Luckily we can’t say that this surprises us and soon after you’ve followed us you’ll know. But we will say it anyway: Intel is going to follow in the footsteps of AMD and will have to increase the size of its caches in its new Raptor Lake processors if it wants to compete for performance in content creation and gaming. The data is confirmed, up to 50% more total cache memory in its Core 13.
You may have initially thought that the leak we saw in January was a bit “crazy” since it was too early for something so specific, but it was totally real. The filtering was completely correct and what we are going to see is a 50% increase in the L2 and L3 caches, curiously maintaining the L1.
Intel Raptor Lake and cache: +50% in Core 13
The first thing we must take into account is that we are talking about a greater number of total cores, specifically up to 24 between E-Cores and P-Cores. Specifically, the highest-end Core 13 (assumed i9-13900K) will have 8 P-Core and 16 E-Core cores , which doubles the number of the latter and maintains the number of the former.
Now, it appears.
About p core L2 cache latency, non-accuracy test result shown in pic2. https://t.co/Gf8o2V04Pa pic.twitter.com/mLarcP9lhl
– Raichu (@OneRaichu) May 18, 2022
Therefore, we will see how the L1 configuration remains intact with an L1D with 8 x 48 KB + 16 x 32 KB and an L1I with 8 x 32 KB + 16 x 64 KB respectively, making a total of 2,176 KB for the L1 cache. total. This number increases, but it does so because there are more cores as such, maintaining parity with Alder Lake in each of them.
In the L2 things change, we go from 1.25 MB for the P-Core cores to 2 MB, and from 2 MB in the E-Core to 4 MB in a configuration for the first 8 clusters for the first ( 2 MB per core ) and four clusters for the latter ( 4 MB shared per cluster ). As if that were not enough, the L3 is shared and goes from 30 MB to no less than 36 MB in total .
An astonishing total count: 32 MB of L2 and 36 MB of L3
The sum is simple at this point: 68 MB of total cache without adding the L1 where there are no changes as such other than the amount named. In other words, we went from 44 MB in total with Alder Lake for these two caches to 68 MB for Raptor Lake , an increase of 54.54% . Also, Raptor Lake will have better latency than Alder Lake as you scale in cache block size, so performance will take off quite a bit here.
As seen in the diagram in a simpler way, the architecture will actually have the cache assigned per core as follows: 2 MB for each Raptor Cove Core and 1 MB for each Gracemont Core , that is, Intel has set a parity 2: 1 between P-Cores and E-Cores for L2.
On the other hand, with the L3, a different and equal guideline is followed for each 3 MB cluster, but individually speaking we have 3 MB for each Raptor Cove P-Core and 1.33 MB for each Gracemont E-Core. This shows that E-Cores need a lot of L2 cache and are very little dependent on L3.
For this reason, it is speculated that the Core 14 Meteor Lake , being a major architectural leap, will change these proportions again, but only for the E-Cores, which seem to follow a similar path in their L2 and L3, going down from 4 MB to the first to 3 MB , which should increase the size of the second to compensate for the imbalance produced.
