Thursday, February 28th 2019

NVIDIA GeForce GTX 1650 Memory Size Revealed

NVIDIA's upcoming entry-mainstream graphics card based on the "Turing" architecture, the GeForce GTX 1650, will feature 4 GB of GDDR5 memory, according to tech industry commentator Andreas Schilling. Schilling also put out mast box-art by NVIDIA for this SKU. The source does not mention memory bus width. In related news, Schilling also mentions NVIDIA going with 6 GB as the memory amount for the GTX 1660. NVIDIA is expected to launch the GTX 1660 mid-March, and the GTX 1650 late-April.
Source: Andreas Schilling (Twitter)
Add your own comment

37 Comments on NVIDIA GeForce GTX 1650 Memory Size Revealed

#1
GoldenX
Good, let 2GB die in peace.
Posted on Reply
#2
notb
GoldenX said:
Good, let 2GB die in peace.
Let's hope it doesn't. Why pay more for something a particular client group will not use?
I have 750MB GPU memory allocated at the moment (500MB by Firefox for whatever reason). Seriously, 4 GB?

1650 got expensive. I wonder if they'll give us something below in this generation (1630?)
Posted on Reply
#3
Bytales
notb said:
Let's hope it doesn't. Why pay more for something a particular client group will not use?
I have 750MB GPU memory allocated at the moment (500MB by Firefox for whatever reason). Seriously, 4 GB?

1650 got expensive. I wonder if they'll give us something below in this generation (1630?)
Geforce 1629 and 1627 as well as 1628 pro, also 1621 monotanium (1gb ram ddr), and 1622 duatanium (2gb ram DDR2) and 1623 tritanium(3gb ram ddr3).
Also lets not Forget 1625 Pentatanium with 5555 Mb of DDr5 RAM
Posted on Reply
#4
robb
notb said:
Let's hope it doesn't. Why pay more for something a particular client group will not use?
I have 750MB GPU memory allocated at the moment (500MB by Firefox for whatever reason). Seriously, 4 GB?

1650 got expensive. I wonder if they'll give us something below in this generation (1630?)
I think he is talking in context of the x50 level cards. it would be idiotic to have 2gb on a gpu like that in 2019.
Posted on Reply
#5
notb
Bytales said:
Geforce 1629 and 1627 as well as 1628 pro, also 1621 monotanium (1gb ram ddr), and 1622 duatanium (2gb ram DDR2) and 1623 tritanium(3gb ram ddr3).
Unnecessary sarcasm.
Until Kepler we had many chips below the *50. They kept just the 1030 for Pascal and a successor is certainly needed. And since 1650 is positioned higher than 1050 was, they can easily squeeze 2 low-end chips.
It's even more important now since IGP-less CPUs are becoming more popular.
Posted on Reply
#6
GoldenX
notb said:
Let's hope it doesn't. Why pay more for something a particular client group will not use?
I have 750MB GPU memory allocated at the moment (500MB by Firefox for whatever reason). Seriously, 4 GB?

1650 got expensive. I wonder if they'll give us something below in this generation (1630?)
Just use the IGP. This is not intended for light office work, it's both too expensive and too powerful for that.
A 1630ish would be nice thou.
Posted on Reply
#7
Berfs1
Most likely 128 bit, because I highly doubt nvidia would put 256 bit on a 1650 but 192 bit on a 1660 Ti. Also, I doubt it would be 64 bit because it is possible from what I have seen up to 2GB; at 4GB memory sizes you need minimum 128 bit.
Posted on Reply
#8
Bytales
notb said:
Unnecessary sarcasm.
Until Kepler we had many chips below the *50. They kept just the 1030 for Pascal and a successor is certainly needed. And since 1650 is positioned higher than 1050 was, they can easily squeeze 2 low-end chips.
It's even more important now since IGP-less CPUs are becoming more popular.
Oh the sarcasm is very much needed, like bringing in more milkable Plebs into the equation to raise nvidia stock. Having such a wide variety of gpus, for all Ages and Needs, from Kids in the kindergarten which want to Play angry birds in 4k, to granny who wants to check the weather on her wall BFG Display (which she payed for by collecting the life insurance after her late husband died) being powered by nvidia "granny Edition" with hidden specs (grannies dont care about specs), is imperative for Investors, and for the stock Price, as well as market Penetration, making nvidia the undisputable "lider" in the "milking the fulz" industry, so that the board of directors is "plizd", and so that the CEO can get a new "Tasmanian Devil Ledar Jekket", for which a small Population of Devils must be sacrificed, including their Kids, and new borns, and still to be Born specimens, which make fine Buttons, by the way.

Such is the world we live in, my friends !
Posted on Reply
#9
notb
Bytales said:
Oh the sarcasm is very much needed, like bringing in more milkable Plebs into the equation to raise nvidia stock. Having such a wide variety of gpus, for all Ages and Needs, from Kids in the kindergarten which want to Play angry birds in 4k, to granny who wants to check the weather on her wall BFG Display (which she payed for by collecting the life insurance after her late husband died) being powered by nvidia "granny Edition" with hidden specs (grannies dont care about specs), is imperative for Investors, and for the stock Price, as well as market Penetration, making nvidia the undisputable "lider" in the "milking the fulz" industry, so that the board of directors is "plizd", and so that the CEO can get a new "Tasmanian Devil Ledar Jekket", for which a small Population of Devils must be sacrificed, including their Kids, and new borns, and still to be Born specimens, which make fine Buttons, by the way.

Such is the world we live in, my friends !
Honestly, I don't understand this rant, but it might be a cultural barrier of sorts.
You're against Nvidia? Against stocks? Against having a diverse product portfolio?

GoldenX said:
Just use the IGP. This is not intended for light office work, it's both too expensive and too powerful for that.
A 1630ish would be nice thou.
You're using a Ryzen 3. You're an active advocate of the whole Zen architecture.
And you're telling me to use the IGP. That's smart. ;-)
Posted on Reply
#10
SIGSEGV
WTH.
lol.

1660TI 1660 1650 1690 1670 1670Ti 1620Ti 1640?
Posted on Reply
#11
Robcostyle
And I just thought recently....apart from 2080/2080 ti there’s almost 0 improvement in specs, and small perfomance gains. Yet, they raise the prces by x2 for the whole stack, even non rtx cards, or low end 50 card aswell.
That prooves nvidia’s fk-up with mining, and greed from monopoly once again. They could easily release these cards at pascal/maxwell prices.
Posted on Reply
#12
27MaD
GoldenX said:
2GB
It's still kicking if you lower the settings a little bit.
Posted on Reply
#13
trog100
i just loaded farcry 5 onto a 1050ti gpu machine.. at a resolution commensurate with its 21 inch monitor and medium settings it banged out near 90 fps max..

it maintains a steady 60 fps with no problems.. it both looks and plays nicely i was kind of surprised being used to looking at 1440 with a 2080ti powering things..

i think we have all gone over the top with this must use ultra setting kind of thinking that permeates this place..

i know my hardware thinking has definitely gone over the top.. my little farcry 5 with a 1050ti gpu experiment just kind of rubbed it in..

i expected a slideshow but didnt get one..

trog
Posted on Reply
#14
jabbadap
Low profile breed of this card will be the new king for used SFF dell,hp,lenovo etc. PC market. IF it comes without extra power pin. 4GB vram is quite must for 4k video HTPC, and if it has performance near to gtx1060 3GB it will be fairly decent 1080p gaming card.
Posted on Reply
#15
GoldenX
notb said:

You're using a Ryzen 3. You're an active advocate of the whole Zen architecture.
And you're telling me to use the IGP. That's smart. ;-)
Let me get this straight, you want to put an USD 180 card on an office computer, and then complain about the VRAM amount. That's smart.
Also, who said I'm not waiting to the Zen 2 APUs? I got the 1200 to avoid paying extra for the 2200g when it released, a country with huge inflation makes you take some decisions that someone with a stable economy can't comprehend. I paid ARS 1800 for the 1200, the moment the 2200g came out, with the same cost in USD, here costs ARS 3500, and now costs ARS 5500.
Posted on Reply
#16
Imsochobo
trog100 said:
i just loaded farcry 5 onto a 1050ti gpu machine.. at a resolution commensurate with its 21 inch monitor and medium settings it banged out near 90 fps max..

it maintains a steady 60 fps with no problems.. it both looks and plays nicely i was kind of surprised being used to looking at 1440 with a 2080ti powering things..

i think we have all gone over the top with this must use ultra setting kind of thinking that permeates this place..

i know my hardware thinking has definitely gone over the top.. my little farcry 5 with a 1050ti gpu experiment just kind of rubbed it in..

i expected a slideshow but didnt get one..

trog
Amen, people hate on 6 gb rtx2060, why? because internet knows best.
The same people who think HBM on apu's is a smart thing, "it's like saying lets put a GTX1060\rx580 memory system on a GT1030\RX550!"
They also think a 8ghz quadcore could have been made instead of 9900K...
Also, it is impossible to play a game on a RX570 and without freesync\gsync.

Internet is just full of super smart people and thats why these movements start :(
Posted on Reply
#17
notb
GoldenX said:
Let me get this straight, you want to put an USD 180 card on an office computer, and then complain about the VRAM amount. That's smart.
No. I simply want to put a graphics card into an office PC (with a CPU that doesn't have IGP).
And since this PC won't be used for gaming or GPGPU, why would I have to pay for 4GB? I might never use more than 1GB.
Posted on Reply
#18
ONEoo7
Bytales said:
Geforce 1629 and 1627 as well as 1628 pro, also 1621 monotanium (1gb ram ddr), and 1622 duatanium (2gb ram DDR2) and 1623 tritanium(3gb ram ddr3).
Also lets not Forget 1625 Pentatanium with 5555 Mb of DDr5 RAM
Thanks for the free laughs xD
Posted on Reply
#19
GoldenX
notb said:
No. I simply want to put a graphics card into an office PC (with a CPU that doesn't have IGP).
And since this PC won't be used for gaming or GPGPU, why would I have to pay for 4GB? I might never use more than 1GB.
Then an HD7730/R5 240/RX 550/GT 1030/GT 210/GT 710/GT 520/GT 610 is a better choice, either you need a good Nvidia for Cuda, or anything is enough.
Those are all cheap new cards, and with the exception of the Fermi and Tesla ones, all have current driver support.
Posted on Reply
#20
notb
GoldenX said:
Then an HD7730/R5 240/RX 550/GT 1030/GT 210/GT 710/GT 520/GT 610 is a better choice, either you need a good Nvidia for Cuda, or anything is enough.
The models you've mentioned are up to 6 years old. I don't think you can buy a HD7730 today.

Of course GT 1030 is a fine option, but it's a previous gen model. It would make sense for Nvidia to replace it - even if it turns out to be just a sad rebrand.

RX550 is perfect just for video output (likely better IQ than 1030), but the coolers are somehow big and I've seen just a few passive variants. I think they made this GPU too powerful (it's still rubbish for gaming anyway).
It's easier to find something interesting with GT1030 or the mighty GT710.

For acceleration (or conscious GPGPU) Nvidia is a safer choice, because it can do both CUDA and OpenGL. I know you don't like it, but that's the truth. :-)
Posted on Reply
#21
GoldenX
notb said:
The models you've mentioned are up to 6 years old. I don't think you can buy a HD7730 today.

Of course GT 1030 is a fine option, but it's a previous gen model. It would make sense for Nvidia to replace it - even if it turns out to be just a sad rebrand.

RX550 is perfect just for video output (likely better IQ than 1030), but the coolers are somehow big and I've seen just a few passive variants. I think they made this GPU too powerful (it's still rubbish for gaming anyway).
It's easier to find something interesting with GT1030 or the mighty GT710.

For acceleration (or conscious GPGPU) Nvidia is a safer choice, because it can do both CUDA and OpenGL. I know you don't like it, but that's the truth. :)
Of course it's the best choice if you need GPGPU. But you don't need to spend $180 only for displaying a browser, nor you need current hardware for that.
Tell me in what case do you need Mesh Shaders for word processing.
Posted on Reply
#22
Tartaros
notb said:
Of course GT 1030 is a fine option, but it's a previous gen model. It would make sense for Nvidia to replace it - even if it turns out to be just a sad rebrand
The 1030 is 1 year and a half old, don't expect turing refresh for this segment anytime soon. The 710 lasted for a long time, even nvidia skipped a maxwell card for this segment. Just give up.
Posted on Reply
#23
notb
GoldenX said:
Of course it's the best choice if you need GPGPU. But you don't need to spend $180 only for displaying a browser, nor you need current hardware for that.
Tell me in what case do you need Mesh Shaders for word processing.
1) When it comes to GPU acceleration, I have no idea what I need mash shaders for and I shouldn't know. That's the whole point.
2) If you think word processing is what happens in offices, you may not really understand what white collars do for money. :-)
Posted on Reply
#24
GoldenX
notb said:
1) When it comes to GPU acceleration, I have no idea what I need mash shaders for and I shouldn't know. That's the whole point.
2) If you think word processing is what happens in offices, you may not really understand what white collars do for money. :)
If you need a good GPU for real work, a 1650 is too little for it. If you need a GPU for desktop use, a 1650 is too much for it.
The market for the 1650 is games at low resolutions, less than that IGP and low end cards, more than that, Quadro or Titan, with their certified drivers.
Posted on Reply
#25
ArbitraryAffection
notb said:
Let's hope it doesn't. Why pay more for something a particular client group will not use?
I have 750MB GPU memory allocated at the moment (500MB by Firefox for whatever reason). Seriously, 4 GB?

1650 got expensive. I wonder if they'll give us something below in this generation (1630?)
Are you seriously defending 2GB gaming cards... in 2019? Anything with GTX or RTX before the name should be 4GB absolute minimum. GTX 1050 2GB was already being limited by its vram when it launched a couple years ago. 2GB on a card which is going to perform like 1060 6GB in games (if not VRAM bound) is just ridiculous. The GTX 1650 is not aimed at people who want a display output and browse the web. Go get a 1030 for that.
Posted on Reply
Add your own comment