The first review of AMD’s budget RDNA 2 mobile GPU, the Radeon RX 6600M has surfaced, giving us an idea of how the company’s mainstream offering stacks up against the GeForce competition. The 6600M is based on the Navi 23 die, featuring 1792 stream processors (28 CUs/14 DCUs) and 8GB of GDDR6 memory paired with a 128-bit bus. It packs just 32MB of L3 (Infinity Cache), compared to a massive 80MB on the 6700M. Furthermore, it has a Game Clock of just over 2.1GHz and uses the PCIe 4.0 x8 interface.
It’s important to note that in this particular scenario, we’re not exactly looking at an apples-to-apples comparison as the 6600M has a TDP of 100W (120W w/ VR loss) while the RTX 3060 can draw up to 130W under Dynamic Boost 2.0. Both GPUs were paired with the Ryzen 7 5800H and 32GB of DC DDR4-3200 memory.
Regardless, the Radeon RX 6600M performs rather well at both 1080p and 1440p. It’s just under 10% slower than the GeForce RTX 3060 at 1080p with the delta expanding to 112.78% at 1440p. It looks like the 128-bit bus is turning out to be a bit of a bottleneck as the card consistently beats the 3050 Ti but falls behind the 3060.
The RX 6600M nets a lukewarm 27.55 MH/s in Ethermining, short of the 5600M and roughly the same as the RTX 3060 with the mining nerf.
In terms of the power consumption, the 6600M draws up to 105W in Time Spy, a fair bit lower than the RTX 3060 which approaches the 140W mark in the same test. The RTX 3050 Ti also draws just 5-10W lower than the former. Rather an impressive showing from the RX 6600M.