nVidia plans to release its next-generation Ampere series of GPUs in the second half of 2020, adopting 7nm manufacturing and new architectural enhancements to deliver significant performance and power efficiency boosts. nVidia's next-generation graphics cards will reportedly ship with a 50% boost to graphics performance while halving power consumption, providing gamers with vast generational improvements. Sadly, these metrics do not describe what workload nVidia's next-generation architecture is under, as this 50% boost could be referring to fully raytraced games, like Quake II RTX, or traditional rasterised games. What we will say here is that these rumoured numbers don't necessarily mean what they imply. A significant boost in raytracing performance with decreased power consumption is what should be expected from a second-generation RTX product, especially when combined with 7nm. It is also possible that Taiwan Times' source is incorrect, or is overestimating Ampere's performance. Regardless, nVidia has had almost two years since the release of Turing to create a successor, and a shift from 12nm to 7nm will yield significant benefits which will complement any architectural enhancements that are made with Ampere. We are expecting big performance enhancements with nVidia's RTX 30 series, or whatever nVidia decides to call it.

Autor Drug Brko (Post 3373187)

Još 2 mjeseca bulšitanja i nagađanja dok Kožnati zauzme pozornicu i krene svršavat kako je Master Of The Universe. Ali kaj god da taj Ampere bude, opet će AMD samo gledati kad dan danas nemaju karticu s kojom bi skinuli 4 godine starog Pascala 1080Ti.

Moram priznat da mi RTX3080 7nm 10GB jako jako lijepo zvuči ;) bit će sigurno nekih 6-7 kkn garantirano. Ali i mislim da sa njom 5 godina mira i tišine minimalno budući vidimo highend Pascale kako se još drže a olimpijski ciklus su odradile :D

Autor Drug Brko (Post 3373288)

Nekako ne vjerujem. Takvog skoka sa Maxwell na Pascal nebu tak skoro jer konkurencija i dalje nije dovoljno pokazala zube.

Autor Drug Brko (Post 3373486)

Ok...makes sense.
Onda stotare u škrabicu dečki već sad lagano ;)

Autor The Exiled (Post 3373513)

GTC i GDC su u trećem mjesecu, tak da bi nVidia (ako već ne ide konkretno lansiranje proizvoda) to mogla iskoristiti za prve pokazatelje nadolazeće arhitekture, koja je nećemo se lagati već zgotovljena i u rukama HPC/AI/ML društva u obliku prvih Quadro i Tesla čudesa, a kad se u recimo sedmom mjesecu i ostali domognu novih modela - ova scena - bi fino prikazala mišljenje kritike i publike.:):D:chears:

In an article outlining the installation of Indiana University's Big Red 200 supercomputer – which also happens to be the first Cray Shasta supercomputer to be installed – The Next Platform reports that Indiana University has opted to split up the deployment of the supercomputer in to two phases. In particular, the supercomputer was meant to be delivered with Tesla V100s; however the university has instead opted to hold off on delivery of their accelerators so that they can instead have nVidia’s next-generation accelerators, which would make them among the first institutions to get the new accelerators. At any rate, according to Indiana University the group expects to have their new accelerators installed by later this summer, with Big Red 200 running in CPU-only mode for now. The Next Platform article goes on to state that the newer accelerators will deliver “70 percent to 75 percent more performance” than nVidia’s current V100 accelerators, which assuming it’s accurate, would make for a hefty generational upgrade in performance. Though as always, with multiple modes of compute involved – everything from straight FP32 vector math to tensor operations to low precision operations – the devil is in the details on where those performance gains would most be realized. In the meantime, nVidia’s next GTC event is scheduled for mid-March. So if nVidia is planning to launch a new Tesla, then I would certainly expect to see it there.

Autor The Exiled (Post 3363680)

Kaj se rekordne zarade tiče, postoji još jedna firma:) koja već jedno dulje vrijeme također bilježi rekordne rekorde u području koje Intel očajno želi i nadaju se da će im njihove Xe kartice omogućiti barem mali dio tog kolača. Jedina razlika između te firme i Intela je u tome da je prva firma alfa i omega u tom tržišnom segmentu, a po pitanju zajebanosti drži lekcije Intelovim Ryan Shrout PR mudrolijama. Firma je to koja u svom arsenalu ima ama baš sve i još koji komad više da Intel u narednim godinama i popratnim pokušajima dočeka na kojekakve komade noževa, a da pritom ne trpne dvaput dok im njihov CEO smješta od svega pomalo u tak jasno najnovijim proizvodnim procesima (1 - 2). Ne bi me čudilo da se Intel ubrzo nađe u situaciji u kojoj CEO one firme s početka donosi odluku da se njihovi proizvodi u OEM formatu isporučuju isključivo i samo u kombinaciji s AMD EPYC rješenjima, jer Intel jednostavno više nije ne samo najbolji, nego ni dovoljno dobar. Intel bi se tako vrlo lako i u rekordnom roku mogel naći u raznim i ne baš previše bajnim situacijama.:)