![]() |
|
|||
|
Hi,
Wanted to kickoff this thread, as I feel that it's my turn to learn from Hax. So.., Mr. Hax, please keep us updated while your PS3 is compiling away as to how you've you are customizing your own fresh install of Gentoo without relying on the stage4 tarball. We would loooove to learn! BTW: You can check the build flags on the -G option. Use the -pvG option. The USE flags will show. The USE flag that return after this command will represent how that binary was compiled (which USE flags were used). Conversely, if you use -pv only, you will see how that app will be installed based on default options +, -, your make file configuration. So, -pvG will always show how that binary is installed regardless of your make file settings. Keep us posted on your progress. As I firmly believe that this is the real way of doing it. Afterall, the stage4 tarball that we use are compiled using PowerMac G5 as confirmed by the creator (rangerb) of these tarballs. I'd be real curious to see how well it performs. You're taking a great step for all of us. Not to say this hasn't done before, but I don't think anyone has done this recently, leveraging the latest toolchains. I think people like tokka over in Japan did this first, but back then the toolchains were very immature. |
|
|||
|
Good stuff. One thing I am looking forward to is an update toolchain that allows you to enable more SPUs, without the memory limitation issues during compile time. If I understand correctly there memory issues to be consider for compiling these apps with more SPUs enabled, as well as running these apps after they've been compiled...
After you do all that, you should backup your system as a disk image, and then see how much memory you're actually using at normal usage environment by typing: free I'd like to see some delta of the memory usage. Anyway, all good stuff. I can't wait to hear your results. |
|
|||
|
Yo Hax!
So, what's the verdict? Should we all recompile our own kernel and rebuild everything from ground up? Did you notice any changes in performance? Somehow I doubt it.... Prove me wrong. ![]() |
|
||||
|
Hey whatsup guys, Asong26... Props to you guys for getting it done on the site here.
So after 4 builds of Gentoo, I think my current setup is squeezing the most that I can get out of the PS3 so far, but it is a long process. Gentoo on the PS3 = super slow compiling, but the goal is to have just what I want rather than loading up prebuilt packages. I would caution that it takes perseverance and some Linux knowledge before I would recommend anyone else trying to custom compile their setup. Asosng could easiily do it, but others who are just getting used to Linux and don't want to trial and error setups and don't delve into emerge without the -G or -g, and -umerge and --depclean etc. on installations (re-installations in some cases) but want windows-type software installations, then stay away from this. What are the results? 1. In startup mode in KDE, am I running with less RAM usage than say, Asong26? Tough question, depends on my apps and services in memory compared to Asong26. But I am willing to bet that my KDE base system takes up less memory (physical and logical), but the kernel is probably exactly the same. 2. Do I have all standard WM apps installed? no, just apps that I want. My WM (KDE in my case) apps are very few, by design. Example? My KDE menu for System apps has 3, just 3 apps in the menu. How many does yours have? 3. We should compare apps and services in memory after startup. Remove all user selected apps from startup and lets compare tables and RAM. Lets use KDE as the base WM. I won't custom hack out any KDE basic services, just remove superkaramba plugins and Amarok from my startup. I will post my table here later and Asong26 will put up his. More later. Oh, one thing that I have developed over time is a complete migration to using the package use keywords approach to compiling and updating versus the make.conf use flags. The concept of global use flags versus app-specific use flags is a very important one that I learned the hard way. I believe this concept is the most important thing that I have learned so far in Gentoo. Maybe we should try to develop here a thread of the best make.conf and the best package.use files for the PS3.
__________________
|
|
|||
|
I am pretty content with my current setup (except for the lack of commercial DVD playback), but I would suggest that you make a backup copy of your install for general backup as well as test bed. You can either tar up a stage4 tarball, or you can use an imaging software for compressed drive image. It's called 'partimage' Once the image is created, you can wipe out the partition and install the official stage4 tarball, and see what memory usage is with just the kdebase installed. Once that's done, you can restore your image, and do another test to see if there's any significant differences.
In my environment I have KDE, GNOME and XFCE4 installed. I've compiled XFCE4 with KDE and GNOME integration, so I am able to use this lightweight WM while having the accessibility to either KDE or GNOME app without bogging down the memory when they're not in use. If you haven't done so already, you should give XFCE4 a try. It very fast, and a lot smoother in my opinion, regardless if I am using KDE / GNOME applications with it or not. One of the first things you'll noticed is the mouse movement, and tighter the controls are. Second, the app load times seem to be faster vs. kdebase. Just out of curiosity, how long did it take you to compile the kernel and the necessary systools to have it up and running? |
|
|||
|
BTW:
This is the memory snapshot on above setup, running XFCE-4 with only the kkonsole (kde terminial console) running.... total used free shared buffers cached Mem: 201136 116896 84240 0 732 52868 -/+ buffers/cache: 63296 137840 Swap: 965624 2948 962676 asong@localhost ~ $ What does your memory look like when you've got nothing running except for the command prompt? You should do a comparison of memory usage with or without your current themes w/ widgets on the desktop. Restore your desktop with no back drop, using a simple color and don't open any apps. Let me know... I'd be really surprised if you could beat these numbers using KDE as your WM.... |
![]() |
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
|
|
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Installation Guide for Ubuntu on PS3 | H.axen | Ubuntu Linux | 17 | Yesterday 07:13 PM |
| How to Install Fedora Core 8 | haxen | Fedora Core 8 | 18 | 12-09-2008 01:38 PM |
| Backup your Blu-Ray Disc | haxen | PS3 & PS2 Games | 11 | 11-03-2008 05:57 PM |
| My experience of install Gentoo using the experimental PPC64 ps3 | 955 | Gentoo Linux | 2 | 09-18-2008 01:03 PM |
| How to Install Gentoo Linux | haxen | Gentoo Linux | 40 | 09-09-2007 09:38 AM |