[llvm-bugs] [Bug 43504] New: Slow matrix vector product

via llvm-bugs llvm-bugs at lists.llvm.org
Mon Sep 30 05:22:49 PDT 2019


https://bugs.llvm.org/show_bug.cgi?id=43504

            Bug ID: 43504
           Summary: Slow matrix vector product
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: david.bolvansky at gmail.com
                CC: craig.topper at gmail.com, llvm-bugs at lists.llvm.org,
                    llvm-dev at redking.me.uk, spatel+llvm at rotateright.com

https://gitlab.com/chriscox/CppPerformanceBenchmarks/blob/master/matrix_vector_product.cpp

clang++-10 -Ofast matrix_vector_product.cpp -o mm
time ./mm 500

Total absolute time for int32_t matrix vector product: 2.63 sec (2.28 sec with
-march=haswell)
Total absolute time for float matrix vector product: 2.43 sec (2.61 sec with
-march=haswell)
Total absolute time for double matrix vector product: 1.77 sec (1.55 sec with
-march=haswell)


g++-10
Total absolute time for int32_t matrix vector product: 2.40 sec (1.68 sec with
-march=haswell)
Total absolute time for float matrix vector product: 1.75 sec (2.05 sec with
-march=haswell)
Total absolute time for double matrix vector product: 1.28 sec (1.49 sec with
-march=haswell)

icc-19

Total absolute time for int32_t matrix vector product: 2.49 sec (1.86 sec with
-march=haswell)

Total absolute time for float matrix vector product: 1.96 sec (1.86 sec with
-march=haswell)

Total absolute time for double matrix vector product: 1.29 sec (1.42 sec with
-march=haswell)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20190930/e1767455/attachment.html>


More information about the llvm-bugs mailing list