Official Nvidia Ampere/Consumer Volta thread

even if they announce new stuff there will be delays, or low stock and high prices due to the lack of production in asia atm.
 
It's got to have a release date in order for it to be late. I just want to know when they expect to launch it and more importantly, what the 3080TI will offer. 5000 or 6000 Cuda cores? 14GB or 16GB or VRAM? What will the memory bus be? That sorta thing is what I want to know. I'm glad the information will still be available at the end of the month.
 
I'm also anxious to see what the info is. I can wait for the actual card, however long that'll be. But at least some info on it would be great.
 
the_sextein said:
It's got to have a release date in order for it to be late. I just want to know when they expect to launch it and more importantly, what the 3080TI will offer. 5000 or 6000 Cuda cores? 14GB or 16GB or VRAM? What will the memory bus be? That sorta thing is what I want to know. I'm glad the information will still be available at the end of the month.
Click to expand...

I mean they could, but then they'd lose some sales to people who don't want to hold out for their next cards.
 
Been away for awhile, hi guys!!!!

I really believe the rumor from the wccftech posted on Jan 20th.

kopite7kimi was the one who posted it but his account has all posts protected. :bleh:

https://twitter.com/kopite7kimi?lang=en

https://wccftech.com/nvidia-ampere-gpu-geforce-rtx-3080-3070-specs-rumor/


Tesla A100? = GA100 8GPC*8TPC*2SM 6144bit (8196 cuda) 24/48GB HBMe?
RTX3080Ti? = GA102 7GPC*6TPC*2SM 384bit (5476 cuda) 12GB GDDR6?
RTX3080? = GA103 6GPC*5TPC*2SM 320bit (3840 cuda) 10GB GDDR6?
RTX3070? = GA104 6GPC*4TPC*2SM 256bit (3072 cuda) 8GB GDDR6?
RTX3060? = GA106 3GPC*5TPC*2SM 192bit (1920 cuda) 6GB GDDR6?
RTX3050? = GA107 2GPC*5TPC*2SM 128bit (1280 cuda) 4GB GDDR6?

Seems very reasonable.

Indiana University's Big Red 200 super computer that is using nvidias "next gen" cards is said to bring 75% more performance per GPU (over V100). They opted out of using V100s and will be using the A100s? and will use a lot less GPUs.

Just speculation. :lol: :lol: :lol: :D
 
I noticed that article from Indiana University and it looks like Nvidia got a nice jump in performance on their new generation of data center GPU's. I hope some of that bleeds down to us bottom feeders:lol:
 
Not going to be waiting for the Ti release and the price, will pickup a 3080 for my 4K gaming, looks to me like the Ti cards will probably come out much later like the old days.
 
You must log in or register to reply here.
Back
Top