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

[PSman1700]

I pretty sure the CPU had nothing to do with the resolution.
Yes it does drop below it on PS4 but a 2500k at 4.2ghz is a lot more powerful than the jaguar in the consoles all things being equal the 2500k should of easily maintained 60fps.

Sometimes resolution has to do with FPS, but i have no idea how that had an effect on PS4 BF1. I think if you mean MP, its kinda hard to compare anyway, try SP on both. In BF4 i had an awful lag/stutter on siege PS4, TDM, when campining with the USAS on the top floor. On a 3550 at stock settings it didnt lag under the same circomestances, whatever that means, since online MP matches vary rather much.
PlanetSide 2 is a good test in that case, it is extremely CPU bound, about all performance is dependend on the CPU. In that game, i get as low as 20fps, or even lower, when in combat areas. I'm sure it is the CPU to fault for. My i7 920 @stock handles it much and much better (close to 60 for most), OC'd to 3.6ghz there are almost no problems regarding FPS. Ive had this CPU since late 2008.
Its a free game on both PS4 and PC. DF did some testing, they got as low as 26fps, ive had lower then that, it almost never pushes above that when in combat.

https://www.eurogamer.net/articles/digitalfoundry-2015-planetside-2-performance-analysis

Like said, those jaguars, in special @ 1.6ghz are rather slow, and for being what it is perform rather well, again for what it is. If the jaguar as in the base machines is comparable to a Q6600 in raw performance overall, less overhead and drawcalls dont lift it to much higher then that i think.
Also, wasnt DX12 supposed to help alot with the draw calls problem?

https://www.eteknix.com/dx12-can-handle-600k-draw-calls-increasing-amd-gpu-performance-600/

 

[Xbat]

Sometimes resolution has to do with FPS,

Yeah but that's graphics card related not CPU.

If the jaguar as in the base machines is comparable to a Q6600 in raw performance overall,

Yes but I'm positive if my 2500k couldn't maintain 60fps in battlefield 1 there's no ways a q6600 could.
What I'm saying is not that the jaguar is More powerful than people think but because it's in a console they were able to get much more from it than I thought possible.

 

[Dictator]

Yes but I'm positive if my 2500k couldn't maintain 60fps in battlefield 1 there's no ways a q6600 could.
What I'm saying is not that the jaguar is More powerful than people think but because it's in a console they were able to get much more from it than I thought possible.

Hm. Well on 64 player games consoles are really no where near 60 in BF, especially on base.

You tried out DX12 then? We have measured extreme CPU gains with that in use.

 

[Svensk Viking]

Sweclockers did a short re-test of the old CPUs last year.
https://www.sweclockers.com/test/28...re-i5-750-tar-nehalem-till-massorna/5#content

Many other sites always test at Ultra, Sweclockers did test most runs on low settings. The i5 750 and i7 920 should definitely still provide a better experience than on consoles.
It's hard to find good tests for Battlefield though, the tests in single player are very misleading for CPU performance.

 

[PSman1700]

Yes but I'm positive if my 2500k couldn't maintain 60fps in battlefield 1 there's no ways a q6600 could.
What I'm saying is not that the jaguar is More powerful than people think but because it's in a console they were able to get much more from it than I thought possible.

My 3550 at stock handled it fine, 60fps as a minumum, 64 slots servers. That CPU is slower then your 2500 @ 4.2ghz.
Q6600/i5 from 2008 is about on par with the jaguars, and should offer about the same experience as the consoles, when accounted for the same settings and fps. Any BF on console delivers a sub-par MP experience. Just try a 64 slot server with everything going on around you. That must be a CPU limitation.

The i5 750 and i7 920 should definitely still provide a better experience than on consoles.

Yeah they should, as both are much more powerfull CPU's then the jaguars in the base and premium consoles. This all goes to say how much underpowered the current gen is in terms of CPU. With next gen, consoles are going to get a CPU that matches mid-end parts.

 

[Xbat]

You tried out DX12 then? We have measured extreme CPU gains with that in use.

Can't remember really, I think directx 12 was buggy or there was some reason like the menu or something being super laggy when I played it.

 

[Xbat]

Hm. Well on 64 player games consoles are really no where near 60 in BF, especially on base.

This is not me saying the game ran better on console it's me saying I expected more from the 2500k if a jaguar could get 40fps.

 

[PSman1700]

This is not me saying the game ran better on console it's me saying I expected more from the 2500k if a jaguar could get 40fps.

That 2500 is/was probably giving you a much better experience then the consoles ever could give you in a like for like situation/settings etc.

 

[Deleted member 13524]

Bloomberg has sources that presumably are specifically talking about the PS5 final design. The github data is not conclusively connected to the final PS5 design. It's that simple.

Aaaand you just invited the wrath of the inquisitors to your doorstep.

Are we back to this again? Here? Isn't there another place to discuss that already?

No! The Inquisition must act on all fronts!
How would they ever be able to spread The Gospel Truth if they were contained in just one thread?

Anyway, the bloomberg leaks only support the github leak.

Ok now I'm genuinely interested in knowing how a Bloomberg article that says:
- The PS5 is too expensive, BoM is at $450.
Is supporting of your Faith that so far said:
- The PS5 uses a smaller GPU because Sony wants to undercut Microsoft in price and it'll be $399 tops.

This sounds like all the religious leaders who started their own mental gymnastics when we got on planes and didn't see any deities living on the clouds, or Mount Olympus.

9TF 5700 XT 9% slower on average on launch day drivers. PS5' GPU is at least half tier above the 5700XT.

Are next gen consoles going to target 1440p?

 

[PSman1700]

This sounds like all the religious leaders who started their own mental gymnastics when we got on planes and didn't see any deities living on the clouds, or Mount Olympus.

PS5 bom actually is close to what ps4 was, the ssd and gddr6 push it abit above that. The leaks confirm cooling issues, which was to be expected at those clocks. DF said github oberon is most likely ps5. Il take that over any 4chan or resetera banned journalist.

If pro targets 4k with its mere 4TF of gcn, then 9tf navi 1.5 will do more then 1440p for most games, in special with checkerboarding etc. 5700xt has abit more raw power but thats also boost clocks. Overall, ps5 will probably close to a 5700xt, in raw power.

 

[Betanumerical]

PS5 bom actually is close to what ps4 was, the ssd and gddr6 push it abit above that. The leaks confirm cooling issues, which was to be expected at those clocks. DF said github oberon is most likely ps5. Il take that over any 4chan or resetera banned journalist.

If pro targets 4k with its mere 4TF of gcn, then 9tf navi 1.5 will do more then 1440p for most games, in special with checkerboarding etc. 5700xt has abit more raw power but thats also boost clocks. Overall, ps5 will probably close to a 5700xt, in raw power.

Which leaks confirmed cooling issues?.

 

[Deleted member 13524]

PS5 bom actually is close to what ps4 was, the ssd and gddr6 push it abit above that.

Lately there's been a bit of a conundrum over what close means.
Is $370 close to $450? Over 20% difference?
Nah.

If pro targets 4k with its mere 4TF of gcn

The Pro targeted a resolution upgrade over the 1.8TFLOPs base PS4. I.e. it's meant to play 8th gen games with 8th-gen graphics at 1440-1800p plus upscaling.

The PS5 will target 9th gen graphics. That means raytracing, more detail, more shaders running at the same time, etc.

Which leaks confirmed cooling issues?.

None that I know of.

 

[PSman1700]

Is $370 close to $450? Over 20% difference?

PS4 BOM was 381. Also, a equal sized SSD/Nand will be more costly then a 5200rpm hdd in late 2013. DRAM prices are rather high too. Then according to Bloomberg, the PS5 is having a 'more expensive cooling' system. The extra cost to BOM doesnt have to be tied to the APU itself, according to 'leaks' it isnt.
Also, the PS4 BOM was supposedly between 390 and 500 dollars, according to 'leaks' before its launch. The PS4 bom leaks was actually pretty close to what we have leaked now for the ps5.

None that I know of.

https://arstechnica.com/gaming/2020...may-make-ps5-cost-more-than-450-report-finds/

Sony is spending more on the cooling system to prevent overheating. 36CU at between 1.8 and 2Ghz will need some good cooling.

 

[Picao84]

PS4 BOM was 381. Also, a equal sized SSD/Nand will be more costly then a 5200rpm hdd in late 2013. DRAM prices are rather high too. Then according to Bloomberg, the PS5 is having a 'more expensive cooling' system. The extra cost to BOM doesnt have to be tied to the APU itself, according to 'leaks' it isnt.
Also, the PS4 BOM was supposedly between 390 and 500 dollars, according to 'leaks' before its launch. The PS4 bom leaks was actually pretty close to what we have leaked now for the ps5.



https://arstechnica.com/gaming/2020...may-make-ps5-cost-more-than-450-report-finds/

Sony is spending more on the cooling system to prevent overheating. 36CU at between 1.8 and 2Ghz will need some good cooling.

Well if Sony really did go for low CU number and very high clocks, that's an incredibly risky strategy... and probably dumb as well. That's why many find it so hard to believe. If the rumors were low TF but due to higher number of CU but very conservative clocks, say 1.4Ghz, it would sound more credible.

 

[Betanumerical]

Well if Sony really did go for low CU number and very high clocks, that's an incredibly risky strategy... and probably dumb as well. That's why many find it so hard to believe. If the rumors were low TF but due to higher number of CU but very conservative clocks, say 1.4Ghz, it would sound more credible.

This is what I don't understand about the low CU's high clocks rumour. Wouldn't the number of APU's that are actually usable at the end of the day be quite low at very high clocks? (say 2.0Ghz+), I can't imagine many of even the discrete cards reach these clocks and thats before you take into account bad chips.

 

[milk]

Ok now I'm genuinely interested in knowing how a Bloomberg article that says:
- The PS5 is too expensive, BoM is at $450.
Is supporting of your Faith that so far said:
- The PS5 uses a smaller GPU because Sony wants to undercut Microsoft in price and it'll be $399 tops.

To be fair, those don't contradict each other, but rather corroborate. Bloomberg didn't say they were AIMING at over $400, but that they ended up above that price accidentaly. So it only makes sence that they'd have a conservative SOC and design to reach the magical $399 price, eventhough they later end up with a $479 product because of unforseen rise in costs of certain parts.
Both point to to 399 TARGET.
I definetly don't trust any rummor 100%, but I'm just pointing out the Bloomberg narrative and the older github leak fit rather well together.

 

[PSman1700]

Well if Sony really did go for low CU number and very high clocks, that's an incredibly risky strategy... and probably dumb as well. That's why many find it so hard to believe. If the rumors were low TF but due to higher number of CU but very conservative clocks, say 1.4Ghz, it would sound more credible.

Maybe it wont run at 2ghz? 1.8ghz perhaps, the testing could have been done to reach the 9.2TF target sony wants, but the final build will contain more CU's at a lower clock, coming in around 9TF.

I definetly don't trust any rummor 100%, but I'm just pointing out the Bloomberg narrative and the older github leak fit rather well together.

They do. Not only that, but Sony apparently having to invest more in the cooling system to prevent overheating suggests higher clocks.
Another thing, PS4 BOM leaks estimated about the same as PS5 BOM leaks do now. It seems that indeed Sony targetted a 399 machine again, since it was a success.
MS with a 2 SKU approach, scalability between platforms, and PC platform, can afford to have their premium XSX at a higher price.

 

[milk]

They do. Not only that, but Sony apparently having to invest more in the cooling system to prevent overheating suggests higher clocks.

Exactly. I'd even add that both point to an over-optimistic design, because I think a mass produced SOC with a high-end-ish zen2 AND a 2.0GHz GPU (from AMD of all all people) was pretty damn optimistic, and quite prone to "woopsies, this ended up more expensive than we had projected"

 

[Picao84]

Maybe it wont run at 2ghz? 1.8ghz perhaps, the testing could have been done to reach the 9.2TF target sony wants, but the final build will contain more CU's at a lower clock, coming in around 9TF.
.

I'm sorry but if you open the door for the GitHub leak to not be the final build and for another one to have more CUs why would still be 9TF? Either you believe that GitHub is the final chip and is 9TF, or you believe it's not and all bets are off then.

 

[PSman1700]

I'm sorry but if you open the door for the GitHub leak to not be the final build and for another one to have more CUs why would still be 9TF? Either you believe that GitHub is the final chip and is 9TF, or you believe it's not and all bets are off then.

The github leak only proofs what the target specs are, the final chip will be in the retail units. It is a testbed for a reason, the leak does imply 9.2TF, but they could alter the design somewhat, not from 36CU to 52, or from 2ghz to 1.5. Slight adjustments i ment.

Exactly. I'd even add that both point to an over-optimistic design, because I think a mass produced SOC with a high-end-ish zen2 AND a 2.0GHz GPU (from AMD of all all ple) was pretty damn optimistic, and quite prone to "woopsies, this ended up more expensive than we had projected"

Rather the NAND/DRAM prices that went up are to blame, if those articles are to be believed.