A preview of Intel’s next-gen CPU architecture, Raptor Lake, appeared on the benchmarking site for SiSoftware Sandra (opens in new tab) (through video cardz (opens in new tab)) and promises some interesting things. First, you’ve got details on the Core i9 13900, plus some early benchmarks to look at too.
The Core i9 13900 is a high-end chip with the same number of performance cores as the current Core i9 12900, but twice the number of efficient cores. That means you’re looking at eight P cores and 16 E cores, for a total of 24 physical cores and 32 threads.
The Turbos of these cores appear to be lower than lake alder (opens in new tab), at 3.7 GHz for the P-cores and 2.76 GHz for the E-cores, but it’s still early days for the chip and it’s probably an engineering show. This is also offset a bit by having twice as many E-cores and also by doubling the L2 cache.
The L3 cache is also 20% larger. Raptor Lake also appears to support faster DDR5; at 5,600 MT/s against 4,800 MT/s. AMD’s Zen 4 (opens in new tab) it is expected to support faster DDR5 right off the bat too, and has already shown the impact of larger cache sizes with Ryzen 7 5800X3D (opens in new tab).
After highlighting the lack of AVX512 support on Raptor Lake – a feature that was initially disabled on Alder Lake and later merged into later chips – and the fact that it was present on Zen 4, the site goes on to deliver the expected performance for the architecture. Intel’s 13th Gen.
While it is important to note that SiSoftware has not directly tested this chip itself: “At the time of publication,” the report says, “the products have not been directly tested by SiSoftware and therefore the accuracy of the benchmark scores cannot be verified; however, they seem consistent and don’t appear to be fake/fake.”
There are certainly some interesting results on display here, even if they are purely synthetic. For example, the Dhrystone Integer test is 33% faster than the Alder Lake, while the FP64 Double Precision Whetstone results are 100% faster. However, these are non-SIMD (single instruction multiple data) tests, which means there are many tests that won’t see any of this kind of lift.
In the SIMD test, the results appear to be much more modest, between 4 and 8% and this is the kind of improvement we expect to see in most real-world applications. You can expect this delta to increase as well, as early engineering samples tend to run slower than final silicon.
Since there are more cores in Raptor Lake, we couldn’t begin to guess how fast the final CPU will run, but you’d expect it to be close to Alder Lake, even if it doesn’t beat it. For reference, the Core i9 12900 has a maximum Turbo of 5.1 GHz on the P-cores and 3.8 GHz on the E-cores, suggesting there is plenty of room for faster clocks for Raptor Lake.
There’s no set date for when Intel will release its 13th-gen Raptor Lake CPUs, although it’s generally believed to arrive in the fall, potentially as early as October.