Don't think it'll be that cheap. Isn't the Xbox One S All-Digital Edition £199.99 (~$260)?
Don't think it'll be that cheap. Isn't the Xbox One S All-Digital Edition £199.99 (~$260)?
lol no.
https://www.eurogamer.net/articles/digitalfoundry-2019-is-amd-gonzalo-the-ps5-processor-in-theory
Look at the bold, they simply don't know.
Because it can play traditional games too..notice I used the word 'intended'.I don't see why you'd have this sort of a device. For streaming you go with a cheap dongle-ish device. Your streaming box at this price would just kill the chances of success of your streaming business.
The problem is that you need intelligence to effectively manage that shared pool. Enter HBCC.anexanhume
M3rcy
Hey guys, ever thought about that you don't need HBCC if you already have a shared memory pool? With HBCC you add extra complexity and die space to just marry a cost intensive HBM2 with slower DDR4 to get back to a pseudo "shared memory" access. Sounds weird to me. Or maybe I am just to dumb to see the advantages ... (cost, complexity, bandwidth ... you name it)
Richard thinks Anaconda to be expensive premium console like XBOX ONE X. But it seems he doesn't expect PS5 to be expensive premium console like Anaconda, so his expectations is around 10TF for PS5.Doesn't Stadia slot in at 10.7TF instead of the mentioned 9.7TF?
So interpreting the article correctly, Eurogamer thinks the PS5 is about 10TF now and Anaconda is about 12TF?
Richard thinks Anaconda to be expensive premium console like XBOX ONE X. But it seems he doesn't expect PS5 to be expensive premium console like Anaconda, so his expectations is around 10TF for PS5.
Okay we have to define "much": in my opinion it points to about 2TF more.Anaconda can definitely end up being more powerful. I don't really see it being "much more" powerful though.
anexanhume
M3rcy
Hey guys, ever thought about that you don't need HBCC if you already have a shared memory pool? With HBCC you add extra complexity and die space to just marry a cost intensive HBM2 with slower DDR4 to get back to a pseudo "shared memory" access. Sounds weird to me. Or maybe I am just to dumb to see the advantages ... (cost, complexity, bandwidth ... you name it)
do you think ms uses vapor chamber cooling than? I think a 12tfps apu has tdp issues otherwise.Okay we have to define "much": in my opinion it points to about 2TF more.
Do I need to modify my baseline? Even german tech sites are now reporting about a much more powerful Anaconda compared to the PS5. Should I still trust the word of Brad Sams on it or should I follow my gut feeling and say that Anaconda would be more powerful? Decisions, decisions ...
I see the use case, but you don't need HBM2 for having HBCC, you can just do it with GDDR6. Cheaper and as efficient in caching slower memory ... my point was more pointing to the HBM2/DDR4 memory setup rather than HBCC. I should have expressed myself better :/It'll be able to cost reduce like crazy though over time, and I think the HBCC was going in anyway to manage the data streaming from the SSD.
Seems like it. But i have to caution, Richard doesn't know anything about PS5 actual specs like Teraflops, clock speeds, ram etc..
What? 7nm was always expected.When I talked about 7nm PS4 people here didn't believe me. Let's see if the reddit post will redeem me.
lol
As unlikely as it may seem I'm still holding out hope for a dual Navi GPU in a chiplet formation :) So if a single Navi can give us say, 12TF, I'm hoping for 18-20 TF (factoring in inefficiencies and all) for a dual setup.
Single stack HBM3.It'll be able to cost reduce like crazy though over time, and I think the HBCC was going in anyway to manage the data streaming from the SSD. HBCC is going to be a part of Navi anyway, because of it's applications in the enterprise space. It's actually not a matter of putting it in to a Navi-based SoC, it's a matter of not taking it out. :)
I will give the rumor more time to grow before I go back to 3 levels of performance (which was the initial prediction just at a lower base).trust your gut. I don't even know what "much more powerful" could possibly mean. From a gamers perspective the difference between 10 TF and 12 TF is probably not even noticeable but from a hardware design standpoint the difference is huge and hard to achieve.
Some people probably think like PS5 is 10TF and XboxNext is like 15TF. Not gonna happen
Don't think it'll be that cheap. Isn't the Xbox One S All-Digital Edition £199.99 (~$260)?
only if nvidia forgos raytracing hardware, which would be dumb. I rather nvidia add mor rt cores at this point.18-20 TFLOPS is about what you'd expect at the top end of PC gaming next year. You're not getting a console with that amount of power, period.
18-20 TFLOPS is about what you'd expect at the top end of PC gaming next year. You're not getting a console with that amount of power, period.
Single stack HBM3.
Does HBCC sufficiently abstract to where you could go down to 2 channels of DDR4 at some point? That would be a neat trick too.
I guess you mean that article on pcgameshardware.de. That's a classic Resetera -> Internet -> resetera. They are explicitly naming their "source": user Klobrille from Resetera.Do I need to modify my baseline? Even german tech sites are now reporting about a much more powerful Anaconda compared to the PS5. Should I still trust the word of Brad Sams on it or should I follow my gut feeling and say that Anaconda would be more powerful? Decisions, decisions ...
I see the use case, but you don't need HBM2 for having HBCC, you can just do it with GDDR6. Cheaper and as efficient in caching slower memory ... my point was more pointing to the HBM2/DDR4 memory setup rather than HBCC. I should have expressed myself better :/
Are all the "leaks" somewhere on Era? There is one of the 1St page. Anything else or it doesn't seem worthy to keep posting them?
I guess you mean that article on PC gameshardware.de. That's a classic Resetera -> Internet -> resetera. They are explicitly naming their "source": user Klobrille from Resetera.
Not too technical, but I did come across this recent Engadget video explaining chiplets. Not a bad video to just get up to speed on the subject.
only if nvidia forgos raytracing hardware, which would be dumb. I rather nvidia add mor rt cores at this point.
To all the people saying that 8-10Tflops would be disappointing. Do you think that GoW is the best looking game ever? If you do, you wont be disappointed by a 8-10Tflops PS5. The PS4 launched with less flops than a what middle-high 2009 AMD GPU's had (A Radeon HD 5850 is rated at 2Tflops).
A 2080 Ti at 1900 MHz is already at 16.5 TFLOPS. No reason to think NV won't be pushing up against 20 TFLOPS after moving to 7nm, imo. Especially for those of us that plan to put them on water.
Right.To all the people saying that 8-10Tflops would be disappointing. Do you think that GoW is the best looking game ever? If you do, you wont be disappointed by a 8-10Tflops PS5. The PS4 launched with less flops than a what middle-high 2009 AMD GPU's had (A Radeon HD 5850 is rated at 2Tflops).
I think the HBM rumor may have some truth. Too many advantages:
- long term cost reduction
- power savings allowing for beefier GPU
- larger total RAM than a unified approach
But I guess it all comes down to the deals that are being made
Not disappointing. Just unrealistically lowTo all the people saying that 8-10Tflops would be disappointing. Do you think that GoW is the best looking game ever? If you do, you wont be disappointed by a 8-10Tflops PS5. The PS4 launched with less flops than a what middle-high 2009 AMD GPU's had (A Radeon HD 5850 is rated at 2Tflops).