diff options
| author | Paolo Bonzini <pbonzini@redhat.com> | 2017-04-18 12:41:18 +0200 |
|---|---|---|
| committer | Radim Krčmář <rkrcmar@redhat.com> | 2017-05-16 21:15:50 +0200 |
| commit | b401ee0b85a53e89739ff68a5b1a0667d664afc9 (patch) | |
| tree | 68a3164a492eb30017c61a727a2c66812d9ed576 /arch/arm/include/asm/kvm_coproc.h | |
| parent | 0780516a18f87e881e42ed815f189279b0a1743c (diff) | |
| download | linux-b401ee0b85a53e89739ff68a5b1a0667d664afc9.tar.gz linux-b401ee0b85a53e89739ff68a5b1a0667d664afc9.tar.bz2 linux-b401ee0b85a53e89739ff68a5b1a0667d664afc9.zip | |
KVM: x86: lower default for halt_poll_ns
In some fio benchmarks, halt_poll_ns=400000 caused CPU utilization to
increase heavily even in cases where the performance improvement was
small. In particular, bandwidth divided by CPU usage was as much as
60% lower.
To some extent this is the expected effect of the patch, and the
additional CPU utilization is only visible when running the
benchmarks. However, halving the threshold also halves the extra
CPU utilization (from +30-130% to +20-70%) and has no negative
effect on performance.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
Diffstat (limited to 'arch/arm/include/asm/kvm_coproc.h')
0 files changed, 0 insertions, 0 deletions