Active Topics
-
Firefox with Leste (5)
to Maemo 7 / Leste by teroyk - 4 days ago -
Handbook (2)
to Maemo 7 / Leste by teroyk - 4 days, 23 hrs ago - more...
|
2012-05-27
, 10:28
|
|
Posts: 3,074 |
Thanked: 12,960 times |
Joined on Mar 2010
@ Sofia,Bulgaria
|
#212
|
Originally Posted by vi_
I tend to support that, I am running /proc/sys/vm/page-cluster = 0 for ages and it definitely makes a difference.
Now here is a VERY interesting paper that advocates the disabling of pre-fetching for massive improvements in VM on flash.
On the other hand it seems there should be some system tweaking for compcache to be effective. Unfortunately I have NFC where and what
to tweak besides what i've already posted.
__________________
Never fear. I is here.
720p video support on N900,SmartReflex on N900,Keyboard and mouse support on N900
Nothing is impossible - Stable thumb2 on n900
Community SSU developer
kernel-power developer and maintainer
Never fear. I is here.
720p video support on N900,SmartReflex on N900,Keyboard and mouse support on N900
Nothing is impossible - Stable thumb2 on n900
Community SSU developer
kernel-power developer and maintainer
| The Following 3 Users Say Thank You to freemangordon For This Useful Post: | ||
|
|
2012-05-27
, 10:32
|
|
Posts: 1,680 |
Thanked: 3,685 times |
Joined on Jan 2011
|
#213
|
Originally Posted by impeham
Uh, why would you set the prioroty of ramzswap LOWER than emmc swap? That completely defeats the point of ramzswap. Now VM will use emmc till it is full THEN use ramzswap.
ok - issue solved - the following made it work:
swapon /dev/mmcblk0p3
insmod /lib/modules/current/ramzswap.ko disksize_kb=65536
swapon /dev/ramzswap0 -p 10
swapoff /dev/mmcblk0p3
it seems that enabling the internal memory's swap was needed to avoid the crash - i don't know why.
Now i have 2 swaps:
/dev/mmcblk1p4 (P -1)
/dev/ramzswap0 (P - 10)
Thanks for the help.
__________________
N900: One of God's own prototypes. A high-powered mutant of some kind never even considered for mass production. Too weird to live, and too rare to die.
|
|
2012-05-27
, 10:43
|
|
Posts: 669 |
Thanked: 433 times |
Joined on May 2010
|
#214
|
Originally Posted by vi_
Why lower? ramzswap priority 10, card swap priority 1 - the "-" char was confusing
Uh, why would you set the prioroty of ramzswap LOWER than emmc swap? That completely defeats the point of ramzswap. Now VM will use emmc till it is full THEN use ramzswap.
.
|
|
2012-05-27
, 11:22
|
|
Posts: 1,680 |
Thanked: 3,685 times |
Joined on Jan 2011
|
#215
|
Originally Posted by freemangordon
From this morning I am using:
I tend to support that, I am running /proc/sys/vm/page-cluster = 0 for ages and it definitely makes a difference.
On the other hand it seems there should be some system tweaking for compcache to be effective. Unfortunately I have NFC where and what
64MB ramzswap
swap on eMMC
echo 3 > /proc/sys/vm/dirty_ratio
echo 3 > /proc/sys/vm/dirty_background_ratio
echo 100 > /proc/sys/vm/dirty_writeback_centisecs
echo 100 > /proc/sys/vm/dirty_expire_centisecs
echo 32 > /proc/sys/vm/min_free_kbytes
echo 95 > /proc/sys/vm/swappiness
echo 666 > /proc/sys/vm/vfs_cache_pressure
echo 0 > /proc/sys/vm/page-cluster
echo 4 > /sys/block/mmcblk0/queue/nr_requests
echo 4 > /sys/block/mmcblk1/queue/nr_requests
All is as smooth as Bulgarian butter so far, and yes I have opened many webpages, have a video playing in the background and conversations+images+file manager.
Cat /proc/swaps shows ramz is well full ages ago.
Will report later.
__________________
N900: One of God's own prototypes. A high-powered mutant of some kind never even considered for mass production. Too weird to live, and too rare to die.
| The Following 10 Users Say Thank You to vi_ For This Useful Post: | ||
 |
|
2012-05-27
, 12:38
|
|
Posts: 139 |
Thanked: 181 times |
Joined on Nov 2011
@ Oulu, Finland
|
#216
|
...again with the butter... but this is quite interesting, do report back! Thanks.
Last edited by Raimu; 2012-05-27 at 12:45.
Last edited by Raimu; 2012-05-27 at 12:45.
|
|
2012-05-27
, 14:53
|
|
Posts: 669 |
Thanked: 433 times |
Joined on May 2010
|
#217
|
Does the "ramzswap" suffer from fragmentation issue same as with the mmc/internal swap? does it need to get swapoff+swapon also to fix it?
|
|
2012-05-27
, 15:57
|
|
Posts: 3,074 |
Thanked: 12,960 times |
Joined on Mar 2010
@ Sofia,Bulgaria
|
#218
|
Originally Posted by impeham
ramzswap ignores Nokia swap algorithm (i.e. the one that tries to write using large blocks of consecutive pages). And there is no penalty for writing non-consecutive blocks, as there is no physical media present. There could be penalty for fragmentation, but the algorithm used in ramzswap for memory allocation/distribution is pretty effective afaik.
Does the "ramzswap" suffer from fragmentation issue same as with the mmc/internal swap? does it need to get swapoff+swapon also to fix it?
__________________
Never fear. I is here.
720p video support on N900,SmartReflex on N900,Keyboard and mouse support on N900
Nothing is impossible - Stable thumb2 on n900
Community SSU developer
kernel-power developer and maintainer
Never fear. I is here.
720p video support on N900,SmartReflex on N900,Keyboard and mouse support on N900
Nothing is impossible - Stable thumb2 on n900
Community SSU developer
kernel-power developer and maintainer
| The Following 4 Users Say Thank You to freemangordon For This Useful Post: | ||
 |
|
2012-05-27
, 18:50
|
|
Posts: 4,117 |
Thanked: 8,901 times |
Joined on Aug 2010
@ Ruhrgebiet, Germany
|
#219
|
vi_
any (good) news?
What I think I forgot to mention in my former post, that when ramzswap was enabled and I had both other (eMMC, SD) swaps enabled, both were used simultaneously (same amount of data written); even the SD swap had higher priority than eMMC.
any (good) news?
What I think I forgot to mention in my former post, that when ramzswap was enabled and I had both other (eMMC, SD) swaps enabled, both were used simultaneously (same amount of data written); even the SD swap had higher priority than eMMC.
|
|
2012-05-27
, 20:12
|
|
Posts: 669 |
Thanked: 433 times |
Joined on May 2010
|
#220
|
Originally Posted by vi_
I've tested same configuration as yours since you posted it - the device's response feels amazing - smooth and fast. i've tested with several applications open which affected the desktop response before and not anymore.
From this morning I am using:
64MB ramzswap
swap on eMMC
echo 3 > /proc/sys/vm/dirty_ratio
echo 3 > /proc/sys/vm/dirty_background_ratio
echo 100 > /proc/sys/vm/dirty_writeback_centisecs
echo 100 > /proc/sys/vm/dirty_expire_centisecs
echo 32 > /proc/sys/vm/min_free_kbytes
echo 95 > /proc/sys/vm/swappiness
echo 666 > /proc/sys/vm/vfs_cache_pressure
echo 0 > /proc/sys/vm/page-cluster
echo 4 > /sys/block/mmcblk0/queue/nr_requests
echo 4 > /sys/block/mmcblk1/queue/nr_requests
All is as smooth as Bulgarian butter so far, and yes I have opened many webpages, have a video playing in the background and conversations+images+file manager.
Cat /proc/swaps shows ramz is well full ages ago.
Will report later.
Thanks for sharing
| The Following 3 Users Say Thank You to impeham For This Useful Post: | ||
swapon /dev/mmcblk0p3
insmod /lib/modules/current/ramzswap.ko disksize_kb=65536
swapon /dev/ramzswap0 -p 10
swapoff /dev/mmcblk0p3
it seems that enabling the internal memory's swap was needed to avoid the crash - i don't know why.
Now i have 2 swaps:
/dev/mmcblk1p4 (P -1)
/dev/ramzswap0 (P - 10)
Thanks for the help.