The sources said that Nvidia is slated to launch GTX 1660 on March 15 and GTX1650 on April 30...

digitimes.com/news/a20190226PD208.html

>The sources said that Nvidia is slated to launch GTX 1660 on March 15 and GTX1650 on April 30, which will bear minimum price tags of US$229 and US$179, respectively.

THANK YOU BASED NVIDIA

nvidia.com/en-us/geforce/news/geforce-gtx-1660-ti-advanced-shaders-streaming-multiprocessor/

Attached: geforce-gtx-1660-ti-architecture-comparison.png (3340x1506, 382K)

Other urls found in this thread:

amazon.de/Sapphire-11276-Grafikkarte-Radeon-Express/dp/B078NPPX3H/ref=sr_1_1?ie=UTF8&qid=1551189076&sr=8-1&keywords=vega 56 sapphire pulse
amazon.de/PNY-GeForceĀ®-Gaming-Overclocked-Grafikkarte/dp/B07NXVKDLW/ref=sr_1_3?ie=UTF8&qid=1551189103&sr=8-3&keywords=gtx 1660 ti
newegg.com/Product/ProductList.aspx?Submit=ENE&N=-1 8000&IsNodeId=1&Description=Vega 56&bop=And&order=BESTMATCH&page=1
twitter.com/AnonBabble

I bet you the 1650 will have 3-4 GB of v-ram

3.5

Hoping for a SFF 2050ti to replace the 1050ti in my optiplex.

Shill

Going by past actions, more like:
2gb, 3gb and 4gb variants with a special 3.5gb for the china market

>a regular 1070, but with 25% less ram, being sold as something new and interesting
Ladies and gentlemen: nVidiots.

And AMD still can't compete.

Except Vega 56, an 18 month old card, is faster for the same money.

Yawn 7nm 4xxx series when?
Im assuming their still skipping gens starting with 3 since geforce 3 has bad memories

>is faster for the same money

He fell for the Vega price drop meme.

who gives a shit
nvidia is dead in 1 year when intel releases.
let's face it, nvidia has always been the fucking worst.

>1060
>4.4 tflops of fp16
sure thing bud

Nvidia nigger cope

Nope, AMD are the worst.
>overpaid for ATi only to dismantle them
All original ATi staff works for nvidia, apple or qcomm now.

Promote to FP32 and bingo.

>being this retarded

is this gonna knock any money off the price of the RX 570 I was planning on buying?

Hopefully, who knows, but honestly, RX500 cards are heavily discounted it seems, and with the game bundle, you could sell the games to a friend, family member, or some randy online for a quick buck and drop the price of the card even further.
Even if you sold the games for $30 (RRP is ~$60 USD), you're still getting $60 off the card because of getting 2 games with the RX570 & 580, or 3 games ($90, assuming you sell the games for half their retail price) with the RX 590 & Vega cards.

There was only one Vega at that price with that shitty blower cooler and it was only during a limited time sale, all other decent variants were and still are at 350/400+

No.

Literal fake news. The Sapphire Pulse model has also been down at 1660 Ti pricing in the first world.

Attached: Sapphire-RX-Vega56-Pulse-Front-Bottom.jpg (2560x1151, 641K)

No, AMD has no reason to.

That model is at 420 where I live: amazon.de/Sapphire-11276-Grafikkarte-Radeon-Express/dp/B078NPPX3H/ref=sr_1_1?ie=UTF8&qid=1551189076&sr=8-1&keywords=vega 56 sapphire pulse
Whilst the GTX 1660 Ti can be found at 305: amazon.de/PNY-GeForceĀ®-Gaming-Overclocked-Grafikkarte/dp/B07NXVKDLW/ref=sr_1_3?ie=UTF8&qid=1551189103&sr=8-3&keywords=gtx 1660 ti

newegg.com/Product/ProductList.aspx?Submit=ENE&N=-1 8000&IsNodeId=1&Description=Vega 56&bop=And&order=BESTMATCH&page=1

AYYMDPOORFAGS STILL LYING ABOUT POOGA 56 PRICE

THERE IS NO VEGA 56 AT THE FAKE PRICE AYYMD BEEN SCAMMING ABOUT

This its so depressing. This is how s3 matrox sgi and
3dfx fags must have felt was slightly before my time (got my first real gaming pc in 99)
Amd ironically lost its last retard savant engineer (raja cucpoopoocurry aka gcn fury Vega hbm fetishist fucktard) to Intels gpu division
2021 is going to be a all out shit fight I hope my old 1080 lasts until then

>Amd ironically lost its last retard savant engineer (raja cucpoopoocurry aka gcn fury Vega hbm fetishist fucktard)
dGPU R&D is hundreds of people, not a single man.

I did say he was a idiot savant not a one man gpu wrecking machine.
Maybe he was just made the figure head to take the blame then sail into the sunset at Intel?
Still I wonder what they have cooked up my bet they are waiting to launch it on their own 10nm or 7nm.
Fuck seeing a 14nm+++ Intel gpu based or larrabee it would be a 600watt monster

I said first world, not third world caliphate.

>I did say he was a idiot savant not a one man gpu wrecking machine.
There's no 'one-man wrecking machines' in semis.
It's always a gigantic, really elaborate team effort.
>Maybe he was just made the figure head to take the blame then sail into the sunset at Intel?
Intel just offered him the infinite money so he went there.

amd has better lincucks drivers, that's already enough for me to switch, what would I even use an nvidia card in linux for?
>dxvk
just virtualbox that shit pajeet.

>what would I even use an nvidia card in linux for?
Machine learning.

> 20% FP32 increase
> 40% transistors count increase
What the shit

I remember autists saying back in the larrabee days before it got canned that it isn't fixed function and it could technically emulate any gpu(cpu/apu) past present of future and add new shit like dx12 dxr & dx13 + whatever comes next in software? If that's true that would be fucking amazing

What's the reason for this? I'm not really familiar with turing at all I tuned out until they can find a use for rt cores outside of wasting render time on pointless shit in vidya

sounds like pure pie-in-the-sky bogus

>I remember autists saying back in the larrabee days before it got canned that it isn't fixed function
Still had TMUs. so no.
>f that's true that would be fucking amazing
That's a very bad idea actually.
No, Larrabee was a genuine thing.
Just a bad one.

RT cores are FP16 and it's roughly the same per CU as with RTX lineup. That means RT cores are blocked for gaming, not cut out.
> What's the reason for this?
Beats me. Wonder if it means Nvidia is pulling Intel on their R&D or it's just a TDP restriction.

TU116 has no RTRT cores.
>RT cores are FP16
They aren't, these are FF blocks responsible for BVH traversal and ray intersection calculation.

lmao ok nerd

Nigger.

Good. 4GB is all that is necessary for my uses. If I wanted more VRAM, I'd get a more expensive card.

rofl u mad?

Double nigger.

lol so mad

GTX 1670 or 1680 when?