The RTX 5050 is back in the news, as new rumours surface concerning the type of memory used. Apparently, the card will be using GDDR6 memory chips, whereas the first rumors referred to GDDR7 VRAM. Please note that these are just rumors, and should be treated with a degree of caution!
RTX 5050: GDDR6 instead of GDDR7!
Things are moving fast regarding this future new entry-level card. According to Benchlife’s initial reports, the card was to be equipped with GDDR7 memory, a claim questioned by another colleague. However, it seems that this questioning has borne fruit, as Benchlife has again questioned its sources, who have confirmed that the card will indeed be in GDDR6. Now, as far as VRAM suppliers are concerned, Samsung will supply the memory chips in the first instance, while SK Hynix is also on the list, but this will come later it seems. In addition, we have confirmation of a card number, namely PG152-F01 SKU 50. Otherwise, there’s no change in the expected configuration, at least for the time being. In fact, we’re still expecting to see a GB207 equipped with 2560 cuda core. On paper, that’s 33% less cuda core than an RTX 5060. Furthermore, the memory bus size remains identical: 128 bits, as does memory capacity. In the end, the only difference will be in the type of VRAM used, which will also have an impact on memory bandwidth, which will be reduced. As GDDR7 operates at higher frequencies than GDDR6, this has an impact on bandwidth, especially on two cards with the same bus width. Finally, early rumors indicate that the card will arrive in July, so we’ll keep an eye on that.
