Final Words

For our final video card review of the year, the GeForce GTX 1650 Super admittedly doesn’t bring any surprises. But then it didn’t need to. As the 4th TU116 card released this year alone, and the 12th Turing GeForce card overall, we generally have a good grip on what the current generation of GeForce cards can do. All that’s really needed to change was for NVIDIA to get a better grip on what the market is looking for in performance for a sub-$200 card, and with the GTX 1650 Super they’re finally doing that.

Looking at things on a pure performance basis, the GTX 1650 Super makes significant strides over the GTX 1650, never mind NVIDIA’s last-generation cards. As it should: it offers a very hefty increase in the number of CUDA cores as well as memory bandwidth, and even with the TU116 GPU in its cut-down state for this card, it’s still just a whole lot more GPU than the TU117 in the GTX 1650. NVIDIA has thrown significantly more hardware at this segment, and the net result is a big performance boost.

And just in time as well. The launch of AMD’s rival Radeon RX 5500 XT immediately took the original GTX 1650 down a peg, as AMD’s card offered a lot more performance for only a small increase in price. So one way or another, the GTX 1650 Super is NVIDIA’s counter-play to AMD’s new card.

Performance Summary (1080p)
  Relative Performance Relative
Price
Relative
Perf-Per-Dollar
GTX 1650S vs GTX 1650 +33% +7% +25%
GTX 1650S vs GTX 1660 -11% -24% +16%
GTX 1650S vs GTX 1050 Ti +77% +14% +55%
GTX 1650S vs RX 5500 XT 4GB 0% -6% +6%

The outcome is surprisingly even-handed between AMD and NVIDIA. With regards to framerates, the $159 GTX 1650 Super and the $169 RX 5500 XT 4GB are at a dead heat. As is usually the case, the cards are anything but equal on a game-by-game basis, constantly trading wins and losses, but at the end of the day they’re fighting over the same market with the same performance. NVIDIA has a slight edge in price, and perhaps an even slighter edge in overall energy efficiency, but that’s it. As a result, the GTX 1650 Super is significantly better than the GTX 1650, but it’s not able to meaningfully pull ahead of AMD’s card.

Meanwhile, although the original GTX 1650 is now well outclassed in terms of performance, the card isn’t going anywhere, and for good reason. It remains NVIDIA’s best option for the 75W market. GTX 1650 Super improves on performance by almost 33%, but it requires just as much additional power to get there. For system builds that aren’t sensitive to power/thermal needs, then this isn’t going to matter, and the GTX 1650 Super is well worth the $10 price premium. Otherwise the original GTX 1650 is still the best card available for the 75W, maximum compatibility PCIe-slot-power-only market.

Though like last week’s Radeon review, I’ll also note here my general hesitation with cards that have 4GB of VRAM. VRAM isn’t cheap, and GDDR6 even less so, so both vendors are using VRAM capacity as product differentiators and to upsell their better cards. But as VRAM capacity in the $150-$200 price range has been pretty stagnant for the last couple of years now, I do have some concerns about the long-term implications for 4GB cards, especially with the next-generation consoles set to launch in a year’s time. With the consoles setting the baseline for most multiplatform games, it’s a reasonable bet that VRAM requirements aren’t going to stay put at 4GB much longer.

Unlike AMD, the situation isn’t quite as black and white in the NVIDIA ecosystem, as NVIDIA doesn’t offer an 8GB GTX 1650 Super – or even an 8GB GTX 1660 series card for that matter. So the answer can’t just be “buy the 8GB card” like it is with the RX 5500 XT. Still, along with offering better performance, the GTX 1660 cards and their 6GB of VRAM stand a better chance of delivering solid, unimpeded gaming performance in a year or two’s time. At the end of the day, I don’t think any 4GB cards are a great choice right now; if you can afford to go higher, you should.

Finally, we have the particularities of Zotac’s card, the Zotac Gaming GeForce GTX 1650 Super. Zotac is one of NVIDIA’s most regular and reliable partners, and it shows, with the company producing a card that may just as well be the official GTX 1650 Super reference card. It is as good of an example of the baseline GTX 1650 Super experience as one could hope for, right on down to the fact that it’ll fit into virtually any PC.

But like it’s predecessor, the original Zotac GTX 1650, I remain unimpressed with the coolers on Zotac’s GTX 1650 series cards. We have seen and tested other similar double-slot cards before that are quieter, including cards from Zotac. So for the Zotac GTX 1650 Super land among the loudest of our low-end cards is unfortunate. While I admire Zotac’s commitment to making such a small card, I can’t help but think that, if nothing else, a single, larger fan would have fared better. Low-end cards are always a design challenge in terms of profit margins, but I think there’s room here for Zotac to do better, even for a baseline card like their Gaming GTX 1650 Super.

 
Power, Temperature, & Noise
POST A COMMENT

67 Comments

View All Comments

  • Marlin1975 - Friday, December 20, 2019 - link

    Little disappointing. Lower performance than a 580 and higher priced than others in its price range.

    If the price was quiet a bit lower then it might make a decent HTPC card for some.
    Reply
  • HarryVoyager - Friday, December 20, 2019 - link

    The 8Gb 580's, both new and used, are going to be the elephant in the room for a while I suspect. A reasonably cared for miner card runs only $100 and will keep you going at least until AMD is competitive enough to drive nVidia's pricing down. Reply
  • Yojimbo - Friday, December 20, 2019 - link

    A reasonably cared for mining card...? How do you assure that?

    The 580s have been so cheap because AMD made too many of them. Nvidia are not so interested in low end market share. They'll sell cards there if they can make money on them. I have a feeling GDDR6 dram has come down in price in the last year more than 7 nm fabrication cost has. Now they can sell in that market segment with good profits.
    Reply
  • Retycint - Friday, December 20, 2019 - link

    I would imagine most miners run their cards undervolted/underclocked for maximum efficiency. And it's a 24/7 load, unlike the periodic low->high->low load of gaming cards, which is actually better for the card in the long run. So miner cards actually tend to be in better condition (save for the fans, of course).

    But then again, there's the risk of miner cards that have had modded BIOSes installed which might have damaged the card? So I suppose it's down to buying from a reputable seller
    Reply
  • Yojimbo - Friday, December 20, 2019 - link

    They would have undervolted and then pushed it to get the most performance they could without crashes and run it 24/7. It would be sitting in a hot case in a hot room. I don't think it's a particularly desirable card. But my question is actually in response to what your conclusion was: How do we judge a mining card? What do you mean a reputable seller? A 3rd party that picked up the card fro some miner? Reply
  • PeachNCream - Saturday, December 21, 2019 - link

    GPU mining was typically done with open air brackets rather than in a densely packed case. However, GPU mining is a lot less commonplace these days given the low profitability. We are past that being a thing really so finding a modern ex-miner GPU is not as easy of a prospect as it was even a couple of years ago. Reply
  • Kangal - Saturday, December 21, 2019 - link

    The thing is Mining Cards really are Gaming Cards.
    There's less than 1% performance and thermal difference between the two. Even Linus TechTips did an experiment on this by comparing a mining card that was used for 4 years constantly, compared to the exact same card (same variant) which was still sealed in the box. No difference.

    At worst case scenario, the used card is not going to have warranty, and you may need to "refurbish" it yourself: clean the case, clean the innards, reapply thermal paste, put in new fans. All up it's going to cost you $5-$20. So when you're saving yourself $50-$100 it's worth it, from a value point.
    Reply
  • MamiyaOtaru - Sunday, December 22, 2019 - link

    nobody thinks there is a performance difference between an ex mining card and a new card. Performance doesn't do a slow fade. The worry is that the mining card could be that much closer to failure. Reply
  • Kangal - Sunday, December 22, 2019 - link

    Actually, heaps of people think mining cards run hotter, use more power, and run slower compared to Brand New. It's actually a very widespread misconception.

    Mining cards being closer to failure is actually a myth as well. If the card is running relatively normally when you buy it, a quick refurb will bring it back to New condition. The same applies to a Gaming Card, if you hear the fans are out of tune: replace fans. If it's running hotter than normal: reapply thermal paste. If it looks dusty: clean the innards and casing.

    As was pointed out, Gaming Cards are usually at Idle, then at Maximum Performance, then Idle, then Max again. That process is actually harsher on the components: fan, tim, logic board, and uneven heat dissipation. A Mining Card is usually run cooler and at lower voltage, and its running smoothly and consistently which doesn't cause as much "micro-cracks" in the thermal paste, or the fan assembly. Not to mention, I think subjectively miners are likely to look after their hardware a little better (it's making $) and used in open air, there's some merit to it.

    However, this stigma has been good for enthusiasts for generating a lot of great mining cards at stupid low prices. The old AMD HD7970, I've seen them go for $50 like 5 years ago. The AMD R9 290, was around $150 as well. Now, we've got plenty of ex-mining RX 470 and RX 580's, that's really killing it for people considering a RX 5500XT or GTX 1660. (Un)fortunately there wasn't too many Vega56 or Vega64 cards manufactured to affect the market too much. The Nvidia cards have seemed to keep their value much better (GTX 980, GTX 1070, etc).
    Reply
  • MASSAMKULABOX - Friday, January 3, 2020 - link

    Some mining cards had NO video outputs .. a slight handicap for gaming ?
    Mining cards were in general looked after well ..but home miners ..not so much...
    quite a lot of the miners factored in th resale value of the cards ,, the prices were high new, so s/h prices would also be high. Only when everyone wants to sell their cards at the same time prices drop.
    Reply

Log in

Don't have an account? Sign up now