build scripts: Made minor modifications to reduce build overhead time.
Review Request #1081 - Created March 5, 2012 and submitted
| Information | |
|---|---|
| Marc Orr | |
| gem5 | |
| default | |
| Reviewers | |
| Default | |
Changeset 8881:042d509574c1 --------------------------- build scripts: Made minor modifications to reduce build overhead time. 1. --implicit-cache behavior is default. 2. makeEnv in src/SConscript is conditionally called. 3. decider set to MD5-timestamp 4. NO_HTML build option changed to SLICC_HTML (defaults to False)
Ran ./util/regress. Seemed to behave as expected.
Posted (March 5, 2012, 11:34 a.m.)
-
src/SConscript (Diff revision 1) -
I don't want to claim that this won't work, but what happens if somebody has a build target that isn't one of the main binaries? Occasionally (very) I've built one of the targets of the ISA parser explicitly to focus on just that bit. What would happen then?
Posted (March 5, 2012, 12:06 p.m.)
Hi Marc. Looks great, but I too have a concern like Gabe mentioned. Not so much in his usecase, but what happens if you try to build one of the unittest binaries or a regression test target (that implicitly depends on a binary)? Otherwise looks great. Thanks for all the effort. Ali
Posted (March 5, 2012, 1:15 p.m.)
Perhaps a better solution is to iterate through BUILD_TARGETS and try to identify for each one which environment it needs (e.g., looking for a suffix like '.debug' or '.do', or 'tests/debug' in the path), and if these heuristics fail then just fall back on including them all.
-
src/SConscript (Diff revision 1) -
Similarly to Gabe's point, I sometimes use scons to drive individual tests, like: scons build/ALPHA/tests/debug/quick/fs/10.linux-boot/alpha/linux/tsunami-simple-timing so the -3 index isn't a given here either.
Review request changed
Updated (March 6, 2012, 3:44 a.m.)
Summary: |
|
|||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Description: |
|
|||||||||||||||||||||||||||||||||
Diff: |
Revision 2 (+47 -16) |
looks good to me.
Other than the extra blank line and making sure you get the right commit message (somehow you lost it in the update), this looks great to me. Thanks!
-
src/SConscript (Diff revision 2) -
Looks like you have a spurious blank line here.
Ship It!
Review request changed
Updated (March 7, 2012, 4:02 a.m.)
Summary: |
|
||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Description: |
|
||||||||||||||||||||||||||||||
Diff: |
Revision 3 (+46 -16) |
