Several upcoming NVIDIA GeForce and AMD Radeon laptop GPUs have been benchmarked in Geekbench 5, including the RTX 4070, RTX 4060, RTX 4050, and RX 7700S.
AMD and NVIDIA both announced their latest laptop discrete GPU lineups earlier this week, featuring the RDNA 3 and Ada Lovelace architectures. The first laptops based on these GPUs are expected to ship in February next month, but benchmarks for a few of these GPUs appear to have leaked ahead of time. But, before we get into the performance, let’s take a look at the specifications of these GPUs.
NVIDIA GeForce RTX 4070 & RTX 4060 Laptop GPUs – 8 GB Designed for the Masses
The GeForce RTX 4070 and RTX 4060 will have 8 GB of GDDR6 memory. Both GPUs will have a standard 35W profile, a TGP of up to 115W, and a Dynamic Boost Range of 25W. The RTX 4070 will have 4608 Cores, 8 GB of memory, up to 2175 MHz boost clocks, and a 128-bit bus, while the RTX 4060 will have 3072 Cores, 8 GB of memory, up to 2370 MHz boost clocks, and a 128-bit bus.
NVIDIA GeForce RTX 4050 6 GB Laptop GPU – Ada For Mobility
Finally, there will be an entry-level GeForce RTX 4050 6 GB with a 35W-115W TGP option (25W Dynamic Boost profile too). At its 115W TGP range, the GPU will have a boost clock of up to 2370 MHz. The GPU will have a 96-bit bus interface and a clock speed of 1605 MHz.
For Mainstream Gamers, AMD Radeon RX 7700S 8 GB Laptop GPU – Navi 33
The Navi 33 GPU is used in the AMD Radeon RX 7700S, which is based on the RDNA 3 architecture. It has 32 Compute Units with 2048 shader processors, 8 GB of 18 Gbps GDDR6 memory, and a 128-bit bus interface. The GPU includes 32 MB of Infinity cache and has a TDP of 75-100W.
When it comes to performance benchmarks, we have performance leaks across multiple laptops, making it difficult to determine which TGP limits were used by each design. Because all GPUs have a variable range, it’s not a good idea to compare them until we know their power figures. These GPUs are also heavily reliant on the cooling capabilities of each laptop, so that must be taken into account as well. You can’t say the RTX 4060 or 4070 is better than the RX 7700S if it’s running at its base-line 75W TDP versus the NVIDIA GPUs’ peak limit of 115W TGPs.
The laptops and maximum clock speeds at which each GPU was tested are listed below:
- RTX 4050: Notebook NP5x Laptop (2280 MHz Peak Clock) / 35-115W
- RTX 4060: Thunderbolt ZERO Laptop (2370 MHz Peak Clock) / 35-115W
- RTX 4070: Thunderbolt ZERO Laptop (2175 MHz Peak Clock) / 35-115W
- RX 7700S: ASUS TUF A16 Laptop (2208 MHz Peak Clock) / 75-100W
As can be seen, the AMD Radeon RX 7700S is the slowest GPU, trailing even the GeForce RTX 4050. Now, OpenCL isn’t the best performance indicator for Radeon GPUs, but the NVIDIA GPUs (RTX 4070, RTX 4060, RTX 4050) are 40%, 25%, and 4% faster than the new Navi 33 GPU, respectively.
The AMD Radeon RX 7600M XT is the fastest GPU AMD plans to offer next month, but it falls short of NVIDIA’s high-end options for laptops, such as the RTX 4090 and RT 4080. It will be fascinating to see how the next-generation mobility GPUs perform on a per-laptop basis. Several NVIDIA RTX 40 & RX 7000M/S designs have been unveiled by a variety of manufacturers, but we’ll see if AMD can actually deliver all-AMD laptops on time this year, as they did with their previous Radeon 5000 and Radeon 6000 series offerings.
Source: Benchleaks
