Next-Generation NVMe SSD and I/O Technology [PC, PS5, XBSX|S]

[davis.anthony]

The challenge I see for next consoles, is OT however, I just don't see how we can get a mid-gen refresh hitting 20TF of power with the appropriate supporting bandwidth for $499. That's one hell of a chip, and we'd likely have to be having some sort of on-chip cache, I don't see how there is another way around it.

I've been thinking for a while that it may be time to start looking at HBM2 again?

24GB max would be a nice increase over 16GB without going to the full 32GB.

The only issue would be cost as we know the bandwidth performance itself shouldn't be a problem as even Radeon VII had 1TB/s of bandwidth.

 

[Globalisateur]

I mean the base consoles are fine. The real question isn't what we have today and whether it's good enough. I think this discussion has done a good job at showcasing this was the right solution. The problem I see is where we go from here.
The next generation of console or mid gen refresh - it's not like getting even more SSD will make things even better. We will need more compute and more bandwidth to increase graphical fidelity. This SSD imo is a 1 time boost. I think PS5 getting a larger step forward here on this front likely means it can sit at this particular speed for a generation or two, maybe more. XBox may find this solution inadequate by next generation, I dunno.

But now that the I/O bottleneck is alleviated, the next generation of consoles is just going to push more on compute and bandwidth until I/O is a problem again, and then we'll fix it and the cycle continues.
The challenge I see for next consoles, is OT however, I just don't see how we can get a mid-gen refresh hitting 20TF of power with the appropriate supporting bandwidth for $499. That's one hell of a chip, and we'd likely have to be having some sort of on-chip cache, I don't see how there is another way around it.

Indeed. Hoping they won't do the same mistake as PS4 Pro.

A chiplet design (40CUs *2) for PS5 Pro would work very well. They could keep the 256bit bus (using a faster memory like 18gbps) and add say 128MB of L3 cache.

 

[snc]

Indeed. Hoping they won't do the same mistake as PS4 Pro.

A chiplet design (40CUs *2) for PS5 Pro would work very well. They could keep the 256bit bus (using a faster memory like 18gbps) and add say 128MB of L3 cache.

with 256bit bus they would keep 16gb of ram (as doubt they would go for 32) so 320bit would be preferable and l3 cache would also help

 

[JPT]

So if an upgrade Pro model comes for Series and PS, what are they supposed to solve/bring to the table?
Last time around, it was 4K, but I doubt 8K is the goal this time? FPS is almost always further down the ladder than pretty pixels, so why make them? I doubt FPS and pretty pixels just a couple of years after launch is viable. Of course if MS wants throw money to get into the lead then sure. But it looks more like MS are widening out with streaming and lower specced devices than a new uber device.

Sony could probably not afford to release an uber or PS6'ish like console now, at least in regards to recouping the cost of the PS5....

 

[snc]

So if an upgrade Pro model comes for Series and PS, what are they supposed to solve/bring to the table?

Headache remedy for removing performance vs quality mode dilemma

 

[iroboto]

Indeed. Hoping they won't do the same mistake as PS4 Pro.

A chiplet design (40CUs *2) for PS5 Pro would work very well. They could keep the 256bit bus (using a faster memory like 18gbps) and add say 128MB of L3 cache.

This is an exact match of the 6800XT but with some more redundant CUs then ? I’m not sure if chipsets are needed then

 

[iroboto]

So if an upgrade Pro model comes for Series and PS, what are they supposed to solve/bring to the table?
Last time around, it was 4K, but I doubt 8K is the goal this time? FPS is almost always further down the ladder than pretty pixels, so why make them? I doubt FPS and pretty pixels just a couple of years after launch is viable. Of course if MS wants throw money to get into the lead then sure. But it looks more like MS are widening out with streaming and lower specced devices than a new uber device.

Sony could probably not afford to release an uber or PS6'ish like console now, at least in regards to recouping the cost of the PS5....

Agreed, there needs to be a generational leap in quality and it doesn’t feel like a midgen refresh will get us there

 

[Silent_Buddha]

Agreed, there needs to be a generational leap in quality and it doesn’t feel like a midgen refresh will get us there

I'm not sure outside of waiting 10-15 years that we'll ever see another generational leap like we used to see prior to this generation (PS1 -> PS2, PS2 -> PS3, PS3 -> PS4). I'm still unconvinced that the PS4 -> PS5 generation is going to result in anything close to a generational leap similar to prior ones.

We'd need one of the following to happen, IMO.

• Consumers to accept a machine closer to Apple pricing (IE - over 1k USD, likely 1.5k - 2k USD consoles) so that the SOC can be very large (multiple times the size of the current SOCs) and on the latest silicon node, while staying in a similar power envelope.
• Consumers to accept a machine that consumes significantly more power (IE 300-400 watt consoles) in order to keep manufacturing costs down.
• Some new breakthrough that allows for either cheaper and more efficient silicon chips or a breakthrough that allows something completely different to be used in the manufacture of computing chips.

Short of those things, it's going to be a long LONG time, IMO, before we see anything approaching past console generational scaling.

Regards,
SB

 

[vjPiedPiper]

I'm not sure outside of waiting 10-15 years that we'll ever see another generational leap like we used to see prior to this generation (PS1 -> PS2, PS2 -> PS3, PS3 -> PS4). I'm still unconvinced that the PS4 -> PS5 generation is going to result in anything close to a generational leap similar to prior ones.

We'd need one of the following to happen, IMO.

• Consumers to accept a machine closer to Apple pricing (IE - over 1k USD, likely 1.5k - 2k USD consoles) so that the SOC can be very large (multiple times the size of the current SOCs) and on the latest silicon node, while staying in a similar power envelope.
• Consumers to accept a machine that consumes significantly more power (IE 300-400 watt consoles) in order to keep manufacturing costs down.
• Some new breakthrough that allows for either cheaper and more efficient silicon chips or a breakthrough that allows something completely different to be used in the manufacture of computing chips.

Short of those things, it's going to be a long LONG time, IMO, before we see anything approaching past console generational scaling.

Regards,
SB

I think we'll continue to see more of the same really, although I'm not hopefully for the next after next generation.
But a 350mm² worth of zen5 + rdna4, in 5nm or smaller, should result in a pretty big leap from where we are today.
That probably gives you double just about everything, perhaps not double memory BW, but likely using more on chip cache to compensate.

 

[JPT]

Headache remedy for removing performance vs quality mode dilemma

Which the average consumer does not care an iota about
But how does that make Sony/MS more money and recoupe the investment into PS5 and Series devices?
To me, it looks like an endeavour that is pure loss for the companies. Especially since the current-gen is so supply limited they barely have made a dent in recouping the investment that went in to it.

 

[Globalisateur]

with 256bit bus they would keep 16gb of ram (as doubt they would go for 32) so 320bit would be preferable and l3 cache would also help

You need to think about (hardware) backward compatibility with PS4 and PS5 games. By the way they could simply double the quantity of ram with clamshell mode but I think their solution to the problem could be a faster I/O bandwith. Like ~20GB/s instead of ~10GB/s (all compressed numbers). Also add some DDR4 ram for some OS functions could save them like 1GB.

This is an exact match of the 6800XT but with some more redundant CUs then ? I’m not sure if chipsets are needed then

If they can do chiplet (if AMD tech is ready) they'll do it as it should dramatically improve yields.

 

[davis.anthony]

I'm not sure outside of waiting 10-15 years that we'll ever see another generational leap like we used to see prior to this generation (PS1 -> PS2, PS2 -> PS3, PS3 -> PS4). I'm still unconvinced that the PS4 -> PS5 generation is going to result in anything close to a generational leap similar to prior ones.

Horizon Zero Dawn (PS4) to Horizon Forbidden West (PS5) is as big of a jump as I've seen on console when viewing both on a 4k display.

 

[Karamazov]

Horizon FW looks awesome on PS4/pro, and while looking better on PS5, it was not a massive jump.
I think the difference between Ratchet PS4/pro and Ratchet PS5 was much bigger.

 

[davis.anthony]

Horizon FW looks awesome on PS4/pro, and while looking better on PS5, it was not a massive jump.
I think the difference between Ratchet PS4/pro and Ratchet PS5 was much bigger.

Base PS4 which is sub 1080p looks a full generation behind the native 4k PS5 version on a 4k display and easily up there with general jumps.

The problem is most have not played or seen both or they've seen a DF video on a phone screen and think they know what the difference is.

Ratchet is also a very good leap over PS4.

 

[snc]

You need to think about (hardware) backward compatibility with PS4 and PS5 games.

You dont have to use same bus width to be bc compatible with ps5 and ps4.

 

[snc]

But how does that make Sony/MS more money and recoupe the investment into PS5 and Series devices?

Probably similarly as ps4pro which sells far more less then base ps4 and still sony see value in it:

In the PS4 Pro section, Executive Vice President of Hardware Engineering Masayasu Ito says that Sony's strategy with the PS4 Pro could be extended with the PS5: "In the past, the life cycle for a new platform was 7 to 10 years. But given the very rapid development and evolution of technology, it is more a question of a six- to seven-year cycle. And on top of that, we can't fully keep up with this rapid development of technology," he says.

 

[davis.anthony]

I imagine with ray tracing now gaining speed PS5 (And Series consoles) will age at a faster rate than base PS4 and Xbox One S did so a PS5 Pro model will be more appreciated and needed to keep up with technology more than PS4 Pro was.

 

[PSman1700]

I'm not sure outside of waiting 10-15 years that we'll ever see another generational leap like we used to see prior to this generation (PS1 -> PS2, PS2 -> PS3, PS3 -> PS4). I'm still unconvinced that the PS4 -> PS5 generation is going to result in anything close to a generational leap similar to prior ones.

We'd need one of the following to happen, IMO.

• Consumers to accept a machine closer to Apple pricing (IE - over 1k USD, likely 1.5k - 2k USD consoles) so that the SOC can be very large (multiple times the size of the current SOCs) and on the latest silicon node, while staying in a similar power envelope.
• Consumers to accept a machine that consumes significantly more power (IE 300-400 watt consoles) in order to keep manufacturing costs down.
• Some new breakthrough that allows for either cheaper and more efficient silicon chips or a breakthrough that allows something completely different to be used in the manufacture of computing chips.

Short of those things, it's going to be a long LONG time, IMO, before we see anything approaching past console generational scaling.

Regards,
SB

Yeah its the general assumption made by the general public aswell. It doesnt really matter anyway, generational leaps have gotten small this time around, though its enough to warrant a new console still.

Horizon FW looks awesome on PS4/pro, and while looking better on PS5, it was not a massive jump.
I think the difference between Ratchet PS4/pro and Ratchet PS5 was much bigger.

HZD to HZFW is abit off anyway as a comparison. Its not a generational leap by any means as we have gotten before (PS1 to PS2 etc). Also usually one compares the last of the PS4's best looking games and then into the new generation what looks best. Later games usually look better, not really huge differences, but refined to say the least.
Rift Apart sure does make a better impression, while not the leaps were used to, it does strike more of a leap than what FW does. Then again FW is more of a openworld title than rift apart is.

Regarding mid-gen refreshes, we will most certainly (as per DFs assumptions) not see one this generation. Theres no '8k' desire and/or enough TV's supporting it, upscaling has gotten more common and theres diminishing returns in hardware for the consoles. Not to forget that this generation has a very slow start, sales just arent doing well atm.

 

[JPT]

Probably similarly as ps4pro which sells far more less then base ps4 and still sony see value in it:

To easy, last time around it was 4K displays, that is something the average person actually can grasp. But there is nothing like that, this time. RT? Most can not even see it without slow motion video or screenshoots from DF? More geometry? Again, most have no clue and so and so.
Which technological progress has the same accessibility for average users as the 4K tv, in this generation?
I know most people on here can see a pixel that is shaded wrong from a mile a way, but you guys/gals are not average in this area.

 

[Karamazov]

Give me the option to link two ps5 for VR, each one rendering one "eye" screen, to have the same fidelity as standard games.