AMD Ryzen 9 7945HX Analysis - Zen4 Dragon Range is faster and more efficient than Intel Raptor Lake-HX

The AMD Ryzen 9 7945HX (Dragon Range) is manufactured in a 5 nm process, while Intel is still using the old 10 nm process (FinFET, Intel 7). Intel's cores require a lot of power especially for the high Turbo clocks, which was already an issue for the previous Alder Lake chips. AMD will also introduce the first 4 nm chips (Phoenix) for the HS and U-series CPUs in a couple of weeks, but they operate in a lower TDP range between ~25-80 Watts.

The following results for the ROG Zephyrus Duo 16 were determined with the Turbo mode (130/100 Watts), for the ROG Strix G17 with the Performance (120/90 Watts) as well as the Turbo mode (125 Watts).

We start with the raw single-core performance and the new Zen4 cores have a clear advantage over the old Zen3+ models like the Ryzen 9 6900HX. Our single-core benchmarks show an average advantage of 22-26 percent for the Ryzen 9 7945HX, which is a very impressive result. This also means AMD is roughly on par with the new Raptor Lake HX chips like the Core i9-13900HX. The Core i9-13980HX on the other hand benefits from the even higher Turbo clock and is slightly faster than the Ryzen 9 7945HX, even though the average advantage of 3 up to 7 percent is not huge anymore.

Looking at the results you should not forget that the AMD Zen4 core requires about 27 Watts in single-core tests, the performance core of the new Raptor Lake HX CPUs about 34-36 Watts, which means the raw CPU efficiency of the AMD chip is much better compared to Intel. We already mentioned that the measurements with an external screen required the dGPU on the AMD systems from Asus, so the results below are not fully representative, but we want to show them nonetheless. If you are a very positive person, you could say that the new Ryzen 9 7945HX is still on par with the Intel Raptor Lake-HX chips (where the iGPU was running) even with the activated dGPU.

The Ryzen 9 7945HX is the fastest chip in our multi-core benchmarks and the chip with the locked 125W is sitting at the top of the comparison charts. The two fastest Intel CPUs in the MSI GT77 Titan (Core i9-13950HX & Core i9-13980HX) are between 5-6 % slower on average. This does not sound like much, but you have to know that the CPUs in the Titan GT77 consume much more than 150W.

The efficiency results for the AMD systems are once again affected by the active dGPU, but the impact is much smaller here due to the higher overall consumption during multi-core loads. The two Ryzen 9 7945HX chips running in Turbo mode are about 35-45 % more efficient than the current Raptor Lake i9-13950HX, and the Ryzen 9 running at 120/90 Watts increases the gap to 60-70 percent compared to Intel. AMD is also very close to the current Apple M2 Max (also 5 nm).

In addition to the maximum performance, we were also interested in the multi-core performance at lower power limits. We used the app Throttle Stop to adjust the power limits of the Intel models, while we used the manual performance mode in the preloaded Armory Crate software for the Asus ROG Strix G17.

One more aspect we want to mention is the limited multi-core performance on battery power. There are obviously differences between individual laptop models, but pretty much every device will limit the CPU performance noticeably when you are running on batteries. The AMD CPU in the ROG Strix G17, for example, can consume up to 90W for short periods, but there are also frequent drops and the CB R23 Multi score of ~15000 points shows that the battery performance is reduced by more than 50 % and pretty much comparable to the Apple M2 Max. However, the situation is not better for the Intel models, either.