geordiemp
Member
Are you talking about L2 cache for CPU or it's for GPU? Because 3900x has just 6MB of L2 cache, caches are really expensive to size of the die.
Only GPU caches I was taling about, RDNA2 has L2 shared, L1 feeding shader arrays and L0 local
CPU is 8 MB on consoles I believe cut down Zen2 and more similar to Zen2 mobile cache.
Last edited: