As many of you know NVIDIA will be launching their first Maxwell-based cards tomorrow. Ahead of the launch our friends over at VideoCardz have released the official specifications and performance for both the GeForce GTX 750 Ti and GeForce GTX 750.
For a run down of the GeForce GTX 750 Series the guys over at video cards have made up the graph below.
Both the GeForce GTX 750 Ti and GTX 750 will be based on the GM107 GPU, which is the first Maxwell processor. It will feature 640 CUDA cores, 40 TMUs, and 16 ROPs. The GeForce GTX 750 Ti will make full use of this GPU which has 5 SMMs (Streaming Multiprocessors Maxwell). The GTX 750 will only make use of 4 SMMs so it will have 512 CUDA cores and 32 TMUs. The rest of the specs are the same although the GTX 750 will have a lower TDP of 55W and less memory. You will see 1GB GTX 750 Ti’s though.
NVIDIA has gone the same route that AMD did with their Hawaii Processor. While it is on the same fabrication process the architecture is improved and more power efficient. Check out the graph below to see the difference between the GK107 (Kepler) and GM107 (Maxwell).
If we look at the performance numbers below we can see that on AMD Title’s like Tomb Raider and Battlefield 4 there is not much of a difference. On most games we would say you are going to see a 10-15% difference. The below graph is comparing the GeForce GTX 750 Ti to the AMD Radeon R7 260X. Keep in mind the R7 260X has 896 stream processors and 56 texture units, so you can see that Maxwell is more powerful on a architectural level with better performance and less power consumption.
While there are no detailed benchmarks for the GeForce GTX 750 (we need to wait for those) there is a graph that shows a comparison to its direct competitor which is the Radeon R7 260. We should mention that the R7 260 has a TDP of 95W while the GTX 750 is only 55W.
Below is another official image of the card. The card will have a one-and-half slot design. The reference board itself will be adopted by a few manufacturers, but don’t expect them to use the reference cooler. Both cards do not have power connectors, power will be drawn directly from the PCI-Express interface.
Because NVIDIA has decided to go with such a low TDP the clock speeds are pretty low. For this reason the overclocking potential is pretty high. You will see cards with 28% factory overclocking, but you will only see this with cards that have additional power connectors.
NVIDIA’s GeForce GTX 750 series is set to launch tomorrow, February 18th.