Jari Ruusu wrote: > Matthias Schniedermeyer wrote: >> So i set the offset to 512 with "/dev/sd<x>" as the backing-store. But >> Performace fell to a crawl (relativly speaking) realative to a >> non-offseted loop. >> >> So i increased the offset by 512 for a few times in the hope that the >> phaenomenom is "curable". And with an offset of 4096 performance was the >> same as without an offset. >> >> As i have a IA32-System it appears to me that the offset has to be >> page-aligned, in case performance matters. So i suggest to put a note >> about that in the README. > > What kernel version are you using? 2.6.19, vanilla, self-compiled (But later this day it will be 2.6.20) > What loop implementation are you using? loop-aes 3.1e (I'm a loop-aes user for 3-4 years) > Mainline loop driver uses page cache for both file backed and device backed > setups. Loop-AES version of loop driver uses page cache only for file backed > setups. Your description sounds like you are using mainline loop. I case numbers matter. CORE2 Duo E6700, 2GB DDR2-800 RAM (Or about the fastest "not extreme" system currently available) "RAW" AES128-v3 throughput this system can reach is about 100MB/s (using a single thread of aespipe) - snip - time (dd if=/dev/zero bs=20480 count=52428 | aespipe -e aes128 -p3 3< <( gpg < key.gpg ) > /dev/null) 52428+0 records in 52428+0 records out 1073725440 bytes (1.1 GB) copied, 10.9219 seconds, 98.3 MB/s real 0m10.924s user 0m7.977s sys 0m0.657s - snip - (Currently my system is working, numbers are a little bit better with 0 load) HDD is a 500GB Seagate/PATA, connected to a onboard jmicron PATA-Controller(apears to be a PCIe device) and driven by the matching libata driver. The HDD delivers a linear throughput of about 70-73MB/s, which doesn't decrease much when i put a aes128-v3-loop over it, and which uses about 1/2 of the available CPU-ressources gpg < key.gpg | losetup -e aes128 -p 0 /dev/loop4 /dev/sdb The same loop, with the "not good" offsets of 512-3584, decreases the throughput to a craw of 5-20MB/s (don't have exact numbers anymore and currently my system is working, so i can't retest) With an offset of 4096 everything is good(tm) again. gpg < key.gpg | losetup -e aes128 -p 0 -o 4096 /dev/loop4 /dev/sdb Bis denn -- Real Programmers consider "what you see is what you get" to be just as bad a concept in Text Editors as it is in women. No, the Real Programmer wants a "you asked for it, you got it" text editor -- complicated, cryptic, powerful, unforgiving, dangerous. - Linux-crypto: cryptography in and on the Linux system Archive: http://mail.nl.linux.org/linux-crypto/
