package_rpm: use zstd's default compression level
PATCH 1/2 package_rpm: use zstd's default compression level
From: Alexander Kanavin <[email protected]>
zstd uses 3 by default, while 19 is the highest and slowest. It's not clear why 19 was picked to begin with, possibly I copy-pasted it from rpm's examples without thinking:
https://git.yoctoproject.org/poky/commit/?h=master-next&id=4a4d5f78a6962dda5f63e9891825c80a8a87bf66
This brings significant speedups in rpm's compression step:
for example compressing webkitgtk takes 11s instead of 36s.
The rpm size increases from 175648k to 234860k. I think it's a worthy default tradeoff.
--- meta/classes-global/package_rpm.bbclass | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
meta/classes-global/package_rpm.bbclass +1 -1
@@ -10,7 +10,7 @@ IMAGE_PKGTYPE ?= "rpm" RPM = "rpm" RPMBUILD = "rpmbuild" -RPMBUILD_COMPMODE ?= "${@'w19T%d.zstdio' % int(d.getVar('ZSTD_THREADS'))}" +RPMBUILD_COMPMODE ?= "${@'w3T%d.zstdio' % int(d.getVar('ZSTD_THREADS'))}" PKGWRITEDIRRPM = "${WORKDIR}/deploy-rpms"
Signature · 1 line
2.39.5
PATCH 2/2 package_rpm: restrict rpm to 4 threads
From: Alexander Kanavin <[email protected]>
TL;DR version:
with this, and the previous compression level changes
I am seeing drastic speedups in package_write_rpm completion times:
webkitgtk goes from 78 seconds to 37 seconds glibc-locale goes from 399 seconds to 58 seconds (!)
The long version:
rpm uses multithreading for two purposes:
- spawning compressors (which are nowadays themselves multi-threaded, so the feature is not as useful as it once was)
- parallel file classification
While the former behaves well on massively parallel CPUs (it was written and verified here :), the latter was then added by upstream and only benchmarked on their very old, slow laptop, apparently:
https://github.com/rpm-software-management/rpm/commit/41f0e214f2266f02d6185ba11f797716de8125d4
On anything more capable it starts showing pathologic behavior, presumably from spawning massive amount of very short-lived threads, and then having to synchronize them. For example classifying glibc-locale takes
5m20s with 256 threads (default on my machine!)
1m49s with 64 threads
59s with 16 threads
48s with 8 threads
Even a more typical recipe like webkitgtk is affected:
47s with 256 threads
32s with 64 threads
27s with 16 or 8 threads
I have found that the optimal amount is actually four: this also means that only four compressors are running at a time, but as they're themselves using threads, and typical recipes are dominated by just two or three large packages, this does not affect overall completion time.
--- meta/classes-global/package_rpm.bbclass | 1 + 1 file changed, 1 insertion(+)
meta/classes-global/package_rpm.bbclass +1 -0
@@ -688,6 +688,7 @@ python do_package_rpm () { cmd = cmd + " --define '_use_internal_dependency_generator 0'" cmd = cmd + " --define '_binaries_in_noarch_packages_terminate_build 0'" cmd = cmd + " --define '_build_id_links none'" + cmd = cmd + " --define '_smp_ncpus_max 4'" cmd = cmd + " --define '_source_payload %s'" % rpmbuild_compmode cmd = cmd + " --define '_binary_payload %s'" % rpmbuild_compmode cmd = cmd + " --define 'clamp_mtime_to_source_date_epoch 1'"
Signature · 1 line
2.39.5