New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Enable USE_SSEVECT only on x86_64 with SSE #6667
Enable USE_SSEVECT only on x86_64 with SSE #6667
Conversation
Patch enables USE_SSEVECT (vector implementation based on SSE intrinsics) only on a proper hardware, i.e, x86_64 with SSE. For the rest of architectures (armv7hl, aarch64, power8, etc.) use USE_EXTVECT. Signed-off-by: David Abdurachmanov <David.Abdurachmanov@cern.ch>
Compiler on x86_64 fine (apart compiler errors in DQM, same in IB). Yet to compile on aarch64. |
A new Pull Request was created by @davidlt for CMSSW_7_4_X. Enable USE_SSEVECT only on x86_64 with SSE It involves the following packages: DataFormats/Math @cmsbuild, @nclopezo, @StoyanStoynev, @slava77 can you please review it and eventually sign? Thanks. |
This is being tested by Jenkins on x86_64 (currently running analyzer). I merged this into my previous aarch64 build and compiling. |
Adding @VinInn |
+1 |
@@ -3,7 +3,11 @@ | |||
|
|||
#if ( defined(IN_DICTBUILD) || defined(__REFLEX__) || defined(__CINT__) || defined(__MIC__)) || (__BIGGEST_ALIGNMENT__<16) | |||
#elif defined(__GNUC__) || defined(__clang__) | |||
#define USE_SSEVECT | |||
# if defined(__x86_64__) && defined(__SSE__) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why is defined(__SSE__)
alone not enough, is it an extra sanity check?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, extra sanity check.
Seems to compile on aarch64. Will retest once approved and in IB. |
I meant to wait for comparisons to show up in jenkins before signing .. technically this should have compiled to exactly the same code and there should be no diffs at all |
This pull request is fully signed and it will be integrated in one of the next CMSSW_7_4_X IBs unless changes (tests are also fine). This pull request will be automatically merged. |
Enable USE_SSEVECT only on x86_64 with SSE
Patch enables USE_SSEVECT (vector implementation based on SSE
intrinsics) only on a proper hardware, i.e, x86_64 with SSE.
For the rest of architectures (armv7hl, aarch64, power8, etc.) use
USE_EXTVECT.
Signed-off-by: David Abdurachmanov David.Abdurachmanov@cern.ch