Luca's meaningless thoughts  

Fuck! I bought this extremely nice ultrabook, the Toshiba Satellite Z830-10J, about an year ago, and I've been experiencing some problems with CPU frequency scaling.

At one point I looked and looked for kernel bugs without much success. I went through several kernel updates in the hope of this being fixed, but never happened.

It seemed that the problem wasn't so bad after all, because I only got the CPU frequency locked down to the minimum when using the ondemand scaling governor, but the conservative was working apparently OK.

Just a little more latency I thought, is not that bad.

Recently I received an update on a related bug and I thought about giving it another shot. This mentioned something about booting with processor.ignore_ppc=1 to ignore some BIOS warning about temperature to overcome some faulty BIOS, so I thought on trying that.

But before doing, if this were a real BIOS problem, I thought about looking for some BIOS update. And there was one. The European Toshiba website offered only a Windows program to do the update though, but fortunately I found in a forum a suggestion about using the non-European BIOS upgrade instead, which was provided also as an ISO image. The problem is I don't have a CD-ROM, but that shouldn't stop me, I still have USB sticks and hard-drives, how hard could it be? I failed with UNetbootin but quickly found a nice article explaining how to boot an ISO image directly with grub.

BIOS upgraded, problem not fixed. So I was a about to try the kernel parameter when I remembered I saw some other article when googling desperately for answers suggesting changing some BIOS options to fix a similar problem.

So I though about messing with the BIOS first instead. The first option I saw that looked a little suspicious was in:

PowerManagement
   -> BIOS Power Management
      -> Battery Save Mode (using custom settings)
         -> Processor Speed
            <Low>

That is supposed to be only for non-ACPI capable OS, so I thought it shouldn't be a problem, but I tried with <High> instead.

WOW!!!

I start noticing the notebook booting much faster, but I thought maybe it was all in my mind...

But no, then my session opened way faster too, and everything was extremely faster. I think maybe about twice as fast. Everything feels a lot more responsive too. I can't believe I spend almost an year with this performance penalty. FUCKING FAULTY BIOS. I didn't make any battery life comparisons yet, but my guess is everything will go well, because it should still consume very little power when idle.

Anyway, lesson learned:

Less blaming to the kernel, more blaming to the hardware manufacturers.

But I still want to clarify that I love this notebook. I found it a perfect combination between features, weight and battery life, and now that it runs twice as fast (at least in my brain), is even better.

Hope this is useful for someone.

The D compiler doesn't provide any information on the static data that the GC must scan, so the runtime/GC have to use OS-dependant tricks to get that information.

Right now, in Linux, the GC gets the static data to scan from the libc's variables __data_start and _end, from which are not much information floating around except for some e-mail from Hans Boehm to the binutils mainling list.

There is a lot of stuff in the static data that doesn't need to be scanned, most notably the TypeInfo, which is a great portion of the static data. C libraries static data, for example, would be scanned too, when it makes no sense to do so.

I noticed CDGC has more than double the static data the basic GC has, just because of TypeInfo (I use about 5 or so more types, one of them is a template, which makes the bloat bigger).

The voronoi test goes from 21KB to 26KB of static data when using CDGC.

It would be nice if the compiler could group all the static that must really be scanned (programs static variables) together and make its limits available to the GC. It would be even nicer to leave static variables that have no pointers out of that group, and even much more nicer to create a pointer map like the one in the patch for precise scanning to allow precise heap scanning. Then only the scan should be scanned in full conservative mode.

I reported a bug with this issue so it doesn't get lost.