Nvidia and partners are said to be making at least three GeForce RTX 30 series graphics cards. mega size GPU (opens in new tab)According to this latest green team rumor, the pipeline will feature an RTX 3060 8GB, a GDDR6X powered RTX 3060 Ti and a GA102 GPU based RTX 3070 Ti. However, the immediate addition of these cards tells us that we still have to wait a few months before his mid-range GeForce RTX 40 series (Ada Lovelace) cards hit the market.
The X060 is often the most popular SKU of any generation, so the introduction of the new GeForce RTX 3060 with 8GB of RAM has the potential to grab the most attention. Of course, this new model will have less VRAM for him than the original (12GB is reduced to 8GB for him), but no other spec changes will shine in the light. Worst case scenario, Nvidia could reduce the bus from 192-bit to 128-bit for 8GB, reducing both memory quota and bandwidth. Without recalibrating the other way around (CUDA cores, clocks, memory speeds), it would simply be a change that would simply underperform the RTX 3060 currently on the market and draw a lot of fire from enthusiasts.
If Zed Wang is correct, the GeForce RTX 3060 Ti update could bring more positive changes. This model is said to have a significant upgrade in memory speed. Use 19 Gbps GDDR6X to go from 8 GB of 14 Gbps GDDR6 to the same quota. Given this information, we can expect about a 35% improvement in memory bandwidth.
Last but not least, the source believes there will be a GeForce RTX 3070 Ti that will utilize the GA102 GPU. This GPU is featured in Nvidia’s flagship cards and seems to be eager to move recently. There will probably be bins full of his GA102 GPU with enough working cores for this new RTX 3070 Ti. Such economics are often the reason behind re-jigging such production.
| RTX3060 | RTX 3060 8GB* | RTX 3060 Ti GDDR6X* | RTX3060Ti | RTX 3070 Ti* | RTX 3070 Ti | |
|---|---|---|---|---|---|---|
| GPUs | GA106 | GA106 | GA104 | GA104 | GA102 | GA104 |
| die size | 276mm² | 276mm² | 392mm² | 392mm² | 628mm² | 392mm² |
| memory bus | 192 bits | 128 bits | 256 bits | 256 bits | 256 bits | 256 bits |
| memory capacity | 12GB GDDR6 | 8GB GDDR6 | 8GB GDDR6X | 8GB GDDR6 | 8GB GDDR6X | 8GB GDDR6X |
| memory clock | 15Gbps | 15Gbps | 19Gbps | 14Gbps | 19Gbps | 19Gbps |
*Specifications are unconfirmed.
It’s almost certain that Nvidia will announce its first Ada Lovelace architecture GPU during its GeForce Beyond special broadcast starting Tuesday for a week. However, it’s a hotly inclined convention for Nvidia to start at the premium his consumer end and then taper off over the next few months. For Ampere, it took five months from the GeForce RTX 3090 to the RTX 3060 launch, and it would be better not to mention the RTX 3050 (add another 11 months).
So at GeForce Beyond on September 20th, we can expect the RTX 4090 launch and possibly more details on pricing and availability. As luck would have it, we have similar information for the GeForce RTX 4080 and RTX 4070.
Tom’s Hardware Graphics editors recently warned against buying the high-end Ampere series.
