NVIDIA RTX 4070 SUPER has extra L2 cache in the end In its legitimate release, the corporate made a mistake.
Now not 36 MB however 48 MB of L2 cache can be to be had at the RTX 4070 SUPER. NVIDIA has made a mistake with the legitimate documentation, which used to be dispensed as a part of the set up demo. The mistake has simply been mounted now. 4 new graphics playing cards were introduced at CES 2024 for desktop PC players. AMD has introduced its Radeon RX 7600 XT GPU for $329 within the mid-range section, whilst NVIDIA is that specialize in fashions that value two times as a lot, with the RTX 4070 SUPER beginning at $599. This style replaces the prevailing RTX 4070 card, which is now discounted to $549. The legitimate documentation indexed within the information displays a 36 MB L2 cache dimension for the RTX 4070 SUPER style. This knowledge used to be additionally to be had within the legitimate advertising slides. Then again, NVIDIA has now up to date the specs to substantiate that this model is the usage of 48 MB of L2 cache. No different adjustments had been reported.
NVIDIA RTX 4070 SUPER Specifications, Supply: NVIDIA The scale of the L2 cache performs a very powerful position within the reminiscence efficiency of the Ada Lovelace GPU structure. This may scale back VRAM pool questions, particularly on playing cards with a 128-bit check in bus within the RTX 4060 sequence. For the RTX 40 sequence, NVIDIA has modified the scale of the L2 cache on all shopper GPUs, offering as much as 96 MB of L2 cache at the AD102 and 32 MB at the AD107/106. The AD104, the place the RTX 4070 sequence is based totally, has 48 MB of overall L2 cache. Because the RTX 4070 SUPER makes use of an AD104 GPU with extra SMs (56 out of 60), it didn’t make sense for NVIDIA to cut back it to 36 MB, which used to be additionally the unique RTX 4070 specification. With that during thoughts, the RTX 4070 SUPER with 20% extra cores and 33% extra L2 cache will for sure display a noticeable development in gaming. Then again, the discussed fashions can also be restricted via the 220W default TDP, which is most effective 20W greater than the RTX 4070. The RTX 4070 SUPER is ready to hit the marketplace on January 17, with the primary feedback arriving sooner or later later. NVIDIA GeForce RTX 40 VideoCardz.com SMs CUDA Cores Reminiscence Bus Width L2 Cache AD102 AD103 AD104 AD106 AD107 Supply: NVIDIA by the use of Wccftech