Next Generation Hardware Speculation with a Technical Spin [2018]

What, what!? I'd say vast majority of console buyers couldn't give a rat's backside about TF is more accurate a statement!

 

I agree. It's more how the launch games look that will have the biggest impact.
Wouldn't surprise me if the vast majority of console gamers don't even know what a TFlop is.

Problem is it can create a narrative on social media depending on the competition that will adversely effect who ever has less TFlops but that depends on lots of other variable factors also such as price and launch date.

So if both machines have similar TFlops it won't effect the mainstreams impression of next gen in my opinion.

I feel native 4k is not really going to happened next gen. Maybe at launch and certain types of games but not AAA games that will be pushing next gen effects. So 12 TFlops running some form of upscaling can deliver impressive looking next gen games.

 

Game benchmarks always respond well to more bandwidth. Personally I would regard only 16GB as a disappointment. Only 2x after 7/8 years? Memory module density has increased 4x since then.

We could get a Scorpio dev kit (24GB, 384 bit) type memory situation and still not need clamshell. This would ensure better bandwidth than Vega, even after accounting for CPU bandwidth and it’s detrimental effect on total GPU bandwidth.

 

$$$$$

Edit:Rather than leave it at that, I'll add some of my own speculation for people to pick at. I think we see AMD's HBCC concepts getting further developed and used to allow more efficient utilization of the memory in the system mitigating the need for massive amounts of memory in the next consoles. It's something that they specifically called out in Vega for this very purpose and having a defined performance spec for the various components, as seen on console, would enable a better implementation.

 

Sony had 16 4Gb chips and still sold for $399. I’m talking about 12 chips.

 

See edit.

 

There are more options than the typical 256 or 384bit. It could easily be in the middle of those two, for example 20GBs of GDDR6 with a 320bit bus. I'm hoping for 24GB personally and would also be disappointed with a 16GB next gen consoles... Having said that over the years my interest for the specs has gone down. There will be great games regardless of the config of the upcoming consoles. I am currently on a holiday at Crete and played Zelda on the Switch on the plane. Amazing and a beautiful game and runs on a rubbish hardware.

 

I think once console gamers get outside the $399 bubble, maybe Sony/MS will offer more powerful hardware (or multiple systems) during a new launch. Honestly, Sony/MS should be offering an entry-level system and premium model at launch (pricing reflective of their respective specs).

 

I think that is a very viable option and should not be discounted. I'm very doubtful that we would get something like a full 384-bit bus, but something in between may work and be sufficient.

320-bit bus with 16 Gbps memory would provide 640GB/sec which should plenty for something a just a tad more powerful than Vega64 and a Zen CPU.

 

Since the precedence has been set, shouldn't they just release an entry model and a premium model 3-4 years later?

 

I figured they'll concentrate more on shrinks during that time, rather than possibly fighting off anyone holding out purchases until the mid-gen refreshes. Plus, Sony/MS has a greater chance at capturing more dollars during a launch (first two years anyhow), rather than mid-way lifecycle.

 

A more obvious Framing Effect might be necessary so that consumers don't feel "forced" to accept that perhaps inflation is somewhat real. Repeat the PS360 launch in a way.

On the one hand, the console makers probably had enough data from the PS360 generation to show that maybe the lower end SKU wasn't bought much, hence launching the current gen with only one SKU, but I think that's a decision made in a bubble as to the psychology of purchasing (not that I'm an expert in purchasing culture). There maybe needs to be a question/internal study of the value of still presenting that lower end "dumb" SKU if they want to push the average MSRP up to accommodate the BOM necessary for a large enough leap where it's perhaps coming to a point where traditional cost reduction plans are no longer a reasonable expectation.

Storage capacity is relatively simple especially if it's just a difference in the high capacity storage (whether they go with an SSD bit or not, that part would be constant).

 

I think this is why the streaming console makes so much sense. The entry point can be half the cost of the premium SKU.

 

Tough to launch with multiple sku's that have different hardware other than HDD and can they?

Will an extra $100-$200 dollars make much difference? Surely there's only so much they can do with 7nm unless the more exspensive sku goes with discrete parts and the cheaper goes with an APU.

 

Absolutely. The die cost is likely going to in the $80–120 range based on previous consoles. They could massively increase the die and GPU size with $100-200 more budget. In reality that would probably be split across GPU, memory, cooling, and power delivery to enable the best incremental improvement. Console price points operate on the part of the performance curve where there’s a large performance gain with a little extra money input. They’re not in diminishing return land as a matter of necessity.

 

I think it could be a good choice to make $349 1080 machine (same everything except GPU core count) and 4k one at $449 initially.
299 and 399 following launch year.
Afaik silicon scaling after 7nm is not going help to lower price.

 

I highly doubt MS will ever go back to a console with a Power Brick, not after they showed they can perform electrical engineering feats too, unless it's more like the docking port of their Surface product where it's minimal and nothing like the power bricks from the X360 days.

 

I think this idea always ignores the potential negative effect the pro ski may have on the cheap one. Cheap asses don't buy day one. Launch day consumers are looking for the latest and greatest. Who among those would want to the cheap one. They'd be left with a model nobody wants because it's perceived as the lesser, and one few can afford because it's too expensive. It's a risky move.

 

Surely it's not that simple, power usage goes up temperature goes up. Now the more exspensive SKU needs a different power supply and better cooling possibly a bigger console. What if they at the limit of an APU and the only way for more TFlops is to go with discrete parts? That means different motherboard.

Remember PS Pro has twice the TFlops of the base PS4 and some people feel it's not enough. Twice the TFlops of the next gen machines would be what 22-28TFlops. I just don't think they can do that with an extra $100 dollars. Maybe an extra 4 TFlops at Max and does that make much of a difference?