• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

NVIDIA RTX 50 SERIES - Technical/General Discussion

Those are qualified as a personal electonic device. A video card is most definitely not.

VAT is still applicable on all items over £390 even if declared as gifts or for personal usage. The savings have been approx. £400 so far. Would be another £3-400 for a 4090, depending on which state it's bought in. But again I'm not bringing a GPU back this time, but if I were I would run the gauntlet again.
 
Last edited:
Yes and no. It's much faster in very simple, low accuracy AI models - like the ones used in DLSS4 etc. where performance matter more than accuracy/quality, as you need to run those hundreds of times a second. With proper AI models, that take seconds+ to process, which are used widely today outside of gaming, it's at most just about 30% faster than 4090 (as benchmarks revealed), so not really that big of an improvement. It has some other (non-AI related) improvements for Pro users over 4000 series, like new video encoders etc. but that's a niche usecase, comparing to numbers of gamers. Pro users will get it mostly because they can't get new 4000 series anymore, so they won't have much of a choice. The moment 6k series comes, they'll move on to that and not to used 5k series, though. Hence, IF 6k series is much better in gaming (at the moment nothing suggest it will be) then 5k could lose value rapidly (so would 4k at that point), as Pro users do NOT buy used GPUs usually, they go for new ones with full proper warranty and long-term support.
Perhaps I could have been more precise with my original statement. The 32GB is the major factor in it's AI appeal. Even if it was just a 4090 with 32GB of VRAM, it would be more attractive for AI use and get gobbled up with a higher pricetag. As I'm sure you know, the more VRAM, the larger the model you can load in to said VRAM, the more power you have at your disposal. Be interesting to see what the 60 series will even be at this point, let alone how much it will cost. It's going to be more and more dificult for Nvidia to justify selling product to the gamer market when the markup for datacentres/AI companies will be so much higher.
 
I'm starting to think that most people, including many with high-end PCs, don't care about latency. It's no wonder Nvidia are going down this route.

Are there any mainstay competitive games out there that don’t already give us incredibly high native FPS and thus lower latency?

I look at all the esports games out there and they all run at high FPS so I doubt MFG is aimed at the online multiplayer crowd.

Even when they start marketing this heavily to that crowd (eventually) I’m assuming Reflex 2 will address the latency issue…
 
Are there any mainstay competitive games out there that don’t already give us incredibly high native FPS and thus lower latency?

I look at all the esports games out there and they all run at high FPS so I doubt MFG is aimed at the online multiplayer crowd.

Even when they start marketing this heavily to that crowd (eventually) I’m assuming Reflex 2 will address the latency issue…

It's definitely meant for single player games right now, and frankly even with the high fps native games you're still reliant on the server tick rate, of the game you're playing and your own low latency connection to the server to have a beneficial advantage.

I just enjoy the motion clarity of the 480Hz experience when available.
 
I'm quite torn which 5090 option to go for (putting aside stock issues). I'm planning to install a waterblock to install it in my custom loop. Originally I was aiming for the TUF (alphacool will have a block for this) and according to reports, I could RMA if it goes faulty (not if its water damage, etc.). However, having read thru the various coil whine shennanigans of the 4090, and the fact that this will be watercooled, it'll make coil whine even more prevailant. The alternative I was thinking was Gainward. This will have both an alphacool and watercool block as it's reference PCB. According to the warranty stickied post, they don't have as good RMA as Asus, and also it's of a 'lower quality' build. I don't know how they handle having installed a custom block on it (of course putting back the original if need to RMA). The other plus of the Gainward, is price, as I don't really care about the aircooling capabilities, and I can always flash a custom vbios. Opinions appreciated!
 
Last edited:
Hi, does anyone have an idea what time on the 30th should I be pressing F5 like a maniac on the OC'ers GPU webpages?

I managed to order my 4090 Strix on the 12th October 2022 (release day) at 10:48, well before they were due to be released at 14:00. Had it delivered the following day, first time for me to be the first 4090 owner on Kaap's owner's club 'roll of honour'. One for the bucket list :cry:

Mistakes can and do happen at popular retailers, I'd recommend having many browsers/profiles setup with pages to the specific card or cards you want and check them as often as you can that day.
 
I'm starting to think that most people, including many with high-end PCs, don't care about latency. It's no wonder Nvidia are going down this route.

It's becoming a movie

games have become more like movies and now our GPUs are too; soon you won't even need to play your game, just watch it play itself, at least it will be 1000fps
 
Last edited:
I've been gaming for decades now and never once though about latency. The games I've played have always felt responsive to me.

It's definitely something I don't want to become sensitive to lol.
 
I suppose it will get to the point that if you want flashy graphics and high FPS, you may as well just scubscribe to Geforce Now. The latency probably won't be very different to MFG and at least you don't have to fork out 1 or 2k to get access to a decent card.


cover4.jpg
 
AD202 to GB202 changes from die

  1. TPC = 2 SM: -6.09%
  2. GDDR PHY 32 bit: +2.54%
  3. L2 2MB Cell: -13.49%
  4. Mem controller 64 bit: +54.38%
  5. Raster Engine/3DFF: +31.89%
  6. Command frontend: -46.77%
  7. PCIe Phy: -24.43%
  8. 5-line display: +23.52%


This explains why the 5090 hates low resolutions and sometimes loses to the 4090
 
Last edited:
Not really. The odd one like Doom, Wolfenstein.

Primarily horror, cinematic third person, RPG's etc.

I always play with a controller as well if that makes a difference.
Controllers make a HUGE difference. They are such an imprecise control device versus a mouse that it masks a lot. Plus FPS games are where you would notice high latency the most. So i think you are safe :p
 
I don't really know all those brands and models so it is unfortunate to only see the Astral and Suprim having their reviews done. They're flagship models and I kind of dont want to pay more than 300EUR max if I actually go for an AIB at all. The Astral is especially ridiculous with it, but both are out of my pricerange.
 
Back
Top Bottom