Threads
- The RTX 5070 has fewer CUDA cores and Texture Mapping Units (TMUs) than the RTX 4070 Super, potentially impacting its raw processing power.
- 5070 consumes more power (250 watts) than its predecessor 4070 while offering potentially less raw performance.
- The GPU is scheduled to launch in February 2025 at a price of $549, featuring 12GB of GDDR7 memory.
According to Nvidia’s white papers, the RTX 5070 might not live up to the performance expectations set by its predecessor, the RTX 4070 Super. This will be especially noticeable when Multi-Frame Generation technology is not in use.
Why it matters: As official reviews and spec sheets come out, it’s becoming apparent that the RTX 50 series does not offer much of a generational leap in terms of raw performance. If we look at the RTX 5080, it’s only 8% faster than the RTX 4080 Super.
Nvidia’s recent spec sheet confirms the RTX 5070 will feature 6,144 CUDA cores, which is 14% fewer than the 7,168 CUDA cores found in the RTX 4070 Super. Additionally, the RTX 5070 will reportedly have 192 Texture Mapping Units (TMUs), compared to the 224 TMUs in the RTX 4070 Super.
Both cards share the same number of Render Output Units (ROPs), but the reduction in CUDA cores and TMUs could lead to a noticeable dip in raw performance.
The higher-tier RTX 5080 is performing slightly better compared to the RTX 4080 Super. It has 5% more CUDA cores, 10.5% more TMUs, and 14.3% more ROPs, which gives it an 8-9% increase in overall performance.
Without Nvidia’s Multi-Frame Generation technology, the RTX 5070 may not perform as well as the RTX 4070 Super in terms of raw performance. MFG uses AI to predict and create extra frames, making games run smoother.
But when it’s turned off, the GPU’s raw power matters more—and the RTX 5070, with fewer cores, might fall behind.
All in all, it’s becoming clear that the RTX 50 series will heavily rely on DLSS and MFG to gain more performance. For instance, the 5090 is only 27% faster than the 4090 without MFG. To put it into perspective, the 4090 was almost 50% faster than its predecessor, the 3090, raw performance-wise.
Anyhow, The RTX 5070 is set to release on February 20, 2025, priced at $549. It will come with 12 GB of GDDR7 memory and use 250 watts of power, which is 14% more than the previous model, the RTX 4070.
For gamers, the big question is whether the RTX 5070’s focus on MFG will be a problem. While this feature might help increase frame rates, some players prefer the consistent performance of raw GPU power.
Do you think it’s worth it to upgrade to the 5070 from the 4070/4070 Super? Let us know in the comments, or join the conversation on the Tech4Gamers forum.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔
Passionate gamer and content creator with vast knowledge of video games, and I enjoy writing content about them. My creativity and ability to think outside the box allow me to approach gaming uniquely. With my dedication to gaming and content creation, I’m constantly exploring new ways to share my passion with others.
Join Our Community
