Kaj se tiče
"rezanja sabirnice", AMD za RDNA 3 kao i za RDNA 2 prije toga, koristi (povećani) Infinity Cache tak da na kraju s užom/manjom sabirnicom dobe (još) bolje performanse.
Primjera radi,
256-bit GTX 1080 i
192-bit + 96MB Infinity Cache RX 6700 XT koje se bez problema (u tom trenutku) nose s prethonicima koji su pak imali 512-bit, 384-bit ili 2048-bit i 4096-bit sabirnice.
Arhitektura je tu bitnija stavka i očito je da nVidia za RTX 4000 seriju sprema po svega više, tako da će na kraju jedna RTX 4060 sasvim lako imati performanse u rangu npr. RTX 3070/3080 sa samo 128-bitnom sabirnicom.
Po pitanju performansi, nova RTX 4000 i RX 7000 generacija budu svakak bolje od ovih aktualnih, možda ne budu duplo brže u svemu, ali skok bude osjetan.
Nažalost, potrošnja, zagrijavanje i cijena, posebice kod nVidia modela, budu taj skok u performansama pratili u stopu.
Citiraj:
|
nVidia is reportedly looking to reduce orders for 5 nm wafers from TSMC as it anticipates a significant drop in demand from both gamers and crypto-currency miners. The same article paints a different picture for AMD: the company has reduced orders for 7 nm and 6 nm nodes; but its 5 nm orders are unaffected. AMD makes not its its next-generation RDNA3 GPUs on 5 nm, but also its next-generation "Zen 4" CPU chiplets.
|
Citiraj:
|
The flagship Radeon RX 7900 XT will reportedly feature a Multi-Chip-Module (MCM) GPU design with 384 MB of Infinity Cache and a staggering 24 GB of GDDR6 memory. Assuming that the company will use 18Gbps GDDR6 modules again, this would result in a memory bandwidth of 864GB/s. We already knew that the GPU might feature an MCM design, but instead of featuring multiple GCDs (Graphics Compute Dies) as previously thought, AMD will reportedly use six MCDs (Memory Compute Dies) with one big GCD. What this means is that each MCD will feature 64MB of L3 cache, combining for a total of 384MB of what AMD calls Infinity Cache.
|