| commit | 5f892cacbdf50322bc3ee2e131c105c71b495086 | |
| author | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | |
| Tue, 15 Jul 2014 16:19:09 +0000 (15 12:19 -0400) | ||
| committer | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | |
| Tue, 30 Sep 2014 12:50:31 +0000 (30 07:50 -0500) | ||
| tree | d7acc81fd96b55a4d0e5a458b6e3f2f8712ecf4d | treesnapshot (tar.gz zip) |
| parent | e6bb56b6914e6435e251814a3a0ccd7fb65a7e36 | commitdiff |
PowerPC: memset optimization for POWER8/PPC64
This patch adds an optimized memset implementation for POWER8. For
sizes from 0 to 255 bytes, a word/doubleword algorithm similar to
POWER7 optimized one is used.
For size higher than 255 two strategies are used:
1. If the constant is different than 0, the memory is written with
altivec vector instruction;
2. If constant is 0, dbcz instructions are used. The loop is unrolled
to clear 512 byte at time.
Using vector instructions increases throughput considerable, with a
double performance for sizes larger than 1024. The dcbz loops unrolls
also shows performance improvement, by doubling throughput for sizes
larger than 8192 bytes.
This patch adds an optimized memset implementation for POWER8. For
sizes from 0 to 255 bytes, a word/doubleword algorithm similar to
POWER7 optimized one is used.
For size higher than 255 two strategies are used:
1. If the constant is different than 0, the memory is written with
altivec vector instruction;
2. If constant is 0, dbcz instructions are used. The loop is unrolled
to clear 512 byte at time.
Using vector instructions increases throughput considerable, with a
double performance for sizes larger than 1024. The dcbz loops unrolls
also shows performance improvement, by doubling throughput for sizes
larger than 8192 bytes.