Does Nvidia have a future?

[starac]

Opet pricate gluposti

http://www.nvnews.net/vbulletin/showthread.php?t=150416

Well, take your pick. Head over to Zipzoomfly, Newegg, and Tiger Direct for all the Fermis you want. It seems the supply is getting a LOT better. I'm sure Charlie D. at SemiAccurate will be happy to hear that.

 

[DariusIII]

Na newegg-u ima GTX470, GTX480 su sve Out Of Stock, zipzoomfly GTX470, zipzoomfly GTX480 a slična je i situacija na Tiger Direct.

Nije to baš tako bajno kako nVidia predstavlja.

 

[Dusan K]

Vecina je u "Notify Me" statusu Kartica nema dovoljno, i najbolji dokaz je cena, koja je visa od MSRP.

 

[starac]

Pa karte dodju i nestanu.

 

[Dusan K]

Pa karte dodju i nestanu.

Naravno, ali je glavno pitanje koliko ih stize Slozices se da je znatna razlika 20, 50, 100 ili 1000 komada dnevno.

 

[ikonoklast]

[YOUTUBE]LTFzRsCj-bg[/YOUTUBE]
[YOUTUBE]2McPMFE1F4A[/YOUTUBE]

Spojler

Nvidia CEO: Netbooks and Tablets to Meld, Hints at Tegra-Powered webOS Devices:

How do you see the netbook evolving? Are tablets just going to take over?

I think that the first generation of netbooks tapped into a growing market for what I call leisure computing. The netbook was first, but the tablet is the new generation of leisure computers. In the beginning people just used these devices to surf the web, and if the performance wasn’t great when they accessed video, or high-definition Flash, they were generally understanding. And the reason for that was because the netbook was inexpensive, it had really good battery life, and it was really liked.

Now the iPad is out, and pretty soon Tegra 2 tablets will be out, and you’ll see high-definition video an Flash, as well as people reading magazines and newspapers on a tablet. And consumers will no longer understand why the netbook can’t do it. Unless netbooks start becoming very good at these things, you’re going to have some trouble. We can fully expect notebooks to become super-accelerated, awesome devices that are still low-power and still very affordable.

...

But by the time Tegra tablets hit the market won’t Apple’s head start be too great to overcome?

No way. We’re one million units into a multi-billion unit opportunity here. Here’s what I believe. In this era of leisure computing the person doesn’t write software, isn’t trying to design a car, isn’t trying to create a movie, isn’t trying to create a PowerPoint slide – just consuming content. And this device can come with either a keyboard, clamshell, i.e. a netbook, or it could come with touch, or it could be a combination of the two. It could be a detachable, wireless keyboard with touch. There are all kinds of interesting formats coming. So the separation between a tablet and a netbook is going to become quite vague.

How would you say your chip compares with Apple’s A4 processor?

Tegra is the world’s first dual-core processor with symmetric multi-processing, and what Nvidia’s most famous for, which is built in GPUs. So the dual-core allows it just to be much snappier. Just look at how dual-core Atom’s going to be much faster than a single-core Atom. And the second thing is because so much of the tablet experience is related to tablets, visual, the whole computer is just one big piece of display. So it’s a visual computer in its ultimate form. A fast GPU will just help you have a much snappier experience, just everything you do from magazines to videos to games.

...

What’s your take on Intel’s Z6 Moorestown chip? Will they be competitive?

No. It’s not possible. You could give an elephant a diet but it’s still an elephant. And when they think about power they think reducing from 20 watts down to 5 watts down to 4 watts down to 2 watts is really good. But you and I both know that in a mobile phone you need to be in a hundred milliwatts, two hundred milliwatts. So they’re still ten times away. So that’s a big challenge for them. And meanwhile Tegra 2 is already much superior to Atom from a performance perspective. And so now we’re already dual-core, and then next year I assume Tegra 3 comes out, and then, you know, here we are increasing performance at a lightning rate and power is incredibly low, so I think it’s going to be tough for them.

...


Given that you’re focused on Android, does that mean you’re not going to support other operating systems?

The first generation we were only doing for Windows Mobile. Since then we’ve added Android. I think the world will want webOS. I think it’s a fabulous operating system. And we’ll continue to do Windows Mobile. And we’re just larger in terms of resources than what we used to be. When we first started with Tegra 1 we only had two hundred people. Now we’ve got a thousand people on the Tegra team.

Can we expect a Tegra chip to be running the first Web OS tablet?

I can’t comment on that. But it sure would be an honor to work on webOS. It’s a great operating system. If you look at the first generation of webOS phones, the Palm Pre, the UI is just brilliant. It’s just too slow. So it needs a faster processor. Otherwise, it’s a great operating system.

iPad set the bar too low

Here at the Netbook Summit, Nvidia product director Bill Henry criticized the iPad during a session on new netbook paradigms. He said Apple set the bar too low, and that the lack of Flash and 1080p video support are major strikes against the device. He also threw up a slide of all the sites users can’t access today on the iPad, including Farmville in Facebook, which has 31 million daily users. Interestingly, one of those screen shots was CNN.com, which last time we checked played video on the iPad just fine. Still, Henry also promised that Nvidia would help facilitate console-quality gaming on tablets.

Slon na dijeti...

 

[Ace Rimmer]

Odlična analogija .
Samo da se konačno pojavi ta Tegra u obliku konkretnih proizvoda, po mogućnosti na Androidu.

 

[Nedjo]

That all sounds good and well, NVIDIA, but it's time to stop talking and start showing some real Tegra 2 phones and tablets.

http://www.engadget.com/2010/05/24/nvidia-intels-moorestown-is-like-an-elephant-on-a-diet-ipad-s/


Realno, pored Samsunga i njihovog Hummingbirda i pored Qualcomma i nihovog Snapdragona (a tu je sada i Apple sa "A4") Tegra 2 ce da ostane samo na nivou ovih prototipova...

da, ne treba zaboraviti ni TexasInstruments i njihovu Dual-Core Coretax-A9 seriju OMAP 4000

Fazon je da u tom biznisu u koji Tegra 2 pokusava da se ugura proizvodjaci smartphone-ova i tableta koriste ili sopstvee cipove (Samsung, APple) ili koriste QUalcomm-ove (HTC, SOnyEricsson, LG) odnosno TexasInstrumentove (Nokia, Archos)... a koriste i jer su to SoC-ovi u pravom smislu reci sa integrisanim radijima. Buducnost tih uredjaja nije u WiFi-ju koji mozes imati sa bilo kojim WiFi chipom, buducnost thi uredjaja je u 3G/4G (LTE, WiMax, HSDPA+) radijima, a to imaju samo ovi sto isporucuju kompletan SoC! Qualcomm sigurno nece NVIDIA-i prodati tehnologiju HSDPA+ radija, a to nece uciniti ni Samsung za LTE...

Ja ne vidim buducnost za Tegru u aktuelnim trzisnim okolnostima i trendovima.

 

[Ace Rimmer]

Za početak, portabilna multimedia/gaming platforma u obliku pad-a ili netbooka. Tako je to i Nvidia zamislila, ako se ne varam . OEMovi mogu dodati modem po želji ovisno o standardu.

You'll notice the one thing missing from NVIDIA's Tegra 2 is a cellular modem. There simply isn't one. NVIDIA's philosophy is to focus on the core compute functions of an SoC that require no carrier or FCC testing. An OEM could mate a Tegra 2 with a tried and true modem, lose out on the integration side but win in time to market. Given the sheer number of different wireless networks in the world, leaving the modem out of the design makes sense to me. But then again I don't make smartphones. It may prevent Tegra 2 from going into the cheapest solutions, but that's not where NVIDIA wants to be in any case.

Izvor:
http://www.anandtech.com/show/2911

 

[Dusan K]

Evo ga i prvi mobile GPU...

GTX 480M

Na ovim klokovima karta ce biti sporiji od Junipera (5870 mobile), uz vecu potrosnju, tj potpuno nekonkurentan po svakom pitanju u ovom segmentu, i pitanje je ko ce od OEM-ova uopste biti zainteresovan. Cudi me da nisu cekali GF104.

 

[Rapture]

Ovo jo dosta sporije i od gtx-a 260, samo *****i ima da se zalete jel je naskrabano gtx480M.

 

[Nedjo]

Evo ga i prvi mobile GPU...

GTX 480M

Na ovim klokovima karta ce biti sporiji od Junipera (5870 mobile), uz vecu potrosnju, tj potpuno nekonkurentan po svakom pitanju u ovom segmentu, i pitanje je ko ce od OEM-ova uopste biti zainteresovan. Cudi me da nisu cekali GF104.

pa nece nijedan! Vidis da se zahvaljuju Clevo-u na saradnji u tom saopstenju! A Clevo je je proizvodjac za "A" brendove.

Na ovaj nacin oni pokusavaju da naprave neku vrstu javne ponude ka "A" brendovima u fazonu, evo mi smo ovo spremili, a Clevo ce da proizvodi, pa vi izvolte kupite.

U prevodu oni nisu osigurali niti jedan "A" brend, a niko nije lood da rizikuje urusavanje svog brenda, ako krenu posle godinu dana da se pojavljuju isprzeni notebookovi.

 

[Dusan K]

pa nece nijedan!

Gotovo izvesno da ce tako biti. Ona neodredjenost je tu reda radi

 

[DAS]

Evo ga i prvi mobile GPU...

GTX 480M

Na ovim klokovima karta ce biti sporiji od Junipera (5870 mobile), uz vecu potrosnju, tj potpuno nekonkurentan po svakom pitanju u ovom segmentu, i pitanje je ko ce od OEM-ova uopste biti zainteresovan. Cudi me da nisu cekali GF104.

Vidi ovako: taj 5870 je u stvari downclockovani Juniper i kao takav, nije nista cudno sto je samo malo brzi od dosadasnje najjace nVIdia mobile karte GTX 285M :

http://www.tomshardware.com/reviews/avadirect-clevo-w860cu-mobility-radeon-hd-5870,2615.html

E sad, malo predvidjanja: ovaj GTX 480M je u stvari GTX 465 sa zesce spustenim klokovima. GTX 465 je po onom testu koji je iscureo nekih 10% brzi od GTX 275, znaci nesto na nivou GTX 285. Dalje, GTX 285 (desktop varijanta) je oko 2x brzi od GTX 285M, znaci taj GTX 480M sa drasticno spustenim klokom bi trebao i dalje da bude primetno brzi od GTX 285M dovoljno da prestigne Ati Mobility 5870.

Ali, slazem se da bi GF104 kao novi, mnogo manji chip bio mnogo bolje resenje za mobile GPU. Ovo je verovatno bas zbog toga da ne bi cekali dotle, a sigurno ce i GF104 biti iskoriscen za mobile GPU, bilo kao zamena za ovaj, ili kao dodatak ponudi.

 

[Dusan K]

Vidi ovako: taj 5870 je u stvari downclockovani Juniper i kao takav, nije nista cudno sto je samo malo brzi od dosadasnje najjace nVIdia mobile karte GTX 285M

Pa i napisao sam da je Juniper

Elem, sto se predvidjanja tice... 5870M je ~10, max 15% sporiji od desktop 5770. Ta 5770 je oko 20% sporija od 5830, sa kojim ce desktop GTX465 biti tu negde. Ova mobilna verzija ce raditi na klokovima koji ce joj odneti ~30-35% performansi, sto je stavlja veoma blizu 5870M. Veoma blizu je tu negde u 5-10% gore/dole. Na kraju krajeva, nisu problem za NVidiu u desktopu performanse, vec druge stvari

Ovaj cip je klasicni PR stunt, da mogu da kazu da i oni to imaju, koji ne moze biti konkurentan sa cenom, perf/watt, perf/mm2 ili po kom god drugom osnovu. NVidia ce morati da ih prodaje sa gubitkom da bi mogli da pariraju ceni po kojoj AMD moze da prodaje Juniper (cenim da on u proizvodnji ne kosta vise od 25 dolara).

Kao sto rekoh, i tu se slazemo, GF104 bi bio mnogo logicnije resenje, sem ukoliko oni znaju nesto sto mi ne znamo Sa ovim GTX480M samo ce nastaviti da gube market share u mobilnom segmentu.

 

[__zevs__]

Mobility 5870 ima tdp 50W.
Ovde je jako upitno koliki ce biti za gtx480m, neki su pominjali i celih 100. To je za laptop ogromna razlika (i zbog grejanja i zbog baterije).

 

[Bogda]

http://www.anandtech.com/show/3740/nvidia-announces-gtx-480m
Anand je preneo nVidijinu najavu, napisao je cirka 1000 reči, ishvalio novi 480m ali je jedva u jednoj rečenici pomenuo tdp od suludih 100W za notebook. "Iznenađujuće" ozbiljan tekst za ovu smejuriju od proizvoda.

 

[pixell]

Mislim da je to tačno.
Čini mi se da trenutno najbrži nVidija mobilni čip GTX285M ima TDP od 75W.

Postavlja se pitanje gde nVidija planira da pozicionira mobilnog Fermija? Sigurno neće biti u laganim prenosnicima i segmentu ispod 16 inča, što dovodi do zaključka da je desktop replacement jedino gde i mogu da računaju. Ali opet, sigurno neće biti brži od CrossFireX kombinacije HD5870 koji ima isti TDP (M17x, na primer).

Znači, mala ali odabrana skupina ljudi koji imaju koristi od CUDA-e, Open CL-a

EDIT: Pretekoše me Bogda i Anand.

 

[gile23]

nVidia kao da sam sebe tera u propast pogotovu zbog svog famoznog mobile GPU-a koji je tezak les, prvo se greje, drugo trosi dosta struje, to je za sve samo nije za prenosni racunar (laptop). CUDA i ostatak tehnologije se nesto ne primenjuje preterano u desktop aplikacije, CUDA se najvise koristi u TESLA sistemima, jer tu ceo posao radi GPU. Digli su toliku frku oko te CUDA tehnologije i na kraju po vecini testova ispade da je ATI Stream brzi, stvarno tuga .

Inace ovaj Fermi je okarakterisan kao Fail, ali ljudi ipak kupuju GTX 4xx kartice, evo jedan decko iz mog grada je uzeo iz Nemacke GTX 470 za 240 evra, sto je realno dzabe .

Pozdrav!

 

[zerbah]

Chitam ovakve postove i smejem se - kod nvidije je sve fail, a kod atija cvetaju ruze :d

Shto se tiche igranja, Ati je sada u prednosti, ali shto se tiche svega ostalog josh uvek zaostaje dobrano. Poshto igrom sluchaja imam dosta kontakta sa ljudima koji se profesionalno bave programiranjem na grafichkim kartama, mogu samo da prenesem njihova zapazanja - svi vec neko vreme programiraju u CUDA-i, a sve ozbiljnije stvari se josh rade u OpenCL-u, ne zbog toga da bi radile na Ati karticama, vec zbog toga shto Intel ima 50% grafichkog trzishta...

 

[Nedjo]

Fermi je "ugasio" i na HPC trzistu!

http://www.theinquirer.net/inquirer/news/1650867/hpc-vendors-ready-dump-nvidia

@zerbah
Da Intelova "grafika" koja zauzima 50% trzista radi, ili barem "odradjuje" OpenCL, Apple bi koristio nove LGA1156 procesore umesto starih LGA775 u novim Macovima

 

[nex999]

NVIDIA Tesla GPUs Power New IBM Servers

 

[Dusan K]

Uopste nije glavno pitanje da li ce NVidia propasti... nece, ili makar ne tako skoro (nikad ne reci nikad, i vece kompanije su bivale likvidirane).

Po meni je pitanje kakvu korist prosecan kucni korisnik ima od NVidie koja bi proizvodila samo karte za profesionalnu upotrebu (Quadro, Tesla), cipsetove za mobilne uredjaje (Tegra), ili cak bila softverska kompanija, kao sto rece CEO pre par meseci. Ne kazem da to nisu dobri koncepti (licno sam bio odusevljen kada sam testirao ION), ali je situacija u kojoj ne mozemo da zivimo bez GPGPU-a jos uvek daleko.

A ako se ovakav trend nastavi i u nekoliko narednih par generacija (mnogo veci, mnogo skuplji, jednostavno mnogo ekonomski neisplativiji), nije ni nemoguce povlacenje sa desktop/notebook trzista. To najmanje odgovara vecini od nas, pa cak i onima kojima zaradjuju pare od CUDA programiranja, renderinga i slicnih stvari, jer ce karte kostati 1500+ dolara i svakako biti mnogo nedostupnije.

Moguce da sam preveliki pesimista, ali poslednje dve generacije su stvarno problematicne po mnogo pitanja.

 

[Ace Rimmer]

Tegra je definitivno zanimljiva za Android slate/netbook tržište koje se tek "otvara".
Za ION šteta s obzirom da Intel na polju grafike nema ništa zanimljivo za ponuditi.

 

[DariusIII]

Udarac za nVidia-u i dobitak za AMD/ATI.

AMD Appoints Manju Hegde to Lead Fusion Experience Program

 

[starac]

Jak udarac

 

[Nedjo]

Jak udarac

ma dok je Starac uz NVIDIA-u, nema zime za njih ;-)

 

[starac]

Pa nije Hedge covjek bez kog se ne moze. Ne volim podcjenjivanje, ali da je toliko bitan, ovi bi ga zadrzali sigurno.

 

[ABET]

Pa i ja sam rekao da nije čovek bez koga će nVidija skapati ali s obzirom da je on začetnik cele Physx priče AMD konačno ima perspektivu da uradi nešto na tom polju u bliskoj budućnosti. Činjenica je da je Physx trenutno najveća prednost nV karti u gejming segmentu i izjednačavanje na tom polju bi bio veliki udarac po nVidiu pogotovo ako nastavi sa trenutnom politikom velikih high end čipova (i zanemarivanjem srednjeg i nižeg segmenta).

 

[Rapture]

Pa nije Hedge covjek bez kog se ne moze. Ne volim podcjenjivanje, ali da je toliko bitan, ovi bi ga zadrzali sigurno.

I obrnuto, da je nv toliko bitna on bi ostao kod njih, i tako u krug...