PlayStation 4 (codename Orbis) technical hardware investigation (news and rumours)

[Shortbread]

ACP stands for Audio Coprocessor.

 

[onQ]

ACP stands for Audio Coprocessor.

I remember someone telling me that the A did not stand for Audio in ACP.


Edit:
Found it.

Doesn't stand for accelleration co-processor.
Doesn't stand for audio co-processor either.

Edit again: it seams that ACP was just the code name for True Audio

7) It's for our TrueAudio Technology (TrueAudio is codenamed "ACP" and it runs a background service that helps manage audio tasks in games.

http://www.overclock.net/t/1485897/...river-available-for-download/70#post_22271409

 

[MrFox]

The CUH-1100 revision is out, there's a few minor tweaks, new chip revisions, but it seems to be just cost cutting. They slimmed down the regulators, CPU is now single phase and GPU 3 phases, I wonder if that means the new SoC revision have an improved stability. The PSU now has only 1.5A of 5V instead of 3A. DDR3L instead of DDR3 for the south bridge. The split density heat sink is gone, that's too bad, it was the weird fun thing hard to understand. I'm sad about this for some irrational reason.

https://translate.google.com/transl.../pkns/2014/10/ps4cuh-1100-cf6.html&edit-text=

 

[Grall]

Since power dissipation seems to be down, it's no longer necessary with as complex a heatsink it would seem. Would be interesting to see an ampmeter hooked up to this revision to check standby/idle/movie playback/gameplay conditions.

 

[MrFox]

Since power dissipation seems to be down, it's no longer necessary with as complex a heatsink it would seem. Would be interesting to see an ampmeter hooked up to this revision to check standby/idle/movie playback/gameplay conditions.

They tested it, they said it's the same power consumption within 1 or 2 watts.

I wonder if they changed the position of the fan to reduce the effect of high/low pressure/flow areas they had before. There's very little wiggle room there.

 

[Grall]

They tested it, they said it's the same power consumption within 1 or 2 watts.

Well, that's a bit disappointing then, but probably to be expected until there's a process shrink of the APU...

Either they realized the current cooling solution was overkill, or they optimized airflow paths etc further - or heck, both.

 

[loekf]

Well, that's a bit disappointing then, but probably to be expected until there's a process shrink of the APU...

Either they realized the current cooling solution was overkill, or they optimized airflow paths etc further - or heck, both.

The typenumber of the APU changed it seems, but probably just a minor change to improve yields (?) or better power distribution => less voltage drop => less VDDC connections needed => cost saving on PSU ?

 

[Globalisateur]

Sunspider & Kraken tests before and after 2.0

I have done Sunspider 1.0.2 and Kraken 1.1 tests on PS4 for both 1.76 and 2.0 versions. The 1.76 tests were done just before the 2.0 (the 2.0 was downloading during the 1.76 tests).

Sunspider 1.0.2
Nothing running, firmware 1.76 : 776ms, 804ms, 800ms
Nothing running, firmware 2.0 : 979ms, 970ms, 975ms

Game running (fan ~100%), firmware 1.76: 915ms, 915ms, 912ms
Game running (fan ~100%), firmware 2.0 : 1129ms, 1116ms

Kraken 1.1
Nothing running, firmware 1.76 : 17017ms, 17051ms, 16928ms
Nothing running, firmware 2.0 : 13520ms, 13956ms, 13452ms

Game running (fan ~100%), firmware 1.76: 22533ms, 22310ms, 22177ms
Game running (fan ~100%), firmware 2.0 : 15282ms, 15313ms, 15287ms


So as you can see, there is a quite big difference before and after the 2.0 update. One test improves (Kraken), but Sunspider in the 2.0 version does worse than before.

I can't even begin to explain those strange results. Sunspider runs ~20% worse than before but Kraken runs ~25% better than before.

 

[Grall]

What would the real-world implications of this be, IE, what sort of workload do the various benchmarks represent?

If nothing, IE, entirely theoretical, then who really cares? Kind of like most every 3dmark, okay, you got 30k (or whatever) points. So what...? Heh.

Anyway, thanks a lot for your thorough investigation!

 

[Deleted member 11852]

I can't even begin to explain those strange results. Sunspider runs ~20% worse than before but Kraken runs ~25% better than before.

Anything going faster has likely benefited from optimisations. Anything running slower is likely due to a security tweaks. It may also have been optimised but it could be the security patch negated the performance improvements.

If you look at the comparative speed tests of Chrome, Firefox and Safari you'll find that optimisations and security patches these are explain virtually all of the performance differences.

There was a time when Firefox was consistently faster than Chrome and Safari but as the security has taken priory over performance, things have changed.

Other browsers are available.

 

[scf]

A vulnerability has been found in PS4 browser and 2.0 fixes it. Might be that they sandboxed the browser or something, in other words improved security, but lost in performance a little.

 

[Deleted member 86764]

Someone should probably test power draw for each update. I noticed that my PS4's fan was blowing much louder yesterday.

 

[pMax]

There was a time when Firefox was consistently faster than Chrome and Safari but as the security has taken priory over performance, things have changed.

Other browsers are available.

Firefox does not sandbox, chrome does, even if it not good enough at.
There was a time a full passive sandbox for firefox was available for windows (jail-like), yet nobody cared at all for it.

Which other open-source browser do you refer about?

 

[patsu]

Anything going faster has likely benefited from optimisations. Anything running slower is likely due to a security tweaks. It may also have been optimised but it could be the security patch negated the performance improvements.

If you look at the comparative speed tests of Chrome, Firefox and Safari you'll find that optimisations and security patches these are explain virtually all of the performance differences.

There was a time when Firefox was consistently faster than Chrome and Safari but as the security has taken priory over performance, things have changed.

Other browsers are available.

... and power usage.

 

[Deleted member 11852]

Firefox does not sandbox, chrome does, even if it not good enough at.

I didn't mention sandboxing

... and power usage.

I don't follow the perceived correlation between browser performance and power usage, can someone explain this? I know it's a metric that Apple have been touting for Safari and their Nitro javascript engine but this is to demonstrate Safari being more power efficient than other browser, which impacts battery life in laptops.

 

[patsu]

They optimized Safari to be power efficient and performant at the same time. There is a tension between the 2 metrics.

 

[Deleted member 11852]

They optimized Safari to be power efficient and performant at the same time. There is a tension between the 2 metrics.

Yeah, full sandboxing between tabs in S8. But why are people using web browser benchmarks on consoles to test power draw?

I don't get it.

 

[patsu]

I think Apple test power draw for everything, not just Safari. So it shows in the web browser too (most notably during video playback).

In general, if the browser is power efficient, whatever that's running at the same time is less impacted. It should be useful for PS4 too.

 

[Deleted member 11852]

I think Apple test power draw for everything, not just Safari. So it shows in the web browser too (most notably during video playback).

Yup, clicking on the battery icon in the OS X menu bar will show any apps using significant amounts of energy and you can get detailed energy usage from the Activity Monitor (jazzed up version Task Manager in Windows).

In general, if the browser is power efficient, whatever that's running at the same time is less impacted. It should be useful for PS4 too.

This is what I don't get. Apple take considerable time and effort to build Safari specifically for power considerations. Sony have no incentive to do this with their browser so I don't think this tells you anything useful.

 

[patsu]

Yup, clicking on the battery icon in the OS X menu bar will show any apps using significant amounts of energy and you can get detailed energy usage from the Activity Monitor (jazzed up version Task Manager in Windows).

I didn't know that !

This is what I don't get. Apple take considerable time and effort to build Safari specifically for power considerations. Sony have no incentive to do this with their browser so I don't think this tells you anything useful.

It depends on the design and what's already in WebKit. Wondering if Sony will benefit from Apple's work one way or another.