icecream95 has joined #panfrost
raster has quit [Quit: Gettin' stinky!]
icecream95 has quit [Ping timeout: 260 seconds]
icecream95 has joined #panfrost
yann has quit [Ping timeout: 272 seconds]
stikonas has quit [Remote host closed the connection]
robmur01 has quit [Ping timeout: 260 seconds]
robmur01 has joined #panfrost
robmur01 has quit [Ping timeout: 260 seconds]
vstehle has quit [Ping timeout: 260 seconds]
davidlt has joined #panfrost
tomboy64 has quit [Ping timeout: 240 seconds]
tomboy64 has joined #panfrost
buzzmarshall has quit [Remote host closed the connection]
vstehle has joined #panfrost
mixfix41 has joined #panfrost
_mixfix41_ has joined #panfrost
kinkinkijkin has quit [Read error: Connection reset by peer]
kinkinkijkin has joined #panfrost
Green has quit [Quit: ...]
Green has joined #panfrost
raster has joined #panfrost
robmur01 has joined #panfrost
robmur01_ has joined #panfrost
robmur01 has quit [Ping timeout: 246 seconds]
robmur01_ is now known as robmur01
robmur01 has quit [Client Quit]
robmur01 has joined #panfrost
<
tomeu>
alyssa: have been looking at depth renders for -brefract, and after making the cmdstream almost identical, I still don't get anything rendered
<
tomeu>
so wonder if there isn't something to be fixed in the BLEND instruction, or something else at fragment out
icecream95 has quit [Quit: leaving]
Ke has quit [Quit: killed]
la-s has quit [Quit: killed]
nhp[m] has quit [Quit: killed]
thefloweringash has quit [Quit: killed]
mani_s has joined #panfrost
showliu has joined #panfrost
thefloweringash has joined #panfrost
andrey-konovalov has joined #panfrost
la-s has joined #panfrost
Ke has joined #panfrost
nhp[m] has joined #panfrost
stikonas has joined #panfrost
chewitt has quit [Read error: Connection reset by peer]
austriancoder has joined #panfrost
chewitt has joined #panfrost
nlhowell has quit [Read error: Connection reset by peer]
yann has joined #panfrost
Werner has joined #panfrost
Werner has quit [Read error: Connection reset by peer]
Werner has joined #panfrost
_mixfix41_ has left #panfrost [#panfrost]
mixfix4one has joined #panfrost
buzzmarshall has joined #panfrost
<
alyssa>
tomeu: so, -bshadow works with one of my hack branches and that's a depth attachment
<
alyssa>
-brefract is special in having a depth
*and* colour attachment so.. slightly diffeernet issue
gcl has quit [Ping timeout: 260 seconds]
gcl has joined #panfrost
<
tomeu>
alyssa: do you know which branch? -bshadow didn't work with your shadow branch here
gcl has quit [Ping timeout: 240 seconds]
<
alyssa>
tomeu: should have been my shadow branch
<
alyssa>
clear_color_* differ I notice
gcl has joined #panfrost
<
alyssa>
.bifrost1.unk1 = 0xa0,
<
alyssa>
^ I'm pretty sure this is wrong
<
alyssa>
It's earily calm
<
alyssa>
Granted we don't know what all those other bits mean
<
alyssa>
and the blob is doing 0x2a0 which is equivalent for this shader
<
tomeu>
0x2a0 didn't work either
<
alyssa>
it wouldn't
<
alyssa>
why not what we have in master?
<
tomeu>
it didn't work, all those changes were done one by one trying to find what made a difference
<
tomeu>
only the ones in mfbd.format did make a difference, btw
<
tomeu>
iirc, they are quiet a few :)
<
alyssa>
(it's just that the blob is doing some crazy optimizations on these shaders that we don't match yet so.)
<
tomeu>
alyssa: I get this:
<
tomeu>
[shadow] <default>:glmark2-es2-drm: ../mesa/src/panfrost/bifrost/bi_ra.c:140: bi_adjust_src_ra: Assertion `off == offset' failed.
<
alyssa>
I swear I fixed that on one of my branches here
<
alyssa>
280b65126e8159e62517828eda12ed3789078aff iirc
<
alyssa>
Oh, here's something
<
alyssa>
rgba16f varying_0.rrrr;
<
alyssa>
varying_0<16>[64] = (memory_22d9000 + 3840);
<
alyssa>
but in the fs
<
alyssa>
rgba32f varying_0.rrrr;
<
alyssa>
varying_0<16>[64] = (memory_22d9000 + 3840);
<
alyssa>
^ defn a bug
<
alyssa>
^ Don't actually use this but this should maybe fix
<
alyssa>
(I'll need to revisit linkage regardless for a proper fix)
<
tomeu>
alyssa: your shadow branch plus that hack still hits the same assertion
<
alyssa>
tomeu: I know, but it might help deqp
davidlt has quit [Remote host closed the connection]
rhyskidd has joined #panfrost
davidlt has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott has quit [Ping timeout: 260 seconds]
cwabbott_ is now known as cwabbott
gcl_ has joined #panfrost
gcl has quit [Ping timeout: 246 seconds]
kinkinkijkin has quit [Remote host closed the connection]
rcf has quit [Ping timeout: 265 seconds]
rcf has joined #panfrost
raster has quit [Quit: Gettin' stinky!]
raster has joined #panfrost
gcl_ has quit [Ping timeout: 260 seconds]
gcl has joined #panfrost
davidlt has quit [Ping timeout: 272 seconds]
raster has quit [Quit: Gettin' stinky!]
<
alyssa>
I just figured out the diff between ld_char2/ld_short on midgard
<
alyssa>
zero-extension vs sign-extension :D
<
alyssa>
so the `abs` bit for scalar
*does* let you signal zext
icecream95 has joined #panfrost
<
alyssa>
I must say, after swimming in bifrost, reading midgard assembly is so.. simple
<
alyssa>
Clauses are hard to read \shrug
<
alyssa>
like I love Bifrost
<
alyssa>
and clauses are
*adorable*
<
alyssa>
but midgard asm is shorter
warpme_ has quit [Quit: Connection closed for inactivity]
megi has quit [Quit: WeeChat 2.8]
<
Lyude>
alyssa: i am glad I am not the only one who describes bifrost as adorable lol
megi has joined #panfrost
<
alyssa>
5 files changed, 7 insertions(+), 87 deletions(-)
<
alyssa>
midg compiler down to 10.7k lines, not counting the disasm