<Company>
and it happened since I recently avoided excess MakeCurrent() calls
<Company>
but only on AMD
<Company>
adding a manual MakeCurrent(NO_CONTEXT), MakeCurrent() at every resize apparently makes the issue go away
<Company>
but I don't have an AMD, so I can't experiment
<Company>
and it doesn't happen on my Intel
<Company>
this feels like some size update of the renderbuffer only happens on MakeCurrent() that should happen earlier, but I don't know Mesa code well enough to know if that's possible
<Company>
or if this is something in GTK that we might be doing wrong
<alyssa>
I'm curious how freedreno performs at `glmark2-es2 -bbuffer:columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata`
<alyssa>
this seems to hit a pathological case with resource shadowing on panfrost (3fps reported on 1 board), wondering if we botched the heuristic
f11f13 has quit [Remote host closed the connection]
f11f13 has joined #dri-devel
<HdkR>
[buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata: FPS: 7 FrameTime: 142.857 ms
<alyssa>
HdkR: .....woof.
<HdkR>
On Snapdragon 888
<alyssa>
HdkR: now try with FD_MESA_DEBUG=inorder
<alyssa>
er wait does that work
<HdkR>
zero change
<alyssa>
Interesting.
<alyssa>
hitting some other pathological case ig
<HdkR>
Oh wait
<HdkR>
It was freaking out since the monitor was off
<HdkR>
[buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata: FPS: 119 FrameTime: 8.403 ms
<alyssa>
Ah
<HdkR>
I think I have a hard cap of 120FPS because of my kernel atm :P
<alyssa>
rip
<alyssa>
Ooh, I.. hm
<alyssa>
i'm starting to see why this case is so bad
<alyssa>
Ooooh.
<alyssa>
Okay. I see where we screwed up.
<alyssa>
And by "we" I mean "me"
<alyssa>
aaaand also why the commit cphealy bisected to is the problem
<alyssa>
rsrc->track isn't getting reset when we shadow
pushqrdx has quit [Ping timeout: 480 seconds]
kts_ has joined #dri-devel
kts_ has quit []
kts has quit [Ping timeout: 480 seconds]
fxkamd has joined #dri-devel
i-garrison has joined #dri-devel
shashank1202_ has quit [Quit: Connection closed for inactivity]
DPA has quit [Ping timeout: 480 seconds]
DPA has joined #dri-devel
mlankhorst has quit [Ping timeout: 480 seconds]
thelounge583 has joined #dri-devel
X-Scale has joined #dri-devel
thelounge583 has quit []
<robclark>
alyssa, HdkR: I'm getting 120fps.. need to resurrect my patch to make DIRTYFB not limit fps to 2x refresh rate
thelounge583 has joined #dri-devel
<robclark>
(on sc7180/a618.. but with the 2x hz limit it really doesn't matter what you're on)
thelounge5837 has joined #dri-devel
thelounge583 has quit [Ping timeout: 480 seconds]
DPA has quit [Ping timeout: 480 seconds]
thelounge5837 has quit []
thelounge5837 has joined #dri-devel
DPA has joined #dri-devel
mlankhorst has joined #dri-devel
DPA has quit [Ping timeout: 480 seconds]
DPA has joined #dri-devel
<alyssa>
robclark: ack
<alyssa>
the bad case for me was due to a bug causing large numbers of useless shadow copies to be made
<alyssa>
but even so .. that test will cause a lot of shadow copies to be made of a fairly large (~256kb) buffer. so even debugged I can imagine there's CPU boundness there.
<robclark>
oh, yeah.. and I suppose that is done on the CPU.. from wc mappings
<alyssa>
yeah...
<alyssa>
not sure if the CPU shadowing path should have a size threshold.. especially because of wc mappings
<alyssa>
otoh if it saves a flush...
thellstrom1 has joined #dri-devel
thellstrom has quit [Ping timeout: 480 seconds]
DPA- has joined #dri-devel
DPA has quit [Ping timeout: 480 seconds]
DPA has joined #dri-devel
DPA- has quit [Ping timeout: 480 seconds]
lemonzest has quit [Quit: WeeChat 3.2]
DPA- has joined #dri-devel
DPA has quit [Ping timeout: 480 seconds]
shashank1202_ has joined #dri-devel
sdutt has joined #dri-devel
DPA- has quit []
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit [Remote host closed the connection]
DPA has joined #dri-devel
gawin has quit [Quit: Konversation terminated!]
mlankhorst has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
gouchi has quit [Remote host closed the connection]
i-garrison has quit []
shashank1202_ has quit [Quit: Connection closed for inactivity]