00:37
FreezingCold has quit [Ping timeout: 252 seconds]
00:53
FreezingCold has joined #linux-rockchip
01:20
Astralix1 has joined #linux-rockchip
01:22
Astralix has quit [Ping timeout: 252 seconds]
02:04
cnxsoft has joined #linux-rockchip
02:47
FreezingCold has quit [Ping timeout: 264 seconds]
02:54
ferric has joined #linux-rockchip
03:01
FreezingCold has joined #linux-rockchip
03:25
hramrach has quit [Write error: Connection reset by peer]
03:41
hramrach has joined #linux-rockchip
04:26
pbacon has quit [Ping timeout: 255 seconds]
05:34
cnxsoft has quit [Ping timeout: 240 seconds]
06:18 _hipboi_ has joined #linux-rockchip
06:18
hipboi_ has quit [Ping timeout: 240 seconds]
06:21
FreezingCold has quit [Ping timeout: 252 seconds]
06:51
cnxsoft has joined #linux-rockchip
07:35
pbacon has joined #linux-rockchip
07:45
pbacon has quit [Ping timeout: 240 seconds]
08:10
pbacon has joined #linux-rockchip
08:36
cnxsoft has quit [Ping timeout: 240 seconds]
09:45
RayFlower has quit [Ping timeout: 240 seconds]
10:11
cnxsoft has joined #linux-rockchip
10:29
rz2k has joined #linux-rockchip
10:37
m1r has joined #linux-rockchip
10:47
bgal has joined #linux-rockchip
11:03
cnxsoft has quit [Ping timeout: 276 seconds]
11:17
cnxsoft has joined #linux-rockchip
11:20
nighty-_ has joined #linux-rockchip
11:51
anatoly has joined #linux-rockchip
11:54
<
anatoly >
Hello. I found several sources of rkflashtool. Like one on sf site, crewrktablets' and Galland's on github. Which one is "official"? Or maybe I missed smth?
12:08
<
ferric >
ahh crap. i bricked my imito qx2 :(
12:10
<
ferric >
anatoly: what're you flashing and on what device?
12:27
<
anatoly >
ferric: just trying to dive in to it an play with rkflashtool. I have rk3066 tablet
12:29
<
anatoly >
I'm not going to flash anything at them moment because I have nothing to flash :-) And want to get a better understanding of it firslt
12:39
cooperok has joined #linux-rockchip
12:41
rz2k has quit [Remote host closed the connection]
13:02
ganbold_ has joined #linux-rockchip
13:42
FreezingCold has joined #linux-rockchip
13:58
m1r has quit [Ping timeout: 250 seconds]
14:13
GriefNorth has joined #linux-rockchip
14:13
m1r has joined #linux-rockchip
14:36
m1r has quit [Ping timeout: 250 seconds]
14:50
m1r has joined #linux-rockchip
15:02
cnxsoft has quit [Quit: cnxsoft]
15:35
RayFlower has joined #linux-rockchip
15:41
RayFlower has quit [Read error: Connection reset by peer]
15:41
RayFlower has joined #linux-rockchip
15:45
RayFlower has quit [Read error: Connection reset by peer]
15:46
RayFlower has joined #linux-rockchip
15:53
FreezingCold has quit [Ping timeout: 252 seconds]
15:58
RayFlower has quit [Read error: Connection reset by peer]
15:58
RayFlower has joined #linux-rockchip
16:08
FreezingCold has joined #linux-rockchip
16:10
RayFlower has quit [Read error: Connection reset by peer]
16:11
RayFlower has joined #linux-rockchip
16:12
bgal has quit [Ping timeout: 255 seconds]
16:17
RayFlower has quit [Read error: Connection reset by peer]
16:17
RayFlower has joined #linux-rockchip
16:22
hramrach has quit [Remote host closed the connection]
16:22
hramrach has joined #linux-rockchip
16:32
bgal has joined #linux-rockchip
16:40
hramrach has quit [Remote host closed the connection]
16:44
hramrach has joined #linux-rockchip
16:47
<
ferric >
so who's running linux on their rk3188 device?
17:08
RayFlower has quit [Read error: Connection reset by peer]
17:08
RayFlower has joined #linux-rockchip
17:15 _massi_ has quit [Quit: Leaving]
17:43
bgal has quit [Ping timeout: 240 seconds]
17:47
RayFlower has quit [Ping timeout: 240 seconds]
18:13
RayFlower has joined #linux-rockchip
18:15
RayFlower has quit [Read error: Connection reset by peer]
18:15
RayFlower has joined #linux-rockchip
18:32
RayFlower has quit [Read error: Connection reset by peer]
18:32
RayFlower has joined #linux-rockchip
18:36
RayFlower has quit [Read error: Connection reset by peer]
18:36
RayFlower has joined #linux-rockchip
18:38
RayFlower has quit [Read error: Connection reset by peer]
18:39
RayFlower has joined #linux-rockchip
18:43
<
cosm >
ferric: I am
18:49
m1r has quit [Ping timeout: 240 seconds]
19:13
RayFlower has quit [Read error: Connection reset by peer]
19:13
RayFlower has joined #linux-rockchip
19:17
RayFlower has quit [Read error: Connection reset by peer]
19:18
RayFlower has joined #linux-rockchip
19:22
RayFlower has quit [Read error: Connection reset by peer]
19:23
RayFlower has joined #linux-rockchip
19:28
else_ is now known as else-
19:35
RayFlower has quit [Read error: Connection reset by peer]
19:35
RayFlower has joined #linux-rockchip
19:37
GriefNorth has quit [Ping timeout: 255 seconds]
20:21
RayFlower has quit [Read error: Connection reset by peer]
20:21
RayFlower has joined #linux-rockchip
20:28
FreezingCold has quit [Ping timeout: 258 seconds]
20:30
FreezingCold has joined #linux-rockchip
20:45
RayFlower has quit [Read error: Connection reset by peer]
20:46
RayFlower has joined #linux-rockchip
20:48
RayFlower has quit [Read error: Connection reset by peer]
20:49
RayFlower has joined #linux-rockchip
20:57
rz2k has joined #linux-rockchip
21:40
tonikasch has joined #linux-rockchip
21:52
pbacon has quit [Ping timeout: 250 seconds]
22:01
FreezingCold has quit [Ping timeout: 245 seconds]
22:10
bgal has joined #linux-rockchip
22:15
FreezingCold has joined #linux-rockchip
22:20
<
cosm >
has anyone tested memory bandwidth on RK3188 devices?
22:24
<
cosm >
and it seems my device (pipo X2) only has about 500 MB/s memory bandwidth
22:25
<
cosm >
but if the framebuffer is set to 32bpp 1080p
22:25
<
cosm >
only roughly half of the memory bandwidth remains available
22:26
<
cosm >
which makes sense if we assume the LCDC reads the framebuffer at roughly 30 fps
22:26
<
cosm >
but it really screws up system performance
22:29
<
ssvb >
cosm: is RK3188 using MCTL or PCTL as the dram controller?
22:29
<
phh >
cosm what you say sounds right. That's the known problem of no dedicated vram
22:30
<
cosm >
ssvb: yeah, that post sounds about right
22:30
<
phh >
(I'd said it's half that bandwidth, ie 1gbps, but well.)
22:31
<
cosm >
ssvb: I don't know what DRAM controller is used
22:32
<
ssvb >
cosm: the code should be somewhere in the kernel sources
22:33
tonikasch has quit [Quit: Bye!]
22:33
<
cosm >
ssvb: ack finds PCTL in lots of register names, so I assume that must be it
22:34
pbacon has joined #linux-rockchip
22:35
<
cosm >
phh: sorry, I'm not sure what do you mean by half the bandwidth (for the FB?)
22:36
<
ssvb >
cosm: PCTL has only a single port and does not support QoS control
22:37
<
ssvb >
cosm: MCTL has 32 host ports with programmable priorities and flexible arbitration
22:38
<
cosm >
ssvb: ok, so it would be impractical to use more than 16 BPP
22:39
<
ssvb >
reducing monitor refresh rate to 50Hz may help
22:41
<
cosm >
oops, found a silly mistake in my microbenchmark
22:42
<
cosm >
I was reporting half the bandwidth
22:44
<
cosm >
ssvb: sure, give me a minute
22:49
<
cosm >
ssvb: so I guess the results in the wiki are with the FB disabled?
22:50
<
ssvb >
cosm: both with and without FB
22:51
<
cosm >
ssvb: I've started it on a few other boards I use
22:52
<
cosm >
ssvb: I've started it on RK3188 first and then on an Exynos4412, guess which one is way ahead
22:54
<
ssvb >
cosm: it does not mean anything, the test runs more iterations if the results are poorly reproducible (caused by some other background activity in the system)
22:55
<
ssvb >
cosm: it's likely to complete faster on an undisturbed system
23:00
<
ssvb >
cosm: is this the RK3188 result with 1080p framebuffer?
23:01
<
cosm >
it's with 1080p and 16bpp
23:01
<
cosm >
ssvb: I'm running it again with 32bpp
23:12
<
cosm >
ssvb: is there no zero-copy extension for XV?
23:14
bgal has quit [Ping timeout: 264 seconds]
23:15
<
ssvb >
cosm: AFAIK there is no zero-copy support for XV, however at least one copy is still needed
23:16
<
ssvb >
cosm: because the video decoder typically wants to read data back from the previous frame, and if you place the buffer into uncached (writecombine) memory, the performance will be very bad
23:18
<
ssvb >
is the 32bpp test still running?
23:22
<
cosm >
ssvb: I'm doing another one with the framebuffer disabled
23:22
<
cosm >
ssvb: anyway, this is quite disappointing
23:23
ganbold_ has quit [Remote host closed the connection]
23:25
<
ssvb >
cosm: maybe the dram controller is just badly misconfigured?
23:25
<
ssvb >
for example, actually running at a very low clock frequency
23:33
<
cosm >
my DRAM is running at 360 MHz when the hardware could do 400 MHz, afaik everything else is correct
23:33
<
ssvb >
is it using 32-bit or 16-bit memory interface?
23:34
<
cosm >
ssvb: there are two unpopulated DRAM footprints
23:34
<
ssvb >
which board is that?
23:37
<
ssvb >
this is really surreal and ridiculous
23:38
<
cosm >
After discovering that I was keen to try my hand at reflowing BGAs :D
23:38
<
ssvb >
who in their mind would combine four fast cpu cores with such a weak memory interface?
23:39
<
ssvb >
but can it properly do 1080p video playback (in android)?
23:40
<
ssvb >
this seems to be one of the advertised features
23:40
<
cosm >
I have no idea, I've only use Android to have a poke around in a terminal
23:41
<
cosm >
I guess some marketer type decided to remove half the memory at some point
23:41
<
cosm >
it was only $40 or so
23:44
<
cosm >
I have a different one with 2 GB of RAM and hopefully a 32bit interface arriving any day now, hopefully it's faster