Next Generation Hardware Speculation with a Technical Spin [post E3 2019, pre GDC 2020] [XBSX, PS5]

Status
Not open for further replies.
That's disappointing, especially from the lads who gave us that 8 GB GDDR5 shocker haha.

It's possible they can clamshell the memory and ship with more than > 16GB, but it'll be a waste of money as SSD already increases effective memory by a magnitude over current gent.

Consider a 30 second game loop.

Current gen: 5GB + 30 x 50MBs = 6.5GB
Next gen: 13GB + 30 x 2GBs = 73GB

More than 10x effective memory in a 30 game time slice.
 
It's possible they can clamshell the memory and ship with more than > 16GB, but it'll be a waste of money as SSD already increases effective memory by a magnitude over current gent.

Consider a 30 second game loop.

Current gen: 5GB + 30 x 50MBs = 6.5GB
Next gen: 13GB + 30 x 2GBs = 73GB

More than 10x effective memory in a 30 game time slice.

The latency is also a lot lower on a SSD. Should be able to get away with caching a lot less data then on a mechanical drive.
 
It's possible they can clamshell the memory and ship with more than > 16GB, but it'll be a waste of money as SSD already increases effective memory by a magnitude over current gent.

Consider a 30 second game loop.

Current gen: 5GB + 30 x 50MBs = 6.5GB
Next gen: 13GB + 30 x 2GBs = 73GB

More than 10x effective memory in a 30 game time slice.
It can't be calculated like this.

Faster storage doesn't help with anything that isn't pure asset loading while moving through the world.
Streaming assets much faster will allow bigger assets, so it needs more memory to take advantage of it.
 
It can't be calculated like this.

Faster storage doesn't help with anything that isn't pure asset loading while moving through the world.
Streaming assets much faster will allow bigger assets, so it needs more memory to take advantage of it.

Wouldn't faster streaming also allow you to cache less data therefore freeing up more ram?. Im not sure how much data this would free, it'd be good to know current gen streaming buffer sizes.
 
It can't be calculated like this.

Faster storage doesn't help with anything that isn't pure asset loading while moving through the world.
Streaming assets much faster will allow bigger assets, so it needs more memory to take advantage of it.

Do we expect to see bigger assets? We're already at 4k assets on XB1X and PC.
 
Do we expect to see bigger assets? We're already at 4k assets on XB1X and PC.
I do expect bigger assets, at least more varied assets in open worlds with more details. It still needs everything in the vicinity to be already loaded and decompressed in ram.

16GB is probably fine, but it depends on what the next engines will need for RT and all the new lighting techniques. I have no idea.
 
Do we expect to see bigger assets? We're already at 4k assets on XB1X and PC.
Yes - the whole pipeline has to improve for someone to see an improvement. 8K resolution on a pure white box won't showcase 8K. If you catch my drift.
 
It's possible they can clamshell the memory and ship with more than > 16GB, but it'll be a waste of money as SSD already increases effective memory by a magnitude over current gent.

Consider a 30 second game loop.

Current gen: 5GB + 30 x 50MBs = 6.5GB
Next gen: 13GB + 30 x 2GBs = 73GB

More than 10x effective memory in a 30 game time slice.

Large parts of the RAM data are not streamable and therefore SSDs won't change anything here. They'll help with streaming but not with data that needs to be stored. Better graphics and AI will need a lot of non-streamable memory. If the RAM remains small the graphical progress will become small.
 
Last edited:
PS4 had only about 160 GB/s of maximum effective bandwidth. 530 GB/s is 3.3x more.
10 TF is over 5x as much processing power though. 13 TF would be 7x the flops with only 3x the bandwidth...unless PS4 had an excess, it's a complete shift in resource balance towards processing over stored data, which conflicts slightly with the fast SSD storage. SSD means virtually more RAM so more data-based resources, but if you're writing constantly to the RAM, you're using up some BW transferring assets from storage. RT techniques could also be hugely impactful on usable BW, with lots of random accesses slowing transfer rates.

Hang on - why is this baseless discussion?
 
The only explanation for it not to be 12 for me would be that Microsoft simply think info about teraflop numbers is ultimately irrelevant and trivial to the average person so they literally wouldn't care if there was confusion online about what it was.

Alternatively, it's still almost a year away from launch and Microsoft might still be waiting to see what the final silicon will do as they get closer to when they'll start manufacturing the consoles for launch.

They likely have a ballpark idea of what the SOC will do, but will get an even better idea when they start getting 100's of thousands of chips rather than the limited quantity they are testing with now.

Similar to how the XBO SOC got a late uplift in CPU frequency relatively close to launch.

Regards,
SB
 
What I really want to know are the tweaks and modifications MS/SONY have made to their CPU/GPU.
Are the CPU tweaks made for XBox One X carried to XBox Serie X or are irrelevant with Zen+/Zen2?
SSDs will be optimized NVME or something tailor made?
RT is from AMD, third party IP or in-hose IP?

Next generation remind me in some aspects 8bit/16bits computers era, non-upgradable computers made mostly on the self components with some tailored parts and new improved models compatible at software level.
 
Has anyone tried to measure the motherboard shown in the reveal? There was apu measurements, but the mobo seemed very traditional in shape, now we have seen the case does the mobo even fit?

What would that mean on mem size, bus width assumptions etc?
 
Alternatively, it's still almost a year away from launch and Microsoft might still be waiting to see what the final silicon will do as they get closer to when they'll start manufacturing the consoles for launch.

They likely have a ballpark idea of what the SOC will do, but will get an even better idea when they start getting 100's of thousands of chips rather than the limited quantity they are testing with now.

Similar to how the XBO SOC got a late uplift in CPU frequency relatively close to launch.

Regards,
SB

I understand they are not releasing the numbers yet because the clocks could change but that wasn't what I was talking about.

I was basically saying given how easy it is to connect the dots between the 12Tflop rumour for Anaconda and Phil saying it's "2x the gpu power of X1X" that the only way it wasn't the case (it being 12Tflops) is they are simply not worried about having the teraflop number being misconstrued now. No need to address, clarify or be exact in your language about anything if the teraflop number is never going to be a focal point of the marketing of the console..or at the very least don't think the majority of consumers care enough or even aware about "teraflops" to ever feel mislead.

I don't believe they(Phil Spencer) are that unaware or "careless" though....which is why I am leaning towards believing the number.
 
I do expect bigger assets, at least more varied assets in open worlds with more details. It still needs everything in the vicinity to be already loaded and decompressed in ram.

16GB is probably fine, but it depends on what the next engines will need for RT and all the new lighting techniques. I have no idea.

You cannot replace a big Ram Amount with streaming via SSD.
SSDs help with streaming, they do not help with data that must be kept available. Do you want ray tracing? You need ray tracing data structures in RAM. Want better lighting? Requires RAM for GI probes or voxels. Better shadows? Requires RAM for larger shadow cache. Better BRFS? More RAM for fatter G-buffers. Better AI? More RAM for game code. All of this cannot be streamed and is independent of the resolution. So yes Next Gen Progress , Game Development is holding back with the 16 Gbyte Ram limitation, and the lowest common Dominator called "Lockart".
 
Status
Not open for further replies.
Back
Top