Guest laum Posted July 7, 2010 Report Share Posted July 7, 2010 Guys, I see that the file system is mounted without noatime option. Could this cause the problem with the delays? So this basically means that every single read from database files updates also the last access time? Link to comment Share on other sites More sharing options...
Guest ssarl Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) So this basically means that every single read from database files updates also the last access time? It seems so, unless it is the default option to not update access time. Paul, what you think? Edited July 7, 2010 by ssarl Link to comment Share on other sites More sharing options...
Guest Kilack Posted July 7, 2010 Report Share Posted July 7, 2010 Hope someone can figure this out: http://www.samsung.com/global/business/sem...4_3_support.pdf KLMAG8DEDD http://www.careace.net/wp-content/uploads/...assembly-14.jpg seems like that chip is... moviNAND, 16G Byte, NAND8 + Microcontroller, MLC 16G Bit, Dual 1.8V(1.65V~1.95V) 3.3V(2.5V ~ 3.6V), S3C49VDX01 blah blah :) so yes that's one 16gig chip of nand... Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 So this basically means that every single read from database files updates also the last access time? By doing this from a SU adb shell, someone who hasn't lag-fixed can try? mount -o remount,noatime /dev/block/mmcblk0p2 /data P Link to comment Share on other sites More sharing options...
Guest rasp Posted July 7, 2010 Report Share Posted July 7, 2010 seems like that chip is... moviNAND, 16G Byte, NAND8 + Microcontroller, MLC 16G Bit, Dual 1.8V(1.65V~1.95V) 3.3V(2.5V ~ 3.6V), S3C49VDX01 blah blah :) so yes that's one 16gig chip of nand... MLC 16G Bit <-- 2GB MLC.. I've found 1.4GB accounted for from the partition table.. 600mb missing, so is this the 2GB? Link to comment Share on other sites More sharing options...
Guest Kilack Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) MLC 16G Bit <-- 2GB MLC.. I've found 1.4GB accounted for from the partition table.. 600mb missing, so is this the 2GB? Which is the size of that chip? First it says its a 16Gbyte for the "NAND density" Then "16G bit" is under the table NAND Small classification..and is multibit level cell (2bit) Edited July 7, 2010 by Kilack Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 MLC 16G Bit <-- 2GB MLC.. I've found 1.4GB accounted for from the partition table.. 600mb missing, so is this the 2GB? How are you breaking down that 1.4GB? P Link to comment Share on other sites More sharing options...
Guest rasp Posted July 7, 2010 Report Share Posted July 7, 2010 How are you breaking down that 1.4GB? P major minor #blocks name 139 0 513024 tfsr0/c 139 1 256 tfsr1 139 2 256 tfsr2 139 3 10240 tfsr3 139 4 1280 tfsr4 139 5 1280 tfsr5 139 6 5120 tfsr6 139 7 7680 tfsr7 139 8 7680 tfsr8 139 9 313856 tfsr9 139 10 116736 tfsr10 139 11 35840 tfsr11 139 12 12800 tfsr12 179 0 16056320 mmcblk0 179 1 14090208 mmcblk0p1 179 2 1966080 mmcblk0p2 179 8 7974912 mmcblk1 179 9 7974847 mmcblk1p1 137 0 513024 bml0/c 137 1 256 bml1 137 2 256 bml2 137 3 10240 bml3 137 4 1280 bml4 137 5 1280 bml5 137 6 5120 bml6 137 7 7680 bml7 137 8 7680 bml8 137 9 313856 bml9 137 10 116736 bml10 137 11 35840 bml11 137 12 12800 bml12 138 3 6400 stl3 138 6 1280 stl6 138 9 304640 stl9 138 10 111104 stl10 138 11 32000 stl11 tfsr - ~512mb bml - ~512mb stl - ~ 455mb that adds up to 1.4-1.5gb so 500-600mb missing if that is the 2GB oneNAND in moviMCP chip Link to comment Share on other sites More sharing options...
Guest Kilack Posted July 7, 2010 Report Share Posted July 7, 2010 Mmmm 500-600 megs... that would be nice to get our hands on. Surely Samsung could better arrange everything on this phone and get it going superfast... Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 major minor #blocks name 139 0 513024 tfsr0/c 139 1 256 tfsr1 139 2 256 tfsr2 139 3 10240 tfsr3 139 4 1280 tfsr4 139 5 1280 tfsr5 139 6 5120 tfsr6 139 7 7680 tfsr7 139 8 7680 tfsr8 139 9 313856 tfsr9 139 10 116736 tfsr10 139 11 35840 tfsr11 139 12 12800 tfsr12 179 0 16056320 mmcblk0 179 1 14090208 mmcblk0p1 179 2 1966080 mmcblk0p2 179 8 7974912 mmcblk1 179 9 7974847 mmcblk1p1 137 0 513024 bml0/c 137 1 256 bml1 137 2 256 bml2 137 3 10240 bml3 137 4 1280 bml4 137 5 1280 bml5 137 6 5120 bml6 137 7 7680 bml7 137 8 7680 bml8 137 9 313856 bml9 137 10 116736 bml10 137 11 35840 bml11 137 12 12800 bml12 138 3 6400 stl3 138 6 1280 stl6 138 9 304640 stl9 138 10 111104 stl10 138 11 32000 stl11[/code] tfsr - ~512mb bml - ~512mb stl - ~ 455mb that adds up to 1.4-1.5gb so 500-600mb missing if that is the 2GB oneNAND in moviMCP chip So what are you saying tfsr and bml are? I don't think you're right on that one i'm afraid. P Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) By doing this from a SU adb shell, someone who hasn't lag-fixed can try? mount -o remount,noatime /dev/block/mmcblk0p2 /data Did it. Still alive. But I'm not sure about lag, looks fast for me. Will try to start some benchmark. BTW, how to remount it to the original state (for comparision)? Edited July 7, 2010 by Archer13 Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) Default state, just after restart: /dev/block/mmcblk0p2 on /data type rfs (rw,nosuid,nodev,vfat,llw,check=no,gid/uid/rwx,iocharset=utf8) Filesystem Total file system score 29.770191 Creating 1000 empty files 42.955 sec Deleting 1000 empty files 24.973 sec Write 1M into file 3.0750308 M/sec Read 1M from file 56.81818 sM/sec SD card perfomance Creating 250 empty files 15.584 sec Deleting 250 empty files 9.908 sec Write 1M into file 7.0077085 M/sec Read 1M from file 227.27272 sM/sec After Paul's command "mount -o remount,noatime /dev/block/mmcblk0p2 /data": /dev/block/mmcblk0p2 on /data type rfs (rw,noatime,vfat,llw,check=no,gid/uid/rwx,iocharset=utf8) First run: Filesystem Total file system score 70.81754 Creating 1000 empty files 41.438 sec Deleting 1000 empty files 29.886 sec Write 1M into file 3.5855145 M/sec Read 1M from file 138.88889 sM/sec SD card perfomance Creating 250 empty files 10.832 sec Deleting 250 empty files 40411 sec Write 1M into file 26.24672 M/sec Read 1M from file 285.7143 sM/sec Second run Filesystem Total file system score 65.270615 Creating 1000 empty files 40.866 sec Deleting 1000 empty files 29.831 sec Write 1M into file 1.4446691 M/sec Read 1M from file 129.87013 sM/sec SD card perfomance Creating 250 empty files 48.84 sec Deleting 250 empty files 8.748 sec Write 1M into file 22.421524 M/sec Read 1M from file 454.54544 sM/sec Edited July 7, 2010 by Archer13 Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) # mount -o remount,noatime,nosuid,nodev /dev/block/mmcblk0p2 /data /dev/block/mmcblk0p2 on /data type rfs (rw,nosuid,nodev,noatime,vfat,llw,check=no,gid/uid/rwx,iocharset=utf8) Filesystem Total file system score 85.69356 Creating 1000 empty files 41.476 sec Deleting 1000 empty files 27.428 sec Write 1M into file 2.9112082 M/sec Read 1M from file 169.49153 sM/sec SD card perfomance Creating 250 empty files 13.012 sec Deleting 250 empty files 17.012 sec Write 1M into file 22.172949 M/sec Read 1M from file 270.27026 sM/sec Edited July 7, 2010 by Archer13 Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 After the LagFix script and restart: Filesystem Total file system score 202.9743 Creating 1000 empty files 3.716 sec Deleting 1000 empty files 3.034 sec Write 1M into file 8.3542185 M/sec Read 1M from file 400.0 sM/sec SD card perfomance Creating 250 empty files 10.708 sec Deleting 250 empty files 15.036 sec Write 1M into file 25.575447 M/sec Read 1M from file 370.37036 sM/sec BIG difference. Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 All previous tests were made with application Benchmark Link to comment Share on other sites More sharing options...
Guest Tyxerakias Posted July 7, 2010 Report Share Posted July 7, 2010 Wow. But on restart it defaults back? Link to comment Share on other sites More sharing options...
Guest Archer13 Posted July 7, 2010 Report Share Posted July 7, 2010 Wow. But on restart it defaults back? Yes, no "atime" after restart. Link to comment Share on other sites More sharing options...
Guest Tyxerakias Posted July 7, 2010 Report Share Posted July 7, 2010 Yes, no "atime" after restart. So is there a way of adding it to the boot sequence? There's no fstab in Android but there must be something else, via init.rc, something to make it stick? Paul must know :) Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 So we can't get close to the internal NAND speed, not surprising... looks like noatime / nosuid help tho? P Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 So is there a way of adding it to the boot sequence? There's no fstab in Android but there must be something else, via init.rc, something to make it stick? Paul must know :) Adding stuff to the boot sequence isn't easy atm, i've been looking into it... P Link to comment Share on other sites More sharing options...
Guest rasp Posted July 7, 2010 Report Share Posted July 7, 2010 So what are you saying tfsr and bml are? I don't think you're right on that one i'm afraid. P this should explain what they are.. http://www.samsung.com/global/business/sem...cation_note.pdf im not sure now if it is possible to use those 500-600mb of unknown space since the FTL ensures that wear leveling is distributed evenly across the media.. Link to comment Share on other sites More sharing options...
Guest gravityz3r0 Posted July 7, 2010 Report Share Posted July 7, 2010 (edited) Adding stuff to the boot sequence isn't easy atm, i've been looking into it... P Paul is the lag fix included actually what we have been doing previously? i.e (moving stuff from data/data to dbdata) Edit : Offtopic : btw Paul your prediction about the result of the match between Germany and Spain is CORRECT YET AGAIN! :) Edited July 7, 2010 by gravityz3r0 Link to comment Share on other sites More sharing options...
Guest Paul Posted July 7, 2010 Report Share Posted July 7, 2010 Yeah... P Link to comment Share on other sites More sharing options...
Guest yingwai Posted July 8, 2010 Report Share Posted July 8, 2010 Not working for I9000ZSJF7 Link to comment Share on other sites More sharing options...
Guest lucasjohny Posted July 8, 2010 Report Share Posted July 8, 2010 tested on my JF4 singtel ROM, work 100% no more stalling issues, thanks bro's... Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now