The GeForce RTX 5070 Ti is going to be a graphics card entry to the high range which will position it one step above the GeForce RTX 5070. Normally the Ti versions usually offer a substantial performance improvement compared to the non-Ti models, but in this case the difference could be even greater than expected, as long as the specifications are met. that have been leaked.
The first and most important thing is the graphic core. The GeForce RTX 5070 will, in theory, use the GB205 core, which in its full configuration will have 6,400 shaders. This GPU leaves very little room for those in green to shape a GeForce RTX 5070 Ti, and that is why it is rumored that the GeForce RTX 5070 Ti will use the GB203 graphics core, which is the same one that will be present in the GeForce RTX 5080.
The GB203 GPU will have a count of 10,752 shaders in its full configuration, which means that in this case NVIDIA does have much more room to use a trimmed version in the Ti model, which is precisely where the latest information that has appeared points. . Its configuration would therefore be much more powerful than the GeForce RTX 5070.
Possible specifications of the GeForce RTX 5070 Ti
- GB203 graphics core manufactured on TSMC’s 4nm node.
- 70 SM units.
- 8,960 shaders at speeds between 2.8 GHz and 3.2 GHz.
- 280 units of textured.
- 128 raster units.
- 70 fourth generation RT cores.
- 280 fifth generation tensor cores.
- 256-bit bus.
- 16 GB of GDDR7 memory at 28 Gbps (896 GB/s bandwidth).
- 64 MB of hidden L2.
Compared to the GeForce RTX 4070 Ti, the GeForce RTX 5070 Ti would have many more shaders (7,680 versus 8,960), and therefore would also have more texturing and raster units, and more tensor cores and RT cores. The difference is considerable, and therefore the jump in terms of gross power would be considerable.
Another important difference is in the configuration of the memory subsystem, since the new NVIDIA graphics card would have a 256-bit bus and 16 GB of GDDR7 memory. This not only represents a increased bandwidth compared to the GeForce RTX 4070 Ti, which has a 192-bit bus and 12 GB of GDDR6X memory, but will also have 4 GB more graphics memory.
Going from 12 GB to 16 GB makes a graphics card much better prepared to move current games in 4K with maximum quality, especially when we activate ray tracing, a technology that significantly increases resource consumption. It would be a very positive change, since it would make the GeForce RTX 5070 Ti a very capable graphics card in 4K.
It is also important to consider the L2 cache. The GeForce RTX 4070 Ti has 48 MB of L2, and the GeForce RTX 5070 Ti could have 64 MB of L2. Those 16 MB extra that the new NVIDIA graphics card will have are going to make a substantial difference, since having more L2 cache increases hit rate by searching data in the cache even at high resolutions, and reducing dependency on graphics memory.
In general, 64 MB L2 cache is an optimal amount for 4K. That is precisely the cache configuration that NVIDIA has used in the GeForce RTX 4080 and GeForce RTX 4080 SUPER, so it is no coincidence, this increase in cache memory is perfectly justified.
The GeForce RTX 5070 Ti will use the new blackwell architectureso we can expect greater performance both at the level of shaders and tensor cores and RT cores, which will be fifth and fourth generation, respectively. The GB203 graphics core will be manufactured in TSMC’s 4nm node, which implies improved performance and efficiency.
Power, performance and equivalences
If the specifications we just saw are confirmed, the GeForce RTX 5070 Ti will have a power of 57,34 TFLOPs en FP32. The GeForce RTX 4070 Ti has a power of 40.09 TFLOPs in FP32, so the GeForce RTX 5070 Ti would be much more powerful, since it would get 17.25 TFLOPs in FP32, a difference that represents an important generational leap, since we talk about 42% more power.
Making a direct comparison in terms of gross power, the GeForce RTX 5070 Ti would be above the GeForce RTX 4080 SUPERwhich reaches 52.22 TFLOPs in FP32, but would not be at the level of the GeForce RTX 4090, which reaches 82.58 TFLOPs in FP32. NVIDIA’s current top of the range is so powerful that it looks like it will only be surpassed by the GeForce RTX 5080.
However, it must be taken into account that gaming performance is not measured solely by the TFLOPs that a graphics card has, and that perhaps the improvements that NVIDIA introduces at the architectural level help the GeForce RTX 5070 Ti get a little closer to the GeForce RTX 4090. All in all, if these specifications are confirmed, the numbers are very clear and indicate that it will be only slightly above the GeForce RTX 4080 SUPER.
In AMD equivalents, the closest would be the Radeon RX 7900 more TFLOPs. I wanted to give you this comparative data to better illustrate what I said before, that when we talk about performance in games TFLOPs are not everything.
The ideal resolution to play with a GeForce RTX 5070 Ti it will be 2160pboth due to its power and the configuration of the memory subsystem and the amount of L2 cache. Those 16 GB of GDDR7 and 64 MB of L2 cache will ensure a good experience even at maximum settings under rasterization.
If our goal is ray tracing, with the GeForce RTX 5070 Ti we should have performance similar or superior to the GeForce RTX 4090 thanks to the improvements that Blackwell will bring, and the new fourth generation RT cores. This is the trend that we have been seeing in previous generations, and I will give you an example, the GeForce RTX 3070 usually performs the same or a little better than the GeForce RTX 2080 Ti in ray tracing.
As I have already told you in previous articles, it is most likely that NVIDIA will take advantage of the launch of the GeForce RTX 50 to introduce some new technology that makes a difference compared to the previous generation. Maybe this will be integrated into the DLSS ecosystem and take advantage of AI, or maybe it will focus on ray tracing, right now it is impossible to guarantee anything.
What processor and what power supply will I need to power a GeForce RTX 5070 Ti
If this graphics card ends up having a TGP of 300 watts, which is very likely, we will need a power supply of at least 600 watts, as long as the rest of the components of our PC do not have very high consumption. For example, a PC configured with a Ryzen 5 7600X, 32GB of DDR5, a 1TB PCIe SSD, six fans, and 240mm liquid cooling would have a consumption of 500 watts in load.
With that same configuration, but changing the processor for an Intel Core i9-14900K, we would have a load consumption of 636 wattsand in this case the 600 watt source would be too small for us. If we have that CPU we will need at least a 700 watt power supply so as not to go too tight and have margin with which to face possible peaks.
It is important that the power supply has a minimum of quality and that it has the 16-pin power connector, which will be the one that the GeForce RTX 5070 Ti will come with. This will make our lives much easier, since It will prevent us from having to use 8-pin adapters. In this case, the possible TGP of 300 watts indicates that we would need two 8-pin connectors, since each one can supply a maximum of 150 watts.
Let’s now talk about the recommended processor to accompany a GeForce RTX 5070 Ti. Assuming that everything I have said in this article is fulfilled, we will need at least one Ryzen 5 5600X o un Intel Core i5-12400F so as not to have a serious bottleneck, although the ideal would be to have a Ryzen 7 7700X o un Intel Core i5-13600K so that this graphics card can develop its full potential.
Release date and possible price
If NVIDIA repeats the current generation strategy, the GeForce RTX 5070 Ti should arrive approximately in March 2025before the launch of the GeForce RTX 5070, which would arrive a little after this. However, it is also possible that the green giant chooses to launch both at the same time, or with a minimal difference.
Regarding the selling price, it is clear that this graphics card is going to be more expensive than the GeForce RTX 5070, but the question is how much more. The most logical thing would be for this to have a price of between 700 and 749 dollars, figures that at the exchange rate in Spain and after applying taxes could become 749 and 799 eurosrespectively.
Those prices would be for the version Founders Editionand represent the best possible scenario. By this I mean that we cannot rule out that the GeForce RTX 5070 Ti ends up being more expensive, and that it exceeds 800 euros. Obviously models with higher quality premium finishes will have a higher price.
Images generated with AI.
Source: www.muycomputer.com
