Why is AMD comparing their 48Gb professional cards to Nvidia’s 24Gb gaming cards? They do know Nvidia makes 48Gb cards, don’t they? (the A5880 & A6000)
They specifically picked models which take up more than 24Gb VRAM, so the Nvidia cards would have to use system RAM. They did the same thing with their Strix Halo benchmark.
This is why no one takes AMD seriously in AI.
I assume comparison was done based on cost, not RAM amount parity, which makes sense because its price/performance you’re most interested in usually.
But on the flip side, the 48GB RDNA 3 GPU is less than half the price of the closest current-generation 48GB Nvidia GPU you can buy today, the RTX A6000 Ada.
Shocking: higher spec new card beats lower spec’d, previous generation cards from competitor. Yet they are mysteriously silent about current generation cards from the competitor.
Rdna 3 is AMD’s previous generation. This card has the same chip as an RX 7900 xtx, actually you could say it’s the workstation version of that, like a Quadro RTX 6000 and a GeForce RTX 2080ti.
You mean the new “paper launch” card with thermal issues and missing ROPs?
Which new generation? It hasn’t been deployed because they can’t make it work (see meta et al *canning blackwell backorders and going for Ada which is stable, reliable, and doesn’t have heat dissipation issues)
