Heard about this a while back and finally experienced it myself.

Ran my battery down to about 20%, plugged in the PSU to charge it and noticed a dramatic drop in performance whilst th ebattery was charging.

This equates to roughly 500 3DMArk 03 points, or a 10-15 FPS drop in the game I was playing at the time.

As soon as I removed the battery the figures shot up again.

Anyone else confirm this?

no one?

hmm i’ve no clue

oh no.. this post is for Type S :shock:
haha srry i didn’t see that.. i dont have type S.. only got TR..
so srryyyy

Sounds a bit worrying… Will see what i can come up with tomorrow.. Will have to get something to capture benchmark with…

Yeah, heard about this over at notebookforums a while back, certainly sounds like something all type S’ suffer from, no idea why though.. since then i make sure my S is either fully charged or take the battery off when playing games… not a show stopper but a bit of a pain.

Interesting - my battery is now fully charged, all power saving features have been disabled completely and I’m looking at a scene in doom 3 whilst running Fraps for an up to date FPS readout. From the 16 fps i was getting whilst charging the readout has now risen to 27, then in a few seconds drops dowm to 18, then back up to 27..and so on -this is without moving or doing anything in the game at all.

After about 5 mins of this the fps stays on 27. Must have been still charging the battery even though the ‘charging’ lamp had gone out.

Really weird thing is - with a fully charged battery and power saving disabled the fps drops down to 9fps when using battery power only.

What on earth is going on I wonder? Seems like power saving is still in operation even though it is fully disabled.

Using glxgears in linux i can get to 13xx fps, on battery power and the CPU half speed (Throttling at T4).  This is great performance…

If I boot up charging then it gets down to 9xx fps, with the CPU at full speed (throttling at T7).

If booting up with batteries and then pluggin in the power cord… the machine crawls and it gets down to 2xx fps, although supposedly the processor is at full speed.

I found this playing with the accelerated 3D driver for linux provided by ATI. ACPI implementation in Linux is still somewhat primitive so I may be confusing the info about CPU speed/or is not being read correctly, but now that the problem is also apparent in Windows I am more worried.

Have you tried benchmarks that do not use the graphics at all? My suspicion was the graphics chip at the start: if the CPU was at full power and generating heat maybe the graphics chip itself was slowing down due to heat. However, this phenomena was the same whether the machine was cold or hot…

So it is not OS dependant..

Well, this is no joke. With the computer unplugged I am getting performances 5 times as fast as with the computer plugged in, using pure CPU benchmarks. I am using the Bytemark benchmarks.

I was fixated in thinking the problem was the ATI chip because I was trying to speed up my 3D simulations… but it seems the problem is the CPU.

Very weird. I did not understand what “Throttling” meant. I assumed, since level T7 was present when plugged in then it was the fastest and T4 was middle because that the setting automatically selected when on batteries….

However, it is the other way around! T0 is the fastest setting (the CPU is not allowed to idle). Forcing it to that state i get my performance back (and a loud Fan).... but it means that the low level hardware seems to be setting the things backwards.

However, as I said before, ACPI control in Linux is under development and all this may be a Linux related artifact.

hmm - ok, how are others getting on? I think adding elements such as linux into the equation may not be helping too much at this stage, as few of us use the OS and Sony wouldn’t listen to us if we identified a problem I guess.

Hopefully Gr00vy may address this issue in his review of the machine?

Wadham,

The only reason I keep Windows installed these days it to be able to troubleshoot OS problems. If I have a problem in both OSs then I know the issues are BIOS/Hardware related. This problem seems to fall in that category.

In this case, I have found a solution that works for me. Linux allows me to go to the raw interface to the ACPI controls. What I found indicated some confusion about the powermodes regarding plugging/un-plugging. If I were you then I will try to force Windows to force the S to behave as it were on batteries even if plugged in, to see if that improves the situation. i.e., I will experiment with “non obvious” power settings since there seems to be a confusion there.

Having said that, the disclaimer still stands: Linux’s implementation of ACPI is incomplete, and my understanding of the whole system lacking.

              Victor

Thanks Victor - point taken.

I hope Sony releases a BIOS patch to fix the issue soon, especially if many of us have this issue!

Sony’s official line is that they tested it with the battery in, and with the battery out on mains power and did not notice a problem.

I explained that the problem only occurs when charging, and that they would need to test the machine whilst the battery is charging.

They called me back today :shock: to say that the engineers have refused to do another test as “the problem should not occur with this model and if it does then you will have to send in your machine for repair, sir” “We don’t need to test it because the problem does not exist.”

Erm.

What can you say to that then?

Not much really.  Aside from the classic, ‘Yeah, it blew up and burned down
my house, you idiots, that’s why the problem DOESN’T EXIST, right?’

But that’s REALLY odd. Wait, I get it.  It’s Sony TR World Forum’s way of saying, “How dare you go to the dark side of the S Type! Return to the land of your forefathers, the TR!”

Anyway, I’ll go off and now and take my Valium.  :twisted:

~nox

It’s official - it’s a ‘feature’ of the machine, and apparently all laptops do it.

Like hell they do - my TR doesn’t and neither does my Toshiba.

I’ll be returning the machine shortly to PC World, and maybe get something different.

I tried and tried, but that response is ridiculous.