]> dgit.raspbian.org Git - git-annex.git/commitdiff
Added a comment
authorAtemu <Atemu@web>
Wed, 9 Mar 2022 11:40:28 +0000 (11:40 +0000)
committeradmin <admin@branchable.com>
Wed, 9 Mar 2022 11:40:28 +0000 (11:40 +0000)
doc/bugs/git-annex_is_slow_at_reading_file_content/comment_8_559e5c3b5ea9c166b3aa4a2e975643c6._comment [new file with mode: 0644]

diff --git a/doc/bugs/git-annex_is_slow_at_reading_file_content/comment_8_559e5c3b5ea9c166b3aa4a2e975643c6._comment b/doc/bugs/git-annex_is_slow_at_reading_file_content/comment_8_559e5c3b5ea9c166b3aa4a2e975643c6._comment
new file mode 100644 (file)
index 0000000..dbe6a0e
--- /dev/null
@@ -0,0 +1,12 @@
+[[!comment format=mdwn
+ username="Atemu"
+ avatar="http://cdn.libravatar.org/avatar/d1f0f4275931c552403f4c6707bead7a"
+ subject="comment 8"
+ date="2022-03-09T11:40:28Z"
+ content="""
+SHA256 is by far the fastest hash on my Celeron J4105 with acceleration. Seconds is BLAKE at around half the speed of accelerated SHA256 or over twice the speed of unaccelerated SHA256.
+
+Seems like coreutils simply uses openssl when available and openssl handles HW acceleration. I just tried using two Haskell OpenSSL wrappers (hopenssl and HsOpenSSL) in my minimal example and it was actually *faster* than the openssl CLI utility by almost 100MiB/ (though still a bit slower than `sha256sum`).
+
+Would using those be an option?
+"""]]