Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]

Status
Not open for further replies.
Seems like new 5600XT cards are already using N7P process. They are using 192bit bus instead of 256bit one, but wattage is definitely down quite a bit.

https://www.techpowerup.com/review/asus-radeon-rx-5600-xt-strix-oc/

Strix OC 5600XT
  • Boost clk 1750MHz
  • 36CUs
  • average TDP - 142W
power-gaming-average.png


Best perf per watt as well.

performance-per-watt_3840-2160.png


For comparison sake ASUS Strix 5700 OC with same 1750MHz boost clock and 36CU will reach 179W while 5700 at ~1650MHz will be at 166W.

https://www.guru3d.com/articles_pages/asus_radeon_rx_5700_rog_strix_review
 
Last edited:
This is the patent. The controller is custom and optimized for read speed.

https://www.resetera.com/threads/ps...nd-sonys-ssd-customisations-technical.118587/

I said it since June 2019

https://www.resetera.com/threads/ne...-ot5-its-in-rdna.120059/page-31#post-21424116

The difference compared to june 2019 I have the SSD speed, the bus, and some other things and it is not this. Maybe it is replaceable by a proprietary SSD by Sony or a standard NVme SSD but the controller is not a Phison controller. Maybe the controller respects the NVME standard but it is not a Phison controller at all.

A good question why do you want a very expensive PCIE4 SSD when you can compress the data and go for a PCIE3 SSD?

This OQA leak is false and a totally made up things because someone heard the SSD is fast.

EDIT: The last SSD patent is 2016, this is not WIP since a long time.
No it won't be a phison I agree the rumor is crap, and it might have a very simple controller. But it could still be on a standard nvme socket to allow upgrades and repairs.

For that to work we need nvme 4.0 to have a direct IO mode as a standard feature and then we can put any consumer drives to upgrade. It's something they want for server so it's not that far fetched. It's called OpenChannel SSD.
 
Last edited:
Seems like new 5600XT cards are already using N7P process. They are using 192bit bus instead of 256bit one, but wattage is definitely down quite a bit.

https://www.techpowerup.com/review/asus-radeon-rx-5600-xt-strix-oc/

Strix OC 5600XT
  • Boost clk 1750MHz
  • 36CUs
  • average TDP - 142W
power-gaming-average.png


Best perf per watt as well.

performance-per-watt_3840-2160.png


For comparison sake ASUS Strix 5700 OC with same 1750MHz boost clock and 36CU will reach 179W while 5700 at ~1650MHz will be at 166W.

https://www.guru3d.com/articles_pages/asus_radeon_rx_5700_rog_strix_review
Of the 24W difference, the memory could explain up to half of the difference. GDDR6 is power hungry.
 
No idea why some want it to be custom/exotic, only complicates things and most likely worse performance.

why take something and spend time and money to make a worse performance? That makes no sense at all...any changes will be to improve performance that is important to PS5
 
Of the 24W difference, the memory could explain up to half of the difference. GDDR6 is power hungry.
I doubt two additional 14Gbps chips would consume anywhere close to 12W. Most likely not even half of that, so with higher boost and same amount of CUs its hard to explain 24W difference just by having 2GB less memory.
 
I doubt two additional 14Gbps chips would consume anywhere close to 12W. Most likely not even half of that, so with higher boost and same amount of CUs its hard to explain 24W difference just by having 2GB less memory.
Avarage clock of ASUS Strix 5700 in techpowerup review is 1723mhz and rx5700 1692mhz so 1.8% higher ;d
 
No it won't be a phison I agree the rumor is crap, and it might have a very simple controller. But it could still be on a standard nvme socket to allow upgrades and repairs.

For that to work we need nvme 4.0 to have a direct IO mode as a standard feature and then we can put any consumer drives to upgrade. It's something they want for server so it's not that far fetched. It's called OpenChannel SSD.

I know how openn Openchannel SSD work it means using the CPU host to do what is done inside the SSD. And Sony did exactly the opposite. They improve the SSD for reduce I/O pressure on the CPU because it is a bottleneck.

Here they did differently and like Remedy told it helps reduce the I/O pressure on the CPU side.

With the extra burden of a detailed destruction system, it’s safe to say we hit the limit on what could be achieved on older CPUs. When it comes to the PS5, faster hardware is always appreciated and will make life easier in the short term. But it’s the new SSD that really stands out; essentially streaming will become something that we don’t really have to worry so much about and it will free up some extra CPU bandwidth in the process.
 
Avarage clock of ASUS Strix 5700 in techpowerup review is 1723mhz and rx5700 1692mhz so 1.8% higher ;d
Yea, but I would say 24W is relatively big difference considering 2 GDDR6 chips at 14Gbps shouldnt cross ~5W mark (I think GDDR5x was 2.2-2.5W per chip, and it is more power hungry then GDDR6).

In the end, projected gain from N7 to N7P is 10% lower TDP at iso power, or 7% higher performance for same power, so we should be looking at around 16W lower TDP in 5700 case. In theory ofcourse...
 
  • Like
Reactions: snc
SC load 11 times faster on SSD. It is the same during the demo they present to investors. And the demo to the Wired Journalist it is 19 faster on a low speed devkit.

Completely different games, then spiderman can't be compared to SC's detail, amount of assets, textures etc.

A single sentence from the OsirisBlack heretic who's info must obviously be false because he says The Gospel doesn't tell the full story so we must nitpick all the words he writes:

Klee down, another raises, and we happily hang on to them.

"It's time to play" was already being used by PS in adverts back in November.

I suspect this is Klee himself, going on with the trolling, and people happily fall for it, so desperate.

why take something and spend time and money to make a worse performance? That makes no sense at all...any changes will be to improve performance that is important to PS5

No idea, ask Sony, they have done it before.
 
I know how openn Openchannel SSD work it means using the CPU host to do what is done inside the SSD. And Sony did exactly the opposite. They improve the SSD for reduce I/O pressure on the CPU because it is a bottleneck.

Here they did differently and like Remedy told it helps reduce the I/O pressure on the CPU side.
Not necessarily by the CPU, it could be by the SB or by asic sections. There has to be something custom somewhere managing the FTL, but it doesn't prevent them from using OCSSD 2.0 standard drives (if the standard becomes part of nvme).
 
Last edited:
The console design should be almost finished by now. This card 7.19 TF but it is huge.
So what does AMD have in store we dont know where they managed a 12TF in the XSX in a specific form factor?
 
The console design should be almost finished by now. This card 7.19 TF but it is huge.
So what does AMD have in store we dont know where they managed a 12TF in the XSX in a specific form factor?
I'm not convinced it's 12TF, nor that the ps5 is 9.2TF.

I won't be surprised if they end up being 10 and 7.5 respectively, unless some RDNA2 magic happens, or if they use a dynamic clock fuzzying the math.
 
The console design should be almost finished by now. This card 7.19 TF but it is huge.
So what does AMD have in store we dont know where they managed a 12TF in the XSX in a specific form factor?
The specifics of this card have done nothing but make higher TF counts more feasible due to the reduced power consumption shown. The physical dimensions on this card have 0 bearing on console designs. Consoles only care about die size and total thermal flux.
 
Status
Not open for further replies.
Back
Top