GeForce RTX 50-Series with Blackwell Architecture Gets more specific Memory interface info

Published by

Click here to post a comment for GeForce RTX 50-Series with Blackwell Architecture Gets more specific Memory interface info on our message forum
#6214573
https://forums.guru3d.com/data/avatars/m/239/239175.jpg
So if you want a 16GB/256-bit 5070 Ti, you'll again will have to wait for the "Super" refreshes. NV has mastered the art of milking its customers.
#6214580
data/avatar/default/avatar20.webp
Give me RTX 5080 15% faster than 4090 for 1200 Eur and I will be happy.
#6214583
data/avatar/default/avatar37.webp
RealNC:

So if you want a 16GB/256-bit 4070 Ti, you'll again will have to wait for the "Super" refreshes. NV has mastered the art of milking its customers.
Huh, this is about the 5000 series GPUs. The 4070ti SUPER with 16 GB memory has already been launched in January this year, so your remark seems quite out of date anyway.
#6214585
https://forums.guru3d.com/data/avatars/m/239/239175.jpg
Crazy Joe:

Huh, this is about the 5000 series GPUs. The 4070ti SUPER with 16 GB memory has already been launched in January this year, so your remark seems quite out of date anyway.
Typo.
#6214588
https://forums.guru3d.com/data/avatars/m/246/246088.jpg
The HSF being the size of a shoebox, that's the only thing that's a guarantee.
#6214605
https://forums.guru3d.com/data/avatars/m/186/186805.jpg
RealNC:

So if you want a 16GB/256-bit 5070 Ti, you'll again will have to wait for the "Super" refreshes. NV has mastered the art of milking its customers.
Me: https://media1.giphy.com/media/uP781D6zVOdG0/200w.gif?cid=82a1493blmu1szhu4vgb2tmy0axakw94ggotev6zazhu9uq8&ep=v1_gifs_related&rid=200w.gif&ct=g Also me: https://media1.giphy.com/media/X2xRGTElqdfry/200w.gif?cid=6c09b952xvsxl0ma5rsrpjpb0ufj2zwcyzhzq8yhxj7ik3h9&ep=v1_gifs_search&rid=200w.gif&ct=g :(
MonsterMX99:

Give me RTX 5080 15% faster than 4090 for 1200 Eur and I will be happy.
15%??? You are hopeful. 4090 is like what 30% faster than the 4080 now, and you want 15% more than the 4090 for the 5080. That would mean you want the 5080 45% faster than the 4080..... this is Nvidia we are talking about here bro. You will get 20% more performance for 50% more money and like it.
#6214609
data/avatar/default/avatar31.webp
3080 was 50% faster than 2080, 4080 is 50% faster than 3080 so stop talking BS.
#6214612
https://forums.guru3d.com/data/avatars/m/198/198862.jpg
I expected them to bump memory configurations of midrange cards like 5060/5070. Wasnt gddr7 gonna change that?
#6214613
data/avatar/default/avatar25.webp
CPC_RedDawn:

15%??? You are hopeful. 4090 is like what 30% faster than the 4080 now, and you want 15% more than the 4090 for the 5080. That would mean you want the 5080 45% faster than the 4080..... this is Nvidia we are talking about here bro. You will get 20% more performance for 50% more money and like it.
What? When a new x80 card wasn't faster than previous gen top card? If that happens with 5080 it will be the worst x80 card ever.
#6214617
https://forums.guru3d.com/data/avatars/m/216/216349.jpg
RealNC:

So if you want a 16GB/256-bit 5070 Ti, you'll again will have to wait for the "Super" refreshes. NV has mastered the art of milking its customers.
True and AMD is dying to graduate too...
Undying:

I expected them to bump memory configurations of midrange cards like 5060/5070. Wasnt gddr7 gonna change that?
If GDDR7 provides such a boost to memory bandwith Nvidia will have no reason to increase memory configurations on any card!... The margins must remain absurd! The most important factor in the price of future GPUs is going to be the demand of ML (AI).
#6214622
https://forums.guru3d.com/data/avatars/m/246/246564.jpg
I'd temper expectations. Gaming is not their bread and butter anymore. Sure, we're certainly going to get a performance improvement over Ada, but I doubt it's going to be anywhere near what some people are hoping for. Other than the 5090, obviously. Personally, I'm expecting them to continue the trend, so similar gains to what Ada had over Ampere at the same price points. I don't expect an MRSP increase - they learned that lesson with the 4080. There might be supply issues though, and that will obviously lead us right back to scalping, but - again imo - only in the interim, as they run up production.
#6214627
https://forums.guru3d.com/data/avatars/m/181/181063.jpg
It was said many times here that with fast GDDR you don't need a large bus width. As H83 said: what's the point?
#6214629
https://forums.guru3d.com/data/avatars/m/239/239175.jpg
barbacot:

It was said many times here that with fast GDDR you don't need a large bus width. As H83 said: what's the point?
So they're gonna put 24GB on the 4070 Ti? It's either 24 or 12. And they're gonna go with 12, unless they use a 256-bit bus to support 16GB, which they won't.
#6214632
https://forums.guru3d.com/data/avatars/m/216/216349.jpg
RealNC:

So they're gonna put 24GB on the 4070 Ti? It's either 24 or 12. And they're gonna go with 12, unless they use a 256-bit bus to support 16GB, which they won't.
I`ve read that they will be able to mix and match different memory configurations impossible until now, like 4gb modules together with 2gb ones, opening the possibilty of configurations of 18gb and so on. Of course Nvidia will stick with 12gb for the 70 series, we don`t deserve/need better anyway...
#6214635
https://forums.guru3d.com/data/avatars/m/198/198862.jpg
H83:

I`ve read that they will be able to mix and match different memory configurations impossible until now, like 4gb modules together with 2gb ones, opening the possibilty of configurations of 18gb and so on. Of course Nvidia will stick with 12gb for the 70 series, we don`t deserve/need better anyway...
Micron said gddr7 comes with 3gb module so that could bring 18gb/192bit bus midrange cards like 5070 for instance and yet we still get a 12gb configuration.
#6214639
https://forums.guru3d.com/data/avatars/m/53/53598.jpg
At this point i am expecting Nvidia to treat their GPU side of the business as a hobby or tax write off now, so expect to see a sharp decline in quality with a sharp increase in their future GPU prices. lol
#6214652
https://forums.guru3d.com/data/avatars/m/56/56686.jpg
Undying:

Micron said gddr7 comes with 3gb module so that could bring 18gb/192bit bus midrange cards like 5070 for instance and yet we still get a 12gb configuration.
that wont stop them from drop width more and saying it fine raising prices which i hope the stop everything bellow the 4090/80 is stupid priced only card that sense was 4060 which why alot prlace are now selling under retail price
XP-200:

At this point i am expecting Nvidia to treat their GPU side of the business as a hobby or tax write off now, so expect to see a sharp decline in quality with a sharp increase in their future GPU prices. lol
Can hope in one hand and shit in the other and see which fills first....
#6214653
https://forums.guru3d.com/data/avatars/m/271/271560.jpg
i was hoping Nvidia would pull an "Apple" and sell Nvidia NUCs with the xx80m/xx90m as a SFF. yet another new market that would be hugely popular - a VESA mountable form factor for work and play. that way you could have a whole pc with the performance of the desktop xx70/xx80 for the price of the RTX xx80
#6214660
https://forums.guru3d.com/data/avatars/m/275/275921.jpg
tunejunky:

i was hoping Nvidia would pull an "Apple" and sell Nvidia NUCs with the xx80m/xx90m as a SFF. yet another new market that would be hugely popular - a VESA mountable form factor for work and play. that way you could have a whole pc with the performance of the desktop xx70/xx80 for the price of the RTX xx80
I'd be on this so fast my wallet would never know what hit it

Next Page