Wednesday, September 21st 2022

ICYMI, NVIDIA GeForce RTX 4080 12GB Uses 192-bit Memory Bus

Amid the fog of rapid announcements and AIC graphics card launches, this little, interesting detail might have missed you, but the new GeForce RTX 4080 12 GB graphics card announced yesterday; features a memory bus-width of just 192-bit, which is half that of the RTX 3080 12 GB (384-bit). The card uses 21 Gbps GDDR6X memory, which at 192-bit bus-width, works out to just 504 GB/s bandwidth. In comparison, the RTX 3080 12 GB uses 19 Gbps GDDR6X memory, which at 384-bit bus width, produces 912 GB/s. In fact, even the original RTX 3080 with 10 GB of GDDR6X memory across a 320-bit bus, has 760 GB/s on tap.

The bigger RTX 4080 16 GB variant uses 256-bit memory bus, but faster 23 Gbps GDDR6X memory, producing 736 GB/s of memory bandwidth, which again, is less than that of the original 10 GB RTX 3080. It's only the RTX 4090 that has an unchanged amount of memory bandwidth over the previous generation—1008 GB/s, which is identical to that of the RTX 3090 Ti, and a tad higher than the 936 GB/s of the RTX 3090 (non-Ti). Of course, memory bandwidth is no way to compare the RTX 40-series from its predecessors, there are a dozen other factors that weigh into performance, and what matters is you're getting generationally more memory amounts with the RTX 4080-series. The RTX 4080 12 GB offers 20% more memory than the RTX 3080, and the RTX 4080 16 GB offers 33% more than the RTX 3080 12 GB. NVIDIA tends to deliver significant performance gains with each new generation, and we expect this to hold up.
Add your own comment

81 Comments on ICYMI, NVIDIA GeForce RTX 4080 12GB Uses 192-bit Memory Bus

#76
Bwaze



To be fair, Nvidia stated several times the initial RTX 40- series launch is the "Beyond" part, not a replacement of Ampere cards. That's why we only get RTX 4090 at first, then RTX 4080... So there isn't really a push to replace the existing RTX 30- stock anywhere, they can continue to sell therm for months for the same price they were marketing them in 2020.

Posted on Reply
#77
Xuper
1060 6GB vs 1060 3GB , Ring a bell ? a long time ago, My friend bought 1060 3GB , because it was cheap ( before miner explode ) and after I explain it to him, He got very angry. :laugh::laugh::laugh:
Posted on Reply
#78
defaultluser
Xuper1060 6GB vs 1060 3GB , Ring a bell ? a long time ago, My friend bought 1060 3GB , because it was cheap ( before miner explode ) and after I explain it to him, He got very angry. :laugh::laugh::laugh:
don't hate on these partial cuts; the 4080 12gb card has almost half the shaders of the 4090 (and thus, its fine with half the bus) Ampere desktop capped minimum vram size sat 8gb to avoid such blatant starvation

the 100% cutoff in capacity has always been worse than a 25/50% percent drop in both!
Posted on Reply
#79
openbox1980
hatI'm reminded of the 9600GSO with 192 bit and 128 bit variants...
Or the more current 1060 3gb and 1060 6gb versions.
Posted on Reply
#80
trsttte
sumludus1. Set the goalposts out of reach for RX 7000 by leaning heavily on DLSS 3.0 to promote performance uplifts when they know FSR isn't keeping up.
2. Price the cards out of reach to steer consumers toward the glut of 30 series inventory.
DLSS 3.0 seems to undermine your second point since they're claiming it will only work on 4000 series. I think is they continue their BS of keeping DLSS proprietary and exclusive to nvidia cards (now even segmenting on card generations) they are dooming their own technology since the competition can work on anything. They should learn from the old saying if you love something set it free
Bwaze
From that line up, is the 3090/3090ti being discontinued? Could make sense if 4080 outperforms it but it doesn't replace the higher memory capacity/bandwith.

This shenanigans suck!
Posted on Reply
#81
Vayra86
sumludusGiven how much the cache size has increased vs the 30 series, I really don't think the memory bus being 192-bit matters at all. I don't see this as NVIDIA is being greedy by charging too much or inflating the specs of a x70 card to x80 status. To anyone who plans on boycotting the generation after what they saw yesterday, I'll bet NVIDIA only has one thing to say to you; good.

NVIDIA doesn't want to sell RTX 40 series cards.

Don't get me wrong, if you still wanna buy a 40 series, they'll be more than happy to sell one to you. But given what I can only assume are still poor inventory levels (supply chains, yields, etc.), that press conference wasn't meant to sell 40 series cards. NVIDIA has two objectives with this launch.

1. Set the goalposts out of reach for RX 7000 by leaning heavily on DLSS 3.0 to promote performance uplifts when they know FSR isn't keeping up.
2. Price the cards out of reach to steer consumers toward the glut of 30 series inventory.

Accomplishing 1 will make sure they have the undisputed Halo product, which sells more lower end cards by association. AMD was much closer than I bet NVIDIA anticipated they would be in terms of pure rasterization last generation. It makes sense for them to leverage their proprietary portfolio to the fullest where AMD is not comparable.

As for 2, once the 30 series are depleted they will have a lot of wiggle room to lower the prices of the vanilla 40 series. Just in time for a Super refresh.
This person gets it. That's the strategy here, so once more the best game to play is the waiting game.
Posted on Reply
Add your own comment
Aug 16th, 2024 10:27 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts