NVIDIA GeForce GTX 1650 Specifications and Price Revealed

[trog100]

people seem to forget this card is engineered down to 75 watts and no extra plugs.. it will perform as well as 75 watts can perform.. no more no less..

trog

 

[HenrySomeone]

Exactly and it will (once again) blow anything from team red in this segment out of the water (and people forget, it is quite a big segment). God, when was the last time they were even competitive here? 7750 maybe in 2012?

 

[64K]

If it beats the RX 570 then it will probably sell fine but I just checked Newegg and the prices for the RX 570 start at $130 so for someone will little cash to spend that $20 might matter.

 

[Fluffmeister]

This is looking like a gem of a card, half the power of a RX 570, ouch.

 

[64K]

This is looking like a gem of a card, half the power of a RX 570, ouch.

Where I live I have a cheap rate on electricity (12 cents per kWh) but 20 hours a week gaming over the year that extra 75 watts works out to be around $9 a year. I usually keep a card for 4 years so that's $36. So during the life of my card and the cost to run it then the RX 570 is actually more expensive over all than this card.

 

[Fluffmeister]

Again, it's not about the price of electricity, but what it can plug into without much fuss, in turn less heat and less noise.

It's a shame people continue to play down efficiency. Less heat, less noise, money saved.

 

[HenrySomeone]

Precisely - against the already more efficient 1060 3gb the 570 was still able to compete because of its recently very favorable price, but if 1650 with its even lower power consumption and no power plug is even just close to it (which is almost a given, might actually be faster), the Radeon is dead in the water and with it the last competitive AMD card...

 

[overvolted]

It's for the kids.

 

[lexluthermiester]

Exactly and it will (once again) blow anything from team red in this segment out of the water (and people forget, it is quite a big segment). God, when was the last time they were even competitive here? 7750 maybe in 2012?

AMD is competitive currently. Will you please stop trolling everyone?

 

[kings]

If this card become to sit close to the RX 570/GTX 1060 3GB, it will be quite impressive for a 75W card.

Even more impressive, if we take into account that it uses 12nm/16nm. It makes us wonder what this Turing architecture would be capable of with 7nm.

 

[CheapMeat]

Would it be worthwhile from a GTX 1050 Ti?

 

[notb]

Would it be worthwhile from a GTX 1050 Ti?

Unless a specific scenario (e.g. you want to go SSF and you need a card powered from PCIe), it won't make any sense.
It should offer very similar performance.

 

[CheapMeat]

Unless a specific scenario (e.g. you want to go SSF and you need a card powered from PCIe), it won't make any sense.
It should offer very similar performance.

Interesting enough my current 1050 Ti doesn't have a power connector. It's the EVGA GeForce GTX 1050 Ti SSC. I was thinking since improved architecture that the 1650 might have an advantage.

 

[notb]

Interesting enough my current 1050 Ti doesn't have a power connector. It's the EVGA GeForce GTX 1050 Ti SSC. I was thinking since improved architecture that the 1650 might have an advantage.

If you have a 1050Ti without a 6-pin, then it's a no-brainer really.
1650 is a successor for 1050. You could be interested in 1650Ti. Or wait for next generation.

 

[trog100]

Would it be worthwhile from a GTX 1050 Ti?

no there may be tiny gain but not enough to matter..

trog

 

[lexluthermiester]

Would it be worthwhile from a GTX 1050 Ti?

A 1650 will out perform a 1050ti handily. The upgrade is not an expensive one. I say go for it.

I was thinking since improved architecture that the 1650 might have an advantage.

And this would very likely be correct. Granted, such advice is a conclusion based on observations of other Turing Vs Pascal benchmark comparisons, so it might be wise to wait for review & benchmarks.

 

[Valantar]

Would it be worthwhile from a GTX 1050 Ti?

Generally, no. Paying $150 for a minor (<2x) performance boost over a $150 card makes no sense. The value for money in this market is already terrible, but making single-generation upgrades Judy makes it worse. It'll be faster, but by no means enough to justify the cost.

If you have a 1050Ti without a 6-pin, then it's a no-brainer really.
1650 is a successor for 1050. You could be interested in 1650Ti. Or wait for next generation.

Is an upgrade to the next generation a no-brainer? Really? Paying as much money again within a short time frame for a minor performance upgrade? I'd say if you have that easy access to money, go up a tier. If not, wait for the next generation to come around and get a much bigger performance increase for the same money.

Interesting enough my current 1050 Ti doesn't have a power connector. It's the EVGA GeForce GTX 1050 Ti SSC. I was thinking since improved architecture that the 1650 might have an advantage.

Sure, it'll be faster. It has more CUDA cores, higher clocks and better performance per core per clock. But will it be fast enough to matter, and be worth the outlay? Not likely. If your idea of value is paying the same again in a couple of years for a 30-40% increase in performance, you should raise your expectations. Your money is worth more than that. Saving up for a while and either moving up a tier or holding off for the next generation is likely to be much better value. In general, upgrading to the same tier in the next generation is a bad proposition, and holding off for the next generation again is much smarter.

Also, most 1050 Tis come without PCIe power connectors. It's a 75W card, so only a few models go that route.

 

[lexluthermiester]

It'll be faster, but by no means enough to justify the cost.

As the cards have yet to be released, you can't and don't know that. However, Pascal Vs Turing benchmarks have already shown the Turing GPU's beat out Pascal GPU's by a solid margin. An upgrade from a 1050ti to a 1650 would very likely be a good move especially when you factor in the user selling the 1050ti after to recoup some of the cost.

 

[Valantar]

As the cards have yet to be released, you can't and don't know that. However, Pascal Vs Turing benchmarks have already shown the Turing GPU's beat out Pascal GPU's by a solid margin. An upgrade from a 1050ti to a 1650 would very likely be a good move especially when you factor in the user selling the 1050ti after to recoup some of the cost.

Of course we can't know the exact level of performance, but we do have a decent idea of ballpark performance. And, as with any upgrade, paying full price for an upgrade that's significantly below a 100% performance increase within a year or two is poor value, even counting recouped value from selling the old card (the price of which will crash when this launches). Conventional wisdom on GPU upgrades is quite decent: wait two generations or upgrade to a higher tier. Anything else gives a very poor return on investment (regardless of recouped costs from sales), particularly when intergenerational improvements are unexceptional as with Pascal to Turing. The performance increase from a 1050Ti to a 1650 is likely in the "fine and noticeable but doesn't change much" class. That isn't worth the upgrade IMO, as the change in user experience will be very minor. If a 1050 Ti is giving you 40fps in your favorite game, this likely won't reach 60, and if you're at 60, why upgrade?

 

[trog100]

i would be very surprised if it gives 10 fps over a 1050ti.. not worth spending the money on for the small gains if you already have a 1050ti....

trog

 

[lexluthermiester]

i would be very surprised if it gives 10 fps over a 1050ti.. not worth spending the money on for the small gains if you already have a 1050ti....

That's like saying that the 1080 only got 10FPS over the 980ti. It's total rubbish. And as the 2060 kicks the crap out of a 1060, it's a safe bet the trend will continue with the budget models. A 1650(on paper) cleans a 1050ti's clock, pun intended.

 

[notb]

Is an upgrade to the next generation a no-brainer? Really? Paying as much money again within a short time frame for a minor performance upgrade? I'd say if you have that easy access to money, go up a tier. If not, wait for the next generation to come around and get a much bigger performance increase for the same money.

I was one of the people that suggested not upgrading for reasons you've mentioned.
You make the effort of writing a comment, you could also spend a few seconds reading what you're quoting.

 

[trog100]

That's like saying that the 1080 only got 10FPS over the 980ti. It's total rubbish. And as the 2060 kicks the crap out of a 1060, it's a safe bet the trend will continue with the budget models. A 1650(on paper) cleans a 1050ti's clock, pun intended.

Maxwell to Pascal was special.. Pascal to Turing not so special.. but watt for watt i dont think there is a great difference..

trog

 

[lexluthermiester]

Maxwell to Pascal was special..

True.

Pascal to Turing not so special..

Not by my experience. I went from a 1080 to a 2080 and the difference was impressive and worth the cost. The OP will be equally well served by going from a 1050ti to a 1650. The performance numbers are smaller, but the performance gap will be similar and will give them a reliably solid 60FPS at 1080P, unlike the 1050ti which does not.

 

[John Naylor]

Not going to pay attention until it's released, and we see how it stacks up with real numbers against AMDs offerings with all cards overclocked.