[GTALUG] Fedora 22 Live Workstation Install - no fglrx - no pdftk

[D. Hugh Redelmeier]

| From: Russell Reiter <rreiter91 at gmail.com>

| For now I'v installed lm-sensor which shows:

Those temperatures sure look fine.  I don't know how accurate they
are.  But it's what you've got.

My recollection is that lm-sensors has a large database of motherboards, 
and the plumbing of the sensors for each motherboard. The problem is that 
it is like wikipedia: it may be inaccurate or incomplete.  Even the 
mapping from raw sensor reading to degrees is part of that.

| So far It looks like I don't have the chassis fan plugged into the
| right slot.

The table entry might be wrong.  Look at the board's manual or the
silk screening.

| As you can see from lm-sensor, after the firmware update, I'm within
| operating parameters now. One feature of this motherboard is the
| ability to recover from different bios's. I'm curious beyond making it
| stable now, I may regress the bios and log what was happening at the
| time.

Take your winnings and go!  Unless you have a lot of spare time.

| Apparently ASUS has a MS utility they ship with the board. Ai gear 2,
| which allows overclocking the MB from userspace. The condition seems
| similar to the laptop I experienced where the MS OS had hooks into
| controlling the power management.

Calling it "a MS utility" suggests that Microsoft provided it.  I'm
sure MS had nothing to do with it.  It is true that it will only run
under some MS Windows, but that isn't the same thing.

The "MS OS .. hooks" terminology carries implications that I doubt.
The main way of affecting hardware is through ACPI, which both Linux
and Windows use.  It is true that some ACPI implementations (inside
the firmware) try to detect which OS is being used, but Linux calls
itself Windows, if I remember correctly.  Vendors like ASUS can also
write drivers for Windows that do odd things with the hardware.  But
none of this is by Microsoft as far as I know.

| > What is "CPU access"?
| 
| The little graphic thingy in System Monitor which gives a visual
| output of load as a line graph for each CPU die, as well as for
| Memory/Swap and Network. At this time and at increased resolution the
| load line hovers around 60% as opposed to the old firmware at 800X600
| which showed 85-100% load with video on VLC.

Normally, I'd call this "CPU utilization".  I don't remember how well
it reflects cycles used by the kernel (and there are several ways in
which the kernel uses cycles).

Your computer has one processor chip, a die, but two cores (I'd call them 
two processors, but I'm old fashioned).  The "die" is one of the pieces 
after the silicon wafer is "diced" (cut into bits).  The plural of "die" 
is "dice" but most hardware folks seem to use "dies".

Firmware is mostly cut out of the loop once the system is
bootstrapped.  The major exception is ACPI and perhaps SMM.  Minor
firmware updates should not really change CPU load.  It could affect
thermal throttling (slowing down the CPU to cut down heat generation).
But then cooling should result in *higher* CPU utilization since the
slowed machine has fewer cycles available.

| >
| > | I do note that desktop
| > | sprite animations are slower.
| >
| > What are "desktop sprite animations"?  The mouse pointer?
| 
| Sprites are a term for layers used in animations. So in the case of
| the mouse pointer, it exists on a sprite layer. Any windows animations
| like dragging the widget around aren't dragging a box per-se, they are
| moving the indicator (0,0) to the point where the new sprite is
| created, leaving the old one to be destroyed on the screen redraw.
| 
| In this case in Plasma, when you access the menu it pops up very
| slooooooowly and disappears in the same fashion. I'ts kind of cute.

Sprites historically have been very small patches of pixels and there
usually have been bounds on the numbers of them.  They were from the
era of small and constrained frame buffers.  Think old video consoles.

I have no idea if they remain in current hardware or in current software
abstractions (two different issues).  But I don't expect them to have
much application in non-game situations.  The mouse cursor is it.

When I drag a window, the whole thing moves.  It certainly isn't a
sprite: way too big.

Yeah, there is animation in some desktops.  That's surely not done with 
conventional sprites.  Of course somebody might recycle the term to mean 
something different.

| My error. I used + & - to set the date but tried to type in the time.
| It let me enter the numbers and in fact did display them in the right
| place but didn't commit them.

Odd.  But differently odd.

| At this point I'm still a little at odds with what to do about QT, but
| I can live with this state of things.

After thrashing about a while, leaving unknown randomness in ones
system, sometimes it makes sense to install the system afresh and
apply the distro updates.  The less fiddling one has to do to make it
work, the less fiddling one will have to do in the future.

[Excuse me for the quaint use of "one".  I'd normally use "you" but
that sounds like a command or an accusation.  "One" captures more
accurately what I intend.  Too bad it is vanishing from our
vocabulary.]

I find that it pays to keep a lab book.  I don't always write down
what turns out to be important :-(

Some fiddling is forced upon one.  My ideal behaviour is to document these 
things in bug reports.  Perhaps they get fixed, but at least they are 
documented.  Embarassing fact: I sometimes rediscover my own solutions via 
google.

You might have noticed some of my "war story" postings.

| It looks to me like this bios is heavily vested in Windows gamer land
| with the bios optimized to accept hooks from the ASUS GUI overclocking
| utility.

Usually this can be ignored.  An overclocker-oriented board might have
better margins making it more stable for those who don't use the
capability.