AMD Ryzen Z1 Extreme Handheld - Lighting / White Flash while Gaming #2161
Labels
No labels
accessibility
autorelease: pending
autorelease: tagged
bug
dependencies
documentation
duplicate
enhancement
github_actions
good first issue
help wanted
invalid
question
testing wanted
wontfix
No milestone
No assignees
1 participant
Due date
No due date set.
Dependencies
No dependencies set
Reference
MagicBeanDev/Hookshot#2161
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Hookshot version
v2.0.14
What kind of problem
Graphics (missing, black or wrong-coloured things, lighting, flicker)
What happened
I am helping my friend to report. He said that v2.0.0 would be okay for him with no lighting, but the newer version all got flash / lighting. He would hope the later versions can fix this problem. Thanks
v2.0.14 -- lighting and flash
v2.1.0 -- crash and cannot enter the game
Where in the game
All the time
How often
Every time
Graphics card and driver version
AMD Radeon 780M; Driver vesion: 32.031007.6082
CPU, RAM and operating system
CPU z1e, 24GB RAM, 8GB VRAM, Win11 26h2
Changes from the defaults
Nil
Other software hooked into the game
No, nothing hooked in
hookshot_log.txt
hookshot_log.zip
The log.txt is for v2.0.14
Screenshots or video
Flash video.zip
Before you send it
Same chip and same GPU here (Ryzen Z1 Extreme, Radeon 780M, 24 GB with 8 GB
carved out for the GPU), but on Linux rather than Windows 11 — so this is not a
reproduction of your friend's setup, only a data point from the same hardware.
Two things that may be worth separating, because they look like different
problems.
"v2.1.0 — crash and cannot enter the game"
v2.1.0 does start and play here on this hardware, so whatever happens on his PC
is not every Z1 Extreme. One thing that can look exactly like a hang: the first
start-ups after an update are very long, because the game compiles shaders as it
boots and fills its program cache while it does. Measured here, cold after
updating to v2.1.0:
It took three launches before the title screen appeared in reasonable time, and
hookshot_programs.bingrew to 6.7 GB on the way. If he leaves it running, thelog says how far along it is and eventually estimates the time left:
So the useful question is whether his
hookshot_log.txtstops at a crash orstops while those lines are still advancing — those are very different bugs. The
hookshot_log.txtfield in the report is empty; attaching it (andhookshot_version.txt) would settle it. Worth checking too that the drive hasroom: the caches here are 6.7 GB of programs plus 297 MB of pipelines, on top of
the install.
"v2.0.14 — lighting and flash"
That one sounds like the flicker family tracked in #1005, which notes that the
frame wait is on by default on Windows since v2.0.0 — his "v2.0.0 was okay,
newer versions flash" fits that timeline. #1005 lists two settings that are
quick to try, and either answer is informative:
RX_VK_WAIT_GPU=0(removes thevolumetric-fog brightness sawtooth of #570) and
RX_VK_PER_QUEUE_FENCES=0(fixed it for the reporters of #589). On Windows, set them in the console the
game is started from:
I can measure flashes on this hardware with the renderer's own flash detector
(
RX_VK_FLASH=1withRX_VK_SHOTS=<dir>andRX_VK_SHOT_EVERY=1), whichcompares each presented frame against both its neighbours on a 64x36 grid and
writes the three frames of a one-frame flash, with how many cells jumped and
where —
RX_VK_FLASH_DIFF(48) andRX_VK_FLASH_CELLS(40) set how subtle aflash still counts. If a specific place in
the game is worse than the rest for him, say where and I will run it there and
post the frame counts.
I ran the flash detector in Kokiri Forest on this hardware (Z1 Extreme / Radeon
780M, Linux), since that is where the reports usually put it: 6 one-frame
flashes in 60 s, in two bursts of three consecutive presents, alternating
bright and dark, +6.6 brightness against the neighbouring frames in the bottom
half of the picture (the lit grass and bushes) against +0.7 in the top half.
The numbers and the method are in #1005:
#1005
So the pulse your friend describes does exist on this chip and GPU, on v2.1.0,
with the settings at their defaults — but at a rate that is easy to miss unless
you are looking for it. Whether his is the same thing at a worse rate, or
something else, is what
hookshot_log.txtwould say.Next on this machine: the same minute with
RX_VK_WAIT_GPU=0, then withRX_VK_PER_QUEUE_FENCES=0, to see whether the count goes to zero. I will postthe result here and in #1005 either way.
Hi, my friend just told me.. he uninstalled your program, and reinstall again with all the stuff v2.1.0.... nearly everything is solved!!!
The game would not crash!
But will have lighting and lag sometimes....
and 24GB ram will use 22GB....
Thanks for the report. Reports of the same problem are now collected in #997, which lists this one with your details and links back here, so one fix closes them together and the open list stays readable. Follow #997 for the fix, and add new details there (or here). If this turns out to be a separate problem it will be reopened.