Jim
@AdoredTV
·
19m
Anyone expecting ~4.5GHz boost is gonna be very pleasantly surprised, that's all I'll say about that.

Znači ide i deblje

Sent from my SM-G965F using Tapatalk

[Manuel Calavera]

27. 05. je blizu

[Sn4k3]

Da otvorimo forumsku kladionicu na taktove?

[The Exiled]

@TUM_APISAK 6 hours ago

• Zen2 ES 16 Core
• Base clock 3.3 Ghzฺ
• Boost clock 4.2 Ghz
• MB X570

[Disco]

Ja kažem da će neki OC prosjek sa 4,1-4,3 koliki ja sada ako sam skužio, skočit maksimalno na 4,5ghz, sve preko će biti fešta

[GdjeSuBikovi]

Ako 8 C/ 16 T ryzen na ta procurena 4,2 Ghz bude performansi kao intel na 4,7 Ghz (sa prezentacije od prije koji mjesec), to nije toliko razocaravajuce.
Ja ipak ocekujem 4.5, a nadam se 4,7.
Ustvari bez obzira na takt, ocekujem performanse intela na 4,7 Ghz, a koliko povise toga ocekivati ne znam.

[Gigi1]

Nisam ocekivao nikako na novom nodeu toliki skok ali izgleda da ce neki ryzeni 3xxx serije boostati na 5 ili vrlo blizu 5ghz(+-max 200MHz)

[MadBoY]

Ja cu, kao i obicno, biti preoptimistican i reci da ce 8C16T bez problema ici na 4.8 GHz "all core" i napokon ce pobijediti intel u 1080p@2080Ti okrsaju .

jabi 16 jezgri @ 4.5ghz all core @ 1.2v

[Manuel Calavera]

Autor MadBoY

Ja cu, kao i obicno, biti preoptimistican i reci da ce 8C16T bez problema ici na 4.8 GHz "all core" i napokon ce pobijediti intel u 1080p@2080Ti okrsaju .

Kod igara dosta ovisi o optimizaciji, recimo da kojim slučajem zen2 bude brži u cinebench singlecore od novih intela, nije garancija da će i u igrama biti brži, jer neki enginei imaju samo optimizacije za intel ili bolje optimizacije za intel, neke igre koriste AVX poput assasins creeda i sl.

Naravno to ako b bio sporiji sa 2080ti/1080p ne znači da možda nebi bio bolji deal za gaming u cijelom paketu (cijena/performanse, jezgre/cijena, temperature i potrošnja i sl.)

[Brko]

Kaj vredi i 5GHz na Zen2 cpu kad će onaj poluproizvod PUBG i dalje raditi samo na lntelu. Nije PUBG jedini, ima ih još.
To nije do proca koliko do lijenih ili potkupljenih developera.

Sent from my SM-G965F using Tapatalk

[Manuel Calavera]

Kad vuku prastare engine, eventualno ih updejtaju no baza je stara. Naravno nisu sve igre takve, ako bi stvarno bio brži u singlecoreu - bio bi brži i u dosta igara. Situacija se mijenja, ryzen je bitan na tržištu, pa sve što se pravi novo od 2017 i nadalje ima i ryzen na umu. Eno baš sam linkao u vulkan temi - cryengine 5.6 i nadalje će imati:

CPU Optimization is most likely improved support for high core count CPUs. This will be beneficial to gamers using future Ryzen 3000 CPUs with CRYENGINE taking advantage of Zen 2’s design. The same goes for gamers running high core count Intel Core X-Series CPUs. CRYENGINE 5.6 will lay the foundation for CryEngine 5.7 and will be released within the summer of 2019.

A i ovako je engine već odličan u multithreadu (FX 83xx pobjeđivao i5-ice u crysis 3 i sl.)

Naravno dalje ovisi o recenzijama. Uvijek možeš uzeti review gdje recenziraju 5-6 igara i među njima bude recimo Skyrim i neki Fallout (u bazi predpotopni engine koji obožava singlecore), GTA V, eventualno neki Far Cry i možda par novijih/normalnijih enginea i već si utjecao na rezultate

[Disco]

Kolko sam ja shvatio Ryzen u igram je sporiji od Intela zbog ring busa.
E sad jedino ako se ta razlika može brute forceat. Ali svejedno u većini drugih stvari bi ih AMD trebao sažvakat i ispljunut. Ugl ja čekam rivjue, niš drugo me ne interesira dotada...naravno uvijek škicam ovakve leakove, ali rivju gotovog proizvoda je ono pravo

[Manuel Calavera]

Ne nužno, svaka igra odnosno engine su priča za sebe. Znači nikakav ring bus, ccx-ovi i sl. ne moraju biti razlozi ako je engine ima optimizacije za tu arhitekturu


Npr. za PUBG

I have experience with UE4 & Unity. They are only optimized for Intel architectures due to Intel's long standing dominance. It's important to note not all instructions are equal, so IPC is entirely instruction dependent. These engines compile non-optimal instructions for Zen cores, so even at the same clock speed, an Intel core will be quite a bit ahead, nevermind the lack of multi-threading. The guys responsible for these engines focus on fixing the plethora of bugs and so CPU instruction optimization is the last thing on their to-do list, unfortunately for AMD because these engines are very popular with indie developers.

[shar]

Sve se da brute forcat, pitanje je samo koliko je netko spreman žrtvovat i do kojih ekstrema ići.

[Smartic]

Možda i slabe AVX instrukcije kod 1st/2nd Gen Ryzena čine veću razliku? Vidio sam iz MSI Afterburnera da mi 8600K radi na -100Mhz AVX offsetu 70-80% gameplaya u PUBG-u.

[MadBoY]

Kaze Jim:

I can't leak the details but AMD have been showing at least two CPUs around the mobo makers. The 16C Apisak found is very likely to be one of the chips they were showing.

The other one was 12 cores and really high clock speeds. I don't know if they benched it online though.

— Jim (@AdoredTV) May 10, 2019

Ja se jos uvijek nadam 8c16t@4,8 GHz .

[Manuel Calavera]

Jebemu ako već tako piše, onda stvarno možda ide preko 4.5ghz. Što se tiče cpu ima dobar "track record"

S druge strane za AMD gpu (za razliku od nvidije) nema baš dobar track record, tj. Navi leakove mu baš ne vjerujem full (ne njemu, nego njegovim izvorima) jer neki modeli nemaju smisla tj. preblizu su (npr. V64 + 10% model i rtx 2070 model - to bi trebale biti identične performanse)


Uglavnom Zen2 će biti ubica, al to sam znao i prije godinu-dvije