Do you think there will be a mid gen refresh console from Sony and Microsoft?
- Thread starter invictis
- Start date
You might get staggered buyers, those buying the original flavour and those waiting and only buying the Pro each gen, wtih a few hardcores buying the original and selling it towards the Pro. Is there any business sense in that? Is the cost of developing the Pro worth the investment?
If the competition does a Pro and you don't do you lose out (much)? I kinda feel lower price is more important overall. In the case of Sony ,if they were worried about mid-gen users leaving for PC, by moving software to the PC they partially cover that base without hardware.
Arkham night 2
Newcomer
Toivon, että April Leak on todellinen, sinulla ei ole vielä ps5:tä, joten ps5 Pro on erittäin hyvä.
Not if platforms owners don't allow the targeting of particular skus within a gen. Whats the point of just buying Pro consoles if game support dies for the pro before the pro of the next gen is released. I haven't heard of Sony or MS promising the Pro and One X being supported until the release new premium models. Pro consoles are just an upsell proposition to existing console owners.
Pro models make no sense for this gen as current skus are supply constrained and adding performance is a more costly endeavor now.
English, please!
Arkham night 2
Newcomer
Hope that April Leak is real
Inuhanyou
Veteran
It won't be. Zero chance of liquid cooling in any scenario for an off the shelf consumer device
It has been two years, and I still don't see a compelling reason for a mid-gen refresh. The only thing I can think of is just how disappointing AMD GPUs have been when it comes to raytracing, but I don't see that as a reason to do a mid-gen refresh. Raytracing is something that hopefully will be a paradigm shift of future generations of consoles
Inuhanyou
Veteran
Yep. Save it for next gen when a proper leap can be established in a cost effective manner everyone can take part in
RDNA3's dismal performance improvements have put the nail in the coffin for me in terms of whether I expect a PS5 Pro or XS-Z or whatever. Outside of misleading TF figures, I dont think AMD could achieve the actual GPU improvements to justify such new consoles in the first place.
Inuhanyou
Veteran
Even at the top end people are disappointed in what RDNA 3 can do. Hopefully they take a few years and actually fix what they are doing wrong. And maybe when RDNA4 or 5 comes out they can say they made worthwhile product
davis.anthony
Veteran
Are they? RDNA3 offers a very solid uplift over RDNA2, especially in ray tracing.
People were just disappointed with it because the 4090 was such a monster and offered such an above normal uplift that it made the uplift offered by every other GPU (Including Nvidia's own offerings) look like trash.
In no world could RDNA3 be said to offer a 'very solid uplift' over RDNA2.
They had a long development time, they had a complete architectural overhaul and they had a large process node jump. All for.......35%? That's insanely bad. 35% would be underwhelming for any new generation, much less one with all the advantages they had available to put into this one.
And there's nothing particularly monstrous about the 4090/AD102 that should have caught anybody off-guard. It's actually a smaller die than last generation's GA102. That sort of 60-70% improvement was fully expected by people paying attention. The main thing that really helped Lovelace was moving from Samsung 10nm-family to TSMC 5nm-family. This gave Nvidia more room for improvement, and I'm not suggesting AMD should have been able to match AD102, much less with just 530mm² of silicon. They definitely should have been much closer, though.
Again, AMD had a fair bit going for themselves as well. And it's completely embarrassing that their fully enabled, high end Navi 31 part can only match a cut down, upper midrange AD103 Lovelace part(which is what the 4080 is, basically what the 3070 was in the Ampere lineup). This isn't remotely even 'ok', this is absolutely shockingly bad. I genuinely might think it be in the running for AMD's worst new architecture release ever. Something went badly wrong somewhere.
Last edited:
Ok, but I think it's important to make it clear that RDNA3 is a terrible architecture, and as such, this is going to very much hurt Sony or MS's ability to provide a meaningful upgrade for a mid-gen console.
If you argue that RDNA3 was actually a decent or normal uplift, then it would change the picture.
If you argue that RDNA3 was actually a decent or normal uplift, then it would change the picture.
I would expect there to be RDNA4 out by the time potential newer console happens in middle-to-late 2024 through early 2025. That timing [2024/2025] more closely matches the Xbox One X timing as opposed to PS4 Pro timing.
Switching architectures though would be more of a rolling-generation than any sort of mid-gen refresh.
Switching architectures though would be more of a rolling-generation than any sort of mid-gen refresh.
davis.anthony
Veteran
I'll make this quick so not as to derail
- 5700Xt to 6700XT was only 24% increase on average at 1440p (Techpowerup 6700XT review)
- 5600Xt to 6600XT was only 25% increase on average at 1080p (Techpowerup 6600XT review)
So RDNA2 then was even more underwhelming than RDNA3.
And RDNA3 RT performance uplift is way more than a 35% jump over RDNA2 which is where AMD really needed to focus their improvements on as raster performance was strong vs Nvidia and still is with RDNA3.
You can carry this on via PM if you like.
/done
Even then, I'm not sure if there is real value in doing it because it will almost certainly diminish the impact of the consoles supposedly being released in 2028. 4k was a compelling reason for the last gen refreshes, as it would not change how games were being made significantly. As I said, the only reason I see for refreshes is raytracing, and I don't think there is much that can be done there and continue to support current hardware. IMO AMD should take the opportunity to revamp their architecture. They will likely need to do something that more closely matches what Intel and Nvidia are doing and separate Compute and Tensor-like cores.
It's hard to talk about console refreshes and next-gen consoles without talking about GPU architecture.
Last edited:
Of course, but RDNA3 puts them heavily on the backfoot on this. It would require a seriously miraculous level of improvement to make sense, and AMD's execution just doesn't deserve that kind of benefit of the doubt.
I think it would also require that this RDNA4 revision be TSMC N3E, which makes the cost problem even more problematic.
Ignoring that midrange offerings are heavily shifted around per generation by naming and everything and are NEVER a good way of judging actual generational improvements at all(always compare flagship to flagship), you're not taking into account that there was only like 16 months between RDNA1 and RDNA2, not to mention they're both on the EXACT same process node. These are huge factors to take into account here. AMD absolutely achieved massive architectural improvements with RDNA2, especially in efficiency, which allowed their high end parts to be very competitive.
RDNA3 had a longer development time AND it had a major node process improvement along with it.
"You can carry this on via PM if you like" - haha, you could have PM'd me, but man you needed the last word, eh? lol