http://videocardz.com/59459/nvidia-pascal-editors-day-next-week
>NVIDIA Editors’ Day 2016 will take place next week in USA. It’s a special event for reviewers and technology enthusiasts who are invited for a briefing about new graphics cards. During this day NVIDIA will inform the press about new Pascal-based graphics cards.
>The last major Editors’ Day took place in 2014, a week before Maxwell GM204 announcement. So this should give you a perspective on when to we expect GeForce Pascal cards (I’m thinking maybe second or third week of May).
IT'S HAPPENING
You know after everything's been leaked over the last month I don't hold much enthusiasm hearing Nvidia say the same.
I am so fucking sick of waiting and even in a few months both Nvidia and AMD won't release any decent card.
This FUCKING SUCKS
Is this a bad time to buy a new graphics card?
>>54281696
yes, use the integrated gpu if nothing else until the end of the month
>>54281719
thanks for the advice, padre
Does EVGA usually offer a step-up option as soon as a new GPU drops or will it be a while (say, 1 month)?
>>54281846
it's always a bad time to buy technology. the depreciation is abysmal.
>>54277542
y'all niggas need to learn to use a containment thread.
next time you want to shitpost about your gayme asics pu /gcg/ -graphics card general
into the subject field
>>54281696
wait for the unbelievable deals on prev gen stock wen dey drop
>>54277542
AMD BTFO
Time to watch that motherfucker sink
>>54282605
amd is releasing GPU's within 2months too as far I understand it.
>>54282727
But those GPUs suck. Even AMD admitted that they can't touch the 1070 and the 1080.
>>54282762
As far as I'm consirned they only need to beat them with price. I dont feel like paying $300+ for a freaing 1070, let alone $600 or more for a 1080
>>54281925
Good time to buy core components.
Bad time to buy GPU's an SSD's specifically though
>polaris 11 is for mainstream laptops
>polaris 10 is for mainstream pc and notebook
AMD B T F O
>>54283210
Explain? Shouldnt we wait atleast for zen to buy core components?
>>54284522
>not even Haswell level of IPC
>8 core will cost more than the FX-9590
That's a solid NOPE even before it's on sale
>>54285678
Because you know both its IPC and cost, right.
>>54285702
Not him, but I'm under NDA from AMD but I already know Zen is 15% slower than Haswell per clock and only goes up to 3.9GHz
You're better off buying a 6 core IvyBridge EX for dirt cheap.
>>54285721
Btw my dad works for nintendo
>>54277542
>been saying Pascal is due Q2 2016
>fucking AMDumbs posting [citation needed] erryday because they can't take getting BTFO
Eat shit AMDicks
Keep waiting forever for dogshit.
>>54285721
Not him, but I'm under suicide watch because I'm an AMD fanboy but I already know Zen is going to be a big disappointment like Bulldozer was.
You're better off using a shotgun because helium tanks don't work anymore since they started filling them with 50% nitrogen gas
>>54282605
Time to short AMD stock again :3
>>54283229
>AMD's mainstream performance cards are to be on par with GTX 980Ti
AMDKEKS BTFO
>>54287742
people actually believe this.
kek
>>54287763
investors seemed to believe it
>>54285785
Stop saying short instead of put. It's retarded and there's no explanation for it.
>>54287763
>process shrink and clock it to the moon
>claim victory for a couple weeks until nvidia drops another massive die at you to snatch back the crown
https://www.youtube.com/watch?v=_U8c1_yWS10
>>54288141
>nvidia drops another massive die
Pascal will be titan priced.
Basicly Pascals will be the new Titans and teslas.
AMD promisses to give us 350$ VR capable cards.
If that's the case who even wants that titan when you can get two polarises or a PRO DUO? That would be stupid.
Something's telling me that the future is about modular GPU and AMD will get there first.
If it becomes possible to make a GPu outta many small dies and connect them it will be an absolute fucking breakthrough. and we get to fucking 8k the next day.
And this is what AMD already does for server ZEN modifications while Intel and Nvidia are trying to make massive dies AMD makes many cheap small dies and puts them together.
Modular scalable architectures will be fucking amazing.
>>54288234
>If it becomes possible to make a GPu outta many small dies and connect them it will be an absolute fucking breakthrough.
That's what I'm expecting in future. HBM seems like a bit of a beta test of the idea - if you can mount ram on the passive interposer and use very short wiring to hook it up, why not go further and hook up two gpu modules and maybe even share the ram between them?
nvidia could pull the same shit since they have nvlink now, even a pair of gpus on one board might work better using that
exciting times senpai
>>54288234
The only way to move over the 600mm2 reticle barrier is by putting multiple dies on an interposer.
That, or something other than silicon.
>>54283210
GPU I get but what about SSDs?
I mean the prices are in steady decline, but they probably will be for a while, so you can wait years on end.
>>54288133
short selling is not the same as a put option numbnuts.
>>54288640
The only way multiple dies will work is if they can make the VRAM fully coherent to all of the GPUs.
Granted, AMD has been working towards full memory coherency via their HSA models and APU design. But this only applies to CPU > GPU < CPU crosstalk.
This would mean they need to implement either a complete crossbar switch or very fast bi directional ring bus system between all GPU dies. Another thought on my mind is the possibility of a singular memory distribution hub which all dies can access via their independent memory controllers - this would also imply a greater and faster l2$ system as well as expanded ACE functionality, in a sense.
This mayyyyy be what they hint at with "NexGen memory" in reference to their Navi slide. If these VERY EARLY rumors are true.
>>54290154
>The only way multiple dies will work is if they can make the VRAM fully coherent to all of the GPUs.
this is the exact opposite of correct.
memory interfaces on remote NUMA nodes are OK to write to if you have sufficient interconnect bandwidth, but they murder latency for reads.
what is needed is a standard mechanism, hardware or software, to keep mirrors of working sets local to each die, and pool cold assets across the set.
If each die is maybe as powerful as Fiji/GM200 with 2x HMB2 interfaces (with 8/16GB and 500GB/s per GPU), you really wouldn't need to blow a big proportion of die space on super-wide GPU-to-GPU interconnects nor would redundancy between working sets be a big deal.
>>54290154
Also, NextGen memory is very likely to be 3D stacked, to further increase bandwidth and cut latency.
GP100 only has 4k shaders and a 4MB L2 but 14MB in aggregate register files in order to juggle a shit ton of warps/wavefronts due to memory stalls.
GPUs are "embarrassingly parallel" because there tends to be virtually zero concurrent mutable access to data, so loading up on coherency capacity will do very little good at great expense.
>>54288234
You can already do VR with $300 cards.
Once again, stop failing for the marketing you idiot dumbass, they always word those promises poorly to trick idiots like you and get away with it.
>>54281925
Especially right before new architecture releases. nVidia halted production on 9 series GPUs already, so they're clearly preparing for a new GPU release. Even if it doesn't actually happen until August it's still worth the wait because as soon as it releases 980 Ti and 980 should drop to $3-400 as the new 1070/x70 will be at least on part with the 980, if not a 10% better performer likely for $4-500 with the 1080/x80 being closer to $6-700
>>54288133
Shorts Ellington is sale of stocks prior to a release of new tech for low cost to drop stock prices and get what you can before it's too late nigger
do I sell my 290 or just keep it? it does everything I need and more, but I'm tempted to buy a 490 when they release
>>54282006
>unbelievable deals on prev gen
maybe $20 off, $50 off if you're looking at 980/980ti. deals won't be that great 2bh
still should wait though. it won't hurt
>>54291375
>wanting to trade in a housefire for a 380X-tier card
AMDrones everyone
>>54291388
>AMDrones everyone
I've owned 2 AMD and 1 Nvidia, my only loyalty is to my wallet
lol amd rebranded again, nvidia have an EZ win
>>54291403
So you've made two mistakes in your life?
Don't make a third. Buy Nvidia.
>>54291434
>3 year old 290 is comparable to current gen $500 Nvidia card (in Canada)
Nvidia doesn't provide me with enough benefits to outweigh the 30% price premium
>>54291434
Seriously, AMD BIOS and drivers always take at least 6 months to be unfurled after release. If you're going to get AMD wait till 2017 dude.
Or get nVidia and be good to hook out the gate
>>54291473
>Nvidia doesn't provide me with enough benefits to outweigh the 30% price premium
I can name a few
>better drivers
>better game support on Day 1
>better experience and functionality with GeForce Experience over the buggy, featureless Crimson
>lower power consumption
>less heat
>better quality since AIB partners invest better parts on aftermarket Nvidia cards than AMD cards that barely sell at all
>more SLI support
The list goes on. There is literally zero reason to own an AMD card, and it shows. Nvidia owns 80% of the market for a very damn good reason.
>>54277542
I'm still rocking a nvidia 660
what should I wait for? I want an upper mid level card
>>54291513
If only AMD would rock some corporate espionage vs. Intel and nVidia to unfuck themselves and come up with some worthwhile shit day 1 for once as well as not slashing their R&D budgets constantly
>>54291513
Don't forget Shadowplay.
The fact that NVFBC pulls frames straight out of the buffer means zero input lag which is essential for recording first person shooters. Last I checked AMD uses Raptr which, like most recording methods, uses API injection which causes overhead and input lag.
>>54291513
>>better drivers
never had a problem with AMD's
>>better game support on Day 1
true, but I don't buy day 1 because I'm poor/cheap, hence why I use AMD
>>better experience and functionality with GeForce Experience over the buggy, featureless Crimson
again, no problems
>>lower power consumption
I use my pc maybe 300 hours a year, so the efficiency doesn't matter much (I leave my pc at home when I go away for university)
>>less heat
air conditioning
>>better quality since AIB partners invest better parts on aftermarket Nvidia cards than AMD cards that barely sell at all
Sapphire, MSI, and PC all make great cards
>>more SLI support
if I can't afford one card, why would I ever consider two? also
>SLI
>still needing a bridge
lmao 2010 here
>>54291550
>Shadowplay.
I don't record gameplay so it doesn't matter
all of these are genuine benefits of Nvidia, but like I said, they don't apply to ME
>>54291609
>tl;dr - "i never had those problems/i never use those features/etc"
Typical AMDrone dismissing legitimate and market-sensitive concerns. This is why no one likes AMD, and yet you cover your ears to the criticism. Enjoy your dying brand.
>>54291653
see >>54291609
>all of these are genuine benefits of Nvidia
I appreciate the benefits, but none of them apply to me, you dense retard
>>54288769
Yeah, it struck me as odd advice as well. On the flip side, there has never been a better time to buy RAM than today.
>>54291653
>AMD Dying
Believing that would be great. Yeah no, enjoy your 1000$ 1170 with 13.5/16gb Vram and getting butfucked like crazy by huang
>>54291713
This is sad. You're trying to justify your poorly-thought-out purchase by making excuses as to why you don't need X or why Y isn't a problem for you. The sad reality is, you're not getting the better experience compared to better informed consumers who purchase Nvidia cards for their infinitely better driver support, who don't have to waste more money on electricity (turning on the AC when your PC is running? really?), and can afford to buy better quality cards. ASUS and Gigabyte skim on higher quality capacitor and transistors for their AMD cards, while the top-tier Nvidia cards get the best material and components possible (coil whine, anyone?).
>>54291790
>Better informed consumers
>Yet buying the 970 expierence like no tommorow despite getting seriously skimmed by NV
All these Nvidia shill posts are fucking delicious to read. They know AMD will stomp Nvidia with the new cards, offering better performance, better prices and better longevity across the board, so the shills come out in force and spread around misinformation and lies, hoping to paint AMD in a bad enough light that idiots will continue buying the weaker Nvidia GPUs.
>all this video card enthusiasm
How many of you guys actually spend money on new graphics cards, and how many of you are just shitposting?
How many hours a week do you spend playing games, in order to justify the cost of a flagship card?
>>54292484
My 780s are nearly three years old now, it's time for an upgrade. Want to go to a single card because SLI a shit.
Plus I've got money and it's a tax deduction since I use this pc for contract work
>>54288234
>Pascal will be titan priced.
This is what AMDFans actually believe. Nvidia is gonna price themselves out of the market guys! They're definitely not gonna do what they've always done and make smaller more cut-down versions of big Pascal for the consumer market!
>Buy high end AMD card 3 years ago, enjoy playing new triple A titles on very high for 3 years
>Buy more expensive Nvidia card 3 years ago, enjoy shit performance and purposefully gimped drivers
Remember - with Jews, you lose.
>nvidiashillz.com
