A few months ago, the 12th generation Intel Core processors from the Alder Lake-S family appeared, and recently cheaper models with a locked multiplier and mobile units joined the group. However, Intel does not intend to slow down for the time being, and in the second half of the year we will get the 13th generation codenamed Raptor Lake. It will be the last generation, using a standard processor design – starting with the 14th generation, Intel will switch to tiled construction with Foveros technology. But back to the thirteenth generation. Again, we get a heterogeneous x86 architecture, using two types of cores: performance and efficiency. Intel’s main goal is to increase the number of effective cores. According to unofficial information, the next change is to be the rebuilt cache subsystem.
Intel Raptor Lake processors are scheduled to receive up to 55% of the cache, which, among other things, will translate to better gaming results.
Intel Core i9-13900K Raptor Lake-S – 24 core, 32 thread processor architecture sample tested
Previous Intel Alder Lake processors offered a maximum of 44 MB of L2 and L3 cache – this number consisted of 14 MB of L2 cache and 30 MB of L3 cache, respectively. However, Intel intends to rebuild the memory subsystem, which will be in response to 3D V-Cache, which AMD is working on in cooperation with TSMC. In the case of the 12th generation, each performance core consists of 1.25 MB of L2 cache and 3 MB of L3 cache. The core of the improved performance (Raptor Cove) in the 13th generation will be a 2MB cache and a 3MB L3 cache.
68M cache is interesting.
– Raichu (OneRaichu) January 14, 2022
Total RPL cache (L1 not included).
– Raichu (OneRaichu) January 14, 2022
Intel Digital Linear Voltage Regulator (DLVR) can significantly reduce the power consumption of Raptor Lake processors
In the case of active cores (Gracemont), which in the 13th generation will be combined in 4 groups with a maximum of 4 cores (16 cores instead of 8), you will receive 4 MB of L2 cache and 3 MB of L3 cache for each block. For the 12th generation, the numbers were 3MB and 2MB, respectively. The differences between individual cores/groups are shown in the table below along with a diagram made by a Twitter user pseudonym Tweet embed. In the case of the P-Core at Alder Lake, we got a total of 10MB of L2 cache and 24MB of L3 cache. In the case of the P-Core at Raptor Lake, these numbers should be 16MB respectively for the L2 cache and 24MB for the L3 cache. For E-Core groups (Alder Lake has 2 groups, Raptor Lake has a maximum of 4 groups), the amount of L2 cache was 6MB, and the L3 cache – 4MB. In the case of the Raptor Lake family, it will be 16MB and 12MB in size respectively. All these differences mean that the total size of the L2 + L3 cache in the 13th generation will be 68MB, which is 55% more than in the 12th generation (44MB). This is to translate, inter alia, for better gaming performance.
Rdzeń P-Core (Alder Lake) | Rdzeń P-Core (Raptor Lake) | Cluster E-Core (Alder Lake) | Klaster E-Core (Raptor Lake) | |
L2 . cache | 1.25 MB (x8 cores) | 2 MB (x8 cores) | 3 MB (x2 klastry) | 4 MB x4 klastry |
L3 cache | 3 MB (x8 cores) | 3 MB (x8 cores) | 2MB (x2 klastry) | 3MB x4 klastry |
Cache L2+L3 | 10 MB L2 24 MB L3 34 MB cache |
16 MB L2 24 MB L3 40 MB cache |
6 MB L2 (2 klastry) 4 MB L3 (2 clusters) 10 MB cache |
16 MB L2 (4 klastry) 12 MB (4 klastry) 28 MB cache |
Source: WCCFTech, TwitterOneRaichu, TwitterOlrak
“Prone to fits of apathy. Introvert. Award-winning internet evangelist. Extreme beer expert.”