Add mkFit as an external #5150
Add mkFit as an external #5150
Conversation
|
@cmsbuild, please test (am I powerful enough also here?) |
|
The tests are being triggered in jenkins. |
|
Comparison job queued. |
|
Comparison is ready Comparison Summary:
|
|
No. That gets defined elsewhere (eg gcc.spec)
On 14 Aug 2019, at 00:22, Slava Krutelyov <[email protected]<mailto:[email protected]>> wrote:
@slava77 commented on this pull request.
________________________________
In mkfit.spec<#5150 (comment)>:
@@ -0,0 +1,20 @@
+### RPM external mkfit 2.0.0
+%define tag V2.0.0-1+pr237
+%define branch devel
+%define github_user trackreco
+
+Source: git+https://github.com/%{github_user}/%{n}.git?obj=%{branch}/%{tag}&export=%{n}-%{realversion}&output=/%{n}-%{realversion}.tgz
+Requires: tbb
+
+%prep
+%setup -q -n %{n}-%{realversion}
+
+%build
+TBB_PREFIX=$TBB_ROOT make
can we make the target compiler and arch here explicit?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub<#5150?email_source=notifications&email_token=ABGPFQY34BUFZOFO7GZDNELQEMXXPA5CNFSM4ILOZSHKYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCBO55OA#pullrequestreview-274587320>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ABGPFQZD7TUFZNPNL6OUH2LQEMXXPANCNFSM4ILOZSHA>.
|
please point me to the line where it's done. |
|
Maybe you can be more specific. It doesn’t sound like you really mean compiler and arch but rather something else?
On 14 Aug 2019, at 08:55, Slava Krutelyov <[email protected]<mailto:[email protected]>> wrote:
No. That gets defined elsewhere (eg gcc.spec)
please point me to the line where it's done.
I recall that in some other externals we had to set some flags manually (e.g. in TF and mxnet).
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub<#5150?email_source=notifications&email_token=ABGPFQY3YCZWBUJVAE5ESFLQEOTZXA5CNFSM4ILOZSHKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4H32XQ#issuecomment-521125214>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ABGPFQ4AKHVUQUB2AWMQAO3QEOTZXANCNFSM4ILOZSHA>.
|
|
I meant compiler (gcc) and a specific target architecture (core2 or sse3 or avx etc) |
|
the compiler is defined by other spec files.
options are typically figured out by the configure/cmake scripts of external packages (and overridden via command line options passed to those scripts in the spec file when the default behavior is not grid computing compatible)
… On Aug 14, 2019, at 1:54 PM, Slava Krutelyov ***@***.***> wrote:
I meant compiler (gcc) and a specific target architecture (core2 or sse3 or avx etc)
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
|
|
how does the other spec file know about the fanciness of the mkFit/Makefile.config ? |
|
no other spec file should need to know about some internal makefile of mkFit. That said, Makefile.config looks pretty unfriendly for maintaining a build recipe from. is there a cmake or configure; make solution (of sorts) under development?
… On Aug 14, 2019, at 2:00 PM, Slava Krutelyov ***@***.***> wrote:
how does the other spec file know about the fanciness of the mkFit/Makefile.config ?
About the arch, it looks like -mavx is the default.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
|
I do not quite understand what you mean. I'm so far reading your comments that all arguments specifying which make rule/setting to check is defined in the gcc.spec file and at the same time in no other file except mkfit.spec. |
|
maybe its worth having a look at some other spec files to see the typical interface (eg, root.spec) You'll see the way things like passing in the location of dependencies (gcc, TBB_ROOT) are typically parsed by software packages.
… On Aug 14, 2019, at 3:19 PM, Slava Krutelyov ***@***.***> wrote:
no other spec file should need to know about some internal makefile of mkFit
I do not quite understand what you mean.
My comment was about a line in mkfit.spec "make" call to add more than just TBB_PREFIX=$TBB_ROOT to the build command.
I'm so far reading your comments that all arguments specifying which make rule/setting to check is defined in the gcc.spec file and at the same time in no other file except mkfit.spec.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
|
Just to be clear, I'm making comments in the context that this is an external and this PR can not modify the files in the mkFit repository. |
I'm working on that (thanks for pointing it out @slava77), update should come shortly. |
cmsbuild
added
comparison-pending
tests-pending
and removed
comparison-available
tests-approved
labels
|
Pull request #5150 was updated. |
|
@cmsbuild, please test Added |
|
The tests are being triggered in jenkins. |
|
Comparison job queued. |
|
Comparison is ready Comparison Summary:
|
|
@makortel @smuzaffar @mrodozov |
|
@smuzaffar @mrodozov |
ping |
|
+externals |
cmsbuild
added
externals-approved
fully-signed
and removed
externals-pending
pending-signatures
labels
|
This pull request is fully signed and it will be integrated in one of the next IB/CMSSW_11_0_X/gcc700 IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @davidlange6, @slava77, @smuzaffar, @fabiocos (and backports should be raised in the release meeting by the corresponding L2) |
This PR adds the mkFit as an external package (the CMSSW side will follow).
Pardon my ignorance, but will the contents of a PR to
gcc700propagate automatically to other compiler branches?@slava77