AMD’s flagship Radeon RX 7900M “RDNA 3” laptop GPU has been benchmarked and it is faster than NVIDIA’s RTX 4090 Laptop chip in Vulkan tests.
AMD Radeon RX 7900M “RDNA 3” Flagship Laptop GPU Benchmark Leak: Faster Than 4090 Laptop In Vulkan, Slower Than 4080 Laptop In OpenCL
The AMD Radeon RX 7900M is the flagship gaming GPU for laptops from the Red team and it is also the first laptop GPU ever to make use of a chiplet design. The chip was exclusively announced for the Alienware m18 laptop last month but despite being available since last month, there have been little to no reviews of the new Radeon GPU. We did get a teaser from a YouTuber, GizmoSlipTech, but besides that, there hasn’t been much. But now, we have the latest benchmarks leaked out in Geekbench’s OpenCL and Vulkan tests which should have surprising results.
AMD Radeon RX 7900M “Navi 31” GPU Specifications
The AMD Radeon RX 7900M GPU is based on the RDNA 3 graphics architecture and features a total of 36 WGPs or 72 Compute Units which make up a total of 4608 cores. That’s 20% more cores than the 7800 XT & 10% lower cores than the 7900 GRE. Clock speeds for the GPU peak out at 2090 MHz.
In terms of memory, the GPU is equipped with a total of four MCDs which will enable a 256-bit wide bus interface & 16 GB memory capacities. That will get AMD on par with NVIDIA’s top RTX 40 offerings such as the RTX 4090 which max out at 16 GB memory too. As for the power numbers, the AMD Radeon RX 7900M will be configured at a TGP of 160W & can be pushed up to 200W with SmartShift technologies.
AMD Radeon RX 7900M “Navi 31” GPU Performance
Starting with the numbers, we first want to outline the configuration. The Alienware m18 laptop is configured with the Ryzen 9 7945HX CPU which is the highest you can get and its a really missed opportunity to not include the Ryzen 9 7945HX3D model since this high-end gaming laptop will be perfect with the boost that comes from 3D V-Cache CPUs. The CPU still operates well over 5 GHz & there’s also 32 GB of DDR5-5200 memory.
Performance-wise, the AMD Radeon RX 7900M Laptop GPU scored 171,430 points which makes it faster than both the NVIDIA GeForce RTX 4090 & RTX 4080 laptop GPUs and almost on par with the RTX 4070 Ti Desktop graphics card. So that’s essentially RTX 3090 Ti or RX 6950 XT performance in a laptop form factor. The GPU is 9% faster than the 4090 Laptop & 25% faster than the 4080 Laptop GPUs.
0
42335
84670
127005
169340
211675
254010
In the OpenCL benchmark, we see the exact opposite with the Radeon RX 7900M scoring 144,611 points which ends up being slower than the RTX 4080 Laptop GPU by -11.5%. Here, the card ends up between the 7800 XT and the 6800 XT.
0
53730
107460
161190
214920
268650
322380
We should highlight the Vulkan performance more since that is a more popular API used in games than OpenCL. AMD having the lead there is definitely great to see and more in line with what the company has officially revealed in its own internal benchmarks. The AMD Radeon RX 7900M does feature a higher target TGP than the RTX 4080 but at the same time, laptops featuring the GPU are going to cost less as detailed here. Overall, the Radeon RX 7900M is a strong RDNA 3 Laptop GPU but we would have loved to see it in more laptop configurations and options too.
AMD Radeon RX 7000M RDNA 3 Mobility GPU Lineup:
| GPU Name | GPU Architecture | Process Node | Die Size | GPU Cores | GPU Clock (Max) | Memory Capacity | Memory Bus | TGP |
|---|---|---|---|---|---|---|---|---|
| Radeon RX 7900M | RDNA 3 Navi 31 | 5nm + 6nm | 304.5mm2 | 4608 | 2090 MHz | 16 GB GDDR6 (64 MB Infinity Cache) | 256-bit / 512 GB/s | 160-200W |
| Radeon RX 7600M XT | RDNA 3 Navi 33 | 6nm | 204mm2 | 2048 | 2300 MHz | 8 GB GDDR6 (32 MB Infinity Cache) | 128-bit / 288 GB/s | 75-120W |
| Radeon RX 7600M | RDNA 3 Navi 33 | 6nm | 204mm2 | 1792 | 2070 MHz | 8 GB GDDR6 (32 MB Infinity Cache) | 128-bit / 256 GB/s | 50-90W |
| Radeon RX 7700S | RDNA 3 Navi 33 | 6nm | 204mm2 | 2048 | 2200 MHz | 8 GB GDDR6 (32 MB Infinity Cache) | 128-bit / 288 GB/s | 75-100W |
| Radeon RX 7600S | RDNA 3 Navi 33 | 6nm | 204mm2 | 1792 | 1865 MHz | 8 GB GDDR6 (32 MB Infinity Cache) | 128-bit / 256 GB/s | 50-75W |
News Sources: Benchleaks #1, #2
