Address high CPU usage for some (GUI) simulations #541
Conversation
jondequinor
force-pushed
the
address-high-cpu-usage-380
branch
2 times, most recently
from
7e3f020 to
6f0748b
Compare
jondequinor
force-pushed
the
address-high-cpu-usage-380
branch
4 times, most recently
from
2c51e53 to
0b598a4
Compare
jondequinor
changed the title
jondequinor
force-pushed
the
address-high-cpu-usage-380
branch
5 times, most recently
from
fc7d01c to
6aaee2e
Compare
ert_shared/tracker/utils.py
Outdated
|
|
||
| def _scale(nr_realisations, min_time=5, max_time=15, min_real=1, max_real=500): | ||
| nr_realisations = min(max_real, nr_realisations) | ||
| nr_realisations = max(min_real, nr_realisations) |
There was a problem hiding this comment.
does not quite get what these two assignments does... ? or - at least the second one always overwrite the first one - guess that should not be the case?
There was a problem hiding this comment.
@oyvindeide wanna chime in? =)
There was a problem hiding this comment.
ok, got the explanation which gave sense ... just didn't catch it initially when reading the code.
There was a problem hiding this comment.
Late to the party (and just adding this in case it improves readability), but is the equivalent of np.clip(number, min_real, max_real), or max(min_real, min(number, max_real)), the point is to have a hard upper and lower boundary so that we normalize to 0-1 later.
There was a problem hiding this comment.
yes, I understand - and it makes sense... just me reading through a bit fast and didn't really got the sense of what's going on. One could probably make this obvious by having
def clamp(nr_realisations, max_real, min_real):
clamped = min(max_real, nr_realisations)
clamped = max(min_real, clamped)
return clamped
but - guess that is smak & behag
|
I'm ready to approve, but I think this PR is quite big and it would have been nice having another pair of eyes running through as well. |
ert_gui/simulation/run_dialog.py
Outdated
| @@ -2,7 +2,7 @@ | |||
| import sys | |||
|
|
|||
| try: | |||
| from PyQt4.QtCore import Qt, QTimer, QSize | |||
| from PyQt4.QtCore import Qt, QTimer, QSize, pyqtSignal, pyqtSlot | |||
There was a problem hiding this comment.
When touching this, could we switch to ErtQt?
jondequinor
force-pushed
the
address-high-cpu-usage-380
branch
from
6aaee2e to
8c98903
Compare
Splits *UI rendering into three parts: ticker, progress and details. Each has its own interval at which it renders, which is scaled based on the number of realizations. A simulation tracker has been implemented, and it now controls the flow of information to both the GUI and CLI monitoring through events.
jondequinor
force-pushed
the
address-high-cpu-usage-380
branch
from
8c98903 to
8ed2b5e
Compare
Issue
Resolves #380
Approach
Splits *UI rendering into three parts: ticker, progress and details.
Each has its own interval at which it renders, which is scaled based on
the number of realizations.
A simulation tracker has been implemented, and it now controls the flow
of information to both the GUI and CLI monitoring through events.