AMD asks what you want from RDNA 4. PC gamers reply_ 'er, just make sure we can actually buy it, oh

New GPUs selling out in picoseconds has depressingly become the norm. . Maybe that's why when AMD yesterday asked gamers to say what they were most "excited" about for the , "availability would be a brilliant start," pretty much sums up the sentiment.

Of course, that wasn't the only response to AMD's consumer and gaming rep, Frank Azor, when he posted, "" on X. But the word "availability" does pop up rather a lot.

Pricing is another major theme AMD will need to address . "Don’t want to pay more for my GPU than I paid for my entire high end gaming rig a year ago," was the response from one X user and you get exactly where they're coming from.

Next, upscaling generally and more specifically and as one poster put it, "FSR 4 is a big one." FSR 4 will move AMD into the AI upscaling era, matching the approach Nvidia has been using ever since DLSS was first announced way back in 2018.

The slight snag is that just as AMD finally catches up in that broad regard, Nvidia has just made the jump from a CNN to transformer model for its upscaling, in some ways dramatically improving quality. Oh, and it has added Multi Frame Generation to DLSS, too. As ever, then, it feels like AMD is constantly playing catch up, always taking on Nvidia with a feature set that's a few years [[link]] behind.

On the other hand, AMD can take solace from some of the response on X, many of whom said they just wanted solid raster performance at a great price. "All I want is much better 'real' frames per $," is a comment that probably sums up that line of posting.

Beyond that, one notable absence, relatively speaking, was mention of ray-tracing performance. It's not that nobody mentions it at all, but RT absolutely doesn't rank nearly as highly as availability, general performance, and FSR 4. That's interesting, isn't it?

Perhaps predictably, there's something of a chorus of "give us real frames, not fake frames" along with some posters suggesting that AMD needn't waste its time knocking up an answer to Nvidia's Multi Frame Generation.

MFG is just the latest in a long line of Nvidia technologies that have caused controversy. DLSS upscaling , and MFG is splitting opinions, [[link]] .

Arguably the real problem with MFG is how Nvidia has presented it. Claiming that the new is going to match the old RTX 4090, essentially on the basis of comparing the RTX 5070 running MFG to the 4090 running natively, is at minimum dubious marketing. As we now know, . So, the RTX 5070 will be a long way off.

Among other lesser concerns mentioned in the responses on X are performance per watt, plenty of VRAM, AI performance, driver quality, and video encode and decode features. But more than anything it's that trio of availability, price, and upscaling that seems to matter most. Well, it seems to for the first couple of hundred responses on X.

Jump on over to X if you dare and take a look for yourself. As for me, price is key. I said last year, right from launch. AMD needs to avoid the mistake of pricing too high at launch, suffering poor reviews as a consequence, [[link]] only to lower the price in fairly short order, but not make much impact because the PR damage has already been done.

With that in mind, I think a Radeon RX 9070 XT with near RTX 4080/5080 (let's be honest, they're virtually the same) raster performance for $500 maximum is what I want to see, plus a decent RT uplift and FSR 4 upscaling at least as good as DLSS 3. There's not long to wait...