Last Minute Telegrams: GoG, the future of Radeon vs Nvidia

I love GOG!

CarryTrainer.Editor

Are we ever going to see a return of Crossfire/SLI? And if we do, won't that scenario favour AMD more, as you can comfortably afford two 5700xts for the price of one 2080ti.

Durmij

You could also spend a moment on AMD selling a lot of IP off to Qualcomm, and how Adreno is built on ATi/AMD tech. Maybe also an alt history of "what if AMD never bought ATi".

KarbinCry

Another question. Why is nVidia unable to make any truly really super Cool GPUs, like Radeon VII? Follow-up: can you salivate for a bit about a (impossible) Radeon VIII based on Renoir's CUs? ;)

KarbinCry

How should nVidia react to the coming "next level" APUs from both AMD (Renoir and further, future RDNA APUs, implementation of small HBM (and/or fast DDR5) to solve bandwidth bottleneck) and Intel (Tigerlake, 10nm GPU chiplets)? What is their best strategy going forward, and what you think their actual strategy is going to be?

KarbinCry

As far as i remember, Raja is responsible for RTG bein a semi separate company. Before it was just AMD as a whole doing cpu and gpu (probably still sub groups within amd tho). But RTG is financially dependent on AMD. Dont quote me on that tho, I might be terribly wrong.

Sandro Evertz

Do you think AMD overmanufactures on TSMC 7nm to keep Nvidia (or possibly Intel) from advancing to that node?

•Considering what ATi/AMD has accomplished in the past (5970, 6950/6970, 7970, R9 290X, Ryzen APUs, and even Vega), do you think Big Navi/Nvidia Killer will finally be a legitimate push for the performance crown? •I remember the 5970 holding on as the flagship beast against the GTX 580 while the Cayman GPUs (6950/6970) duked it out with the GTX 570 or the previous Gen GTX 480. Just to keep this point brief, I see a similarity with that era and the Polaris/Vega with Navi going up against Pascal and Turing. But AMD fell down another performance tier against Nvidia with Turing. RDNA is the transitional architecture, somewhat like Cayman helped AMD transition to GCN in my opinion, that couldn't beat Nvidia's best GPU at the time. It took GCN to take the fight to Nvidia, which I hope RDNA 2.0 does nearly a decade later. •With Vega, Nvidia believed the hype and dropped the 1070 Ti And 1080 Ti as hard counters to the Vega 56 and Vega 64, respectively to quash potential high-end competition from AMD. Radeon VII is interesting in its launch due to issues with Navi, but was also a nice prosumer card that was misadvertized as a gaming card. However, RDNA and Renoir show that tweaking Vega with RDNA's improvements is beneficial to Vega. Could you possibly (fingers crossed) see that translate into a Radeon VII prosumer successor alongside RDNA 2.0 cards? Or would AMD simply leave Vega to the AI market and just make workstation versions of RDNA 2.0 cards, similar to RDNA? •Would you see a repeat of the 290X with disruptive 4K Performance? Or the Fury X With a price drop if the biggest GPU couldn't beat the 3080 Ti/Titan?

The Immortal Cameraman

Will Nvidia eventually become irrelevant in low to mid-tier performance because APU's will become sufficient for let's say 4K 120hz.

Tommy

Memory Overclocking: Friend or Foe?? Kappa

Bollocks

I might be mis-remembering this a bit but Jim at AdoredTV mentioned that Raja wanted to split the Radeon Technologies Group off as a separate company. It's probably not something that's going to happen but are there any reasons why it might be a good/bad idea?

RDNA is said to have baked in ray tracing. Turing has raytracing aswell (kind of). Why is there no dedicated raytracing cards? Do you think using separate raytracing accelerators could be a thing? Or will it more likely introduce latency penalties when one card waits for the onther one to be ready, like in crossfire/sli when both cards are involved on rendering the same frame where one part is more demanding than the other half. Having bought a 5700xt, I would actually consider a raytracing addon card, if it worked well.

Sandro Evertz

Do you think AMD can narrow the gap to High-End-GPUs from Nvidia? Many people think Big Navi will be the Nvidia-Killer, but I‘m doubtful about this. Navi is a good architecture and they had some money for the development from Sony and Microsoft, but now they are earning real money with their zen-based products. Do you think AMD can invest this money now and really attack Nvidia in two to three years?

Stefan Bien

to both: looking at the future, do you see shady practices (gpp, hairworks and the like) continue in your head for a marketing perspective, or do you both see generic consumers waking up to such practices? another one to both: how will nvidia tackle the integrated chip approach? will they expand on the tegra line for their apu stuff or interlicence with other semis for say a intel/nvidia hybrid chip? lastly to both again: what kind of performance will we see in 3 years for rasterization? will it level off as we more to get more raytrace performance? p.s. love both of your channels, will repledge to chris in a bit :)

8bit-bunny

Do you think that we will see proper price wars bringing down the entire linup of GPUs in the next year or two, or has AMD's management change necessitated a lawsuit against price fixing or a 3rd competitor?

Nils