Paperweight has a 4bit framebuffer

[twobob]

Originally Posted by geekmaster

Once you find a suitable acid rock, then you need to figure out where to apply it on the DX, and how high to drop it from.

Tee hee hee. You know some idiot paid top dollar for that little cutie dinosaur... (okay that idiot was me)

Don't say bad things about her, she may brick

[knc1]

Haven't crashed it yet:

top - 11:50:54 up 12 days, 11:44,  1 user,  load average: 0.15, 0.11, 0.13
Tasks:  29 total,   1 running,  28 sleeping,   0 stopped,   0 zombie
Cpu0  : 17.0%us,  5.2%sy,  1.9%ni, 72.2%id,  2.5%wa,  0.0%hi,  1.1%si,  0.0%st
Cpu1  : 17.4%us,  7.0%sy,  1.6%ni, 72.4%id,  1.6%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu2  : 14.6%us,  4.3%sy,  1.6%ni, 78.0%id,  1.4%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu3  : 13.8%us,  3.9%sy,  1.5%ni, 79.5%id,  1.2%wa,  0.0%hi,  0.1%si,  0.0%st
Cpu4  : 44.0%us,  6.7%sy,  4.7%ni, 43.2%id,  1.4%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu5  : 34.5%us,  5.1%sy,  3.6%ni, 55.2%id,  1.5%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu6  : 30.1%us,  8.0%sy,  3.0%ni, 57.5%id,  1.2%wa,  0.0%hi,  0.2%si,  0.0%st
Cpu7  : 57.3%us,  6.0%sy,  6.7%ni, 23.6%id,  1.2%wa,  0.0%hi,  5.2%si,  0.0%st
Mem:    921600k total,   252752k used,   668848k free,        0k buffers
Swap:        0k total,        0k used,        0k free,   147324k cached

With TW, the routine is to keep enabling things until the server crashes -
Then from that set, remove a few features.

[twobob]

If I move slow the top nav menu disappears before I to the sub - as an aside

However if I get someone young to help move it fast we can make it

Looking very shiny now

[knc1]

Originally Posted by twobob

If I move slow the top nav menu disappears before I to the sub - as an aside

However if I get someone young to help move it fast we can make it

Looking very shiny now

Still working on it.

I have seen the same thing myself about that top menu bar.
Not yet sure what it may be.

I think it is the browser failing to load all of the java scripts.
The server has a connection limit set, I think at 5 connections per client.

Some browsers, in a search for loading speed, open 6 or more connections per view. Which means one or more of something doesn't get loaded.

I will check the logs when I get a chance, but until then, try a different browser. Or, just try multiple "refresh" actions on the same page - until the browser caches all of the java script.

(A browser other than IE-6, this TW version dropped support for IE-6)

[twobob]

yeah. its a non issue. just reporting, like a good little tester

[miquele]

hi,
don't get what the practical implications of a lower framebuffer are. is that of relevance to a non-hacker? tried to look it up, but unsuccessfully.
thanks.

[knc1]

It is a play on (technical) words.

[geekmaster]

All of the eink kindles (so far) use a 4-bit framebuffer at the hardware level. That means that you can only display pixels in 16 different shades of gray. To display more than 16 colors, you need to use dithering, which mixes those 16-colors spatially to give up to 256 different AVERAGE brightness levels.

In the case of kindles with 8-bit framebuffers, the bottom and top 4-bits must be identical, or unpredictable (hardware dependent) side-effects will result (according to comments in the eink driver GPL source code).

But for smooth animation, you can really only use 2-bit (black and white) and you need to dither only those two colors to get full grayscale. There is some sample C code showing how to do this dithering, in the "newtrix" demo (both 2-bit and 4-bit dithering).

[eureka]

There is the only [techincal] thread about Paperwhite in Developer's Corner at the moment, so I leave this information here: /usr/bin/powerd in Kindle Touch firmware already knows about magnetic sensor. To see it:

strings /usr/bin/powerd | grep -i mag

So there is another trace of Paperwhite hardware support (sensor for going to sleep on cover closing) in current Kindle Touch software.

[twobob]

hmm... or perhaps the hint of a magic button?

There was some talk of using a device internally on the DX (and perhaps the KpW) to enable the mode shift.

This stuff could be easily re-purposed for such a task, as an idle thought.

[twobob]

on the 3.4:

[root@kindle root]# strings /usr/bin/powerd | grep -i mag

system/daemon/powerd/DRIVER_MAGSENSOR
W def:magnostate::Invalid mag sensor event - no state information
D def:mag senror event. state=%s (%s:%s:%d)
I def:magopen::Mag sensor open event
I def:magclose::Mag sensor closed event
W def:msgstate:state=%s:Invalid mag sensor event - unknown state
e_mag_sensor_open
e_mag_sensor_closed

[eureka]

Originally Posted by twobob

on the 3.4:

[root@kindle root]# strings /usr/bin/powerd | grep -i mag

Thanks for checking.

Yep, the same I've seen on KT. So it means I was wrong when I thought it's implemented specifically for Paperwhite. Looks like it's an artifact of support of earlier Kindle model. Sorry for false alarm.

[NiLuJe]

@eureka: Or that they're using the same codebase for the low-level stuff .

[twobob]

Originally Posted by NiLuJe

@eureka: Or that they're using the same codebase for the low-level stuff .

^ My thoughts.

[twobob]

Originally Posted by eureka

Thanks for checking.

Yep, the same I've seen on KT. So it means I was wrong when I thought it's implemented specifically for Paperwhite. Looks like it's an artifact of support of earlier Kindle model. Sorry for false alarm.

Heck, no apology needed that's great info!!

next up. we need someone to write some code to monitor that and wave magnets around a DX / Kpw might be fun : D

meh okay it might not be fun, would be nice to know though : )