How is ExpressCache 1.3.2 working?

It would be great if SanDisk can take a active role in this thread and address these issues.

It would be great if SanDisk can take a active role in this thread and clarify/address these remaining EC issues.

The finish line is within sight… I hope.

It would be great if SanDisk could take an active role in this thread and clarify/address these remaining EC issues.

From another post:

My cache is filled up as the screenshot shows and it take 1.35 minutes to get to the login screen!!! 

this is with 1.3.2.  I actually think my PC might boot up just as fast or faster without ReadyCache.

cache - https://www.dropbox.com/s/hgpj4cvmjhompan/cache.JPG

boot - https://www.youtube.com/watch?v=40FpvBixpiI

Motherboard - GA-MA78GPM-UD2H (rev. 1.0)  AMD 780G + SB700 Chipset

Windows 7 64-bit

This needs to get fixed pronto or I am going to have to remove this drive from my PC.

Constant resets, and yes no imprvement in performance, might actually a drag esp considering the never ending defrag

Beyond the lack of performance i think its beating itself to death.   SSDs have finite lifespans, usually expect a few gb in writes a day as normal usage for their life estimates…and thats on real ssds with good controllers, this is more a high end flash thumb drive.  So flush just 3 times a day and thats 100gb.  Its just not going to last…while doing nothing either  

It also keeps other drives running.

I’m at my wits end on this

Hi all,

I politely ask slotmonsta, gls, drlucky and other SanDisk employees to urgently help and comment here.

You can also PM them and ask that they reply in this forum. Procrastination prevents progress.

Flavio

This morning (system on overnight) ExpressCach had an error that the drive was not available. I restarted the GUI without rebooting, and ExpressCache came right up. The error is not something that I have ever seen before.

This past week, my system has been hanging with frustrating frequency. I’m not ready to blame ExpressCach, but neither am I ready to rule it out.

had significantly fewer errors with this release, and though it might be something i’ve done i have noticed the cache fills with a lot more data as well as a lot faster after a reset.

seems overall better than the last version.

After instalation, most problems of 1.3.1 remained on my SB850 board (I’ve tested this system >12h prime stable multiple times). The delayed boot with cache >20GB, and a tendency for a system crash on boot when cache is about - 28GB filled (stuck loading windows with hdd led on, both on w7 sp1 or w8.1 x64), or even stuck when idle after some hours (probably when ready cache is doing some maintenance). Excluding drives doesn’t make a difference for this behaviour, but disconnecting the SSD to make Readycache stop working restores system stability.

Right after setting cache to only 16GB (and also excluding 2 hdd I don’t want cached), I haven’t had any problems when cache is nearly filled , either the usual freezed boot delay or system crashes for the last two weeks.

Couldn’t this be some Sandforce firmware issue that causes trouble with an almost full drive with some SATA controllers, and Readycache simply triggers it?

I don’t seem to have “cache has been reset” problems. When it happens it’s usually due an understandable reason that triggers it (a manual or scheduled disk defrag to cached drives, updates on boot, large file changes, etc.).

Also to look in too, if not solved already, when I installed 1.3.0 on this pc that had 4 hdds with 7 visible partitions Readycache wouldn’t work. Cache would always stay at 0.07GB. After reducing the number of partitions (by merging on the same drives) to 4, it started caching normally.

1.3.2 resulted in 3 blue screens of death in one week.  Prior to installing  1.3.2/1.3.110, I had disabled caching, and my Windows 7 system was rock solid, if a little slow.  Since installing 1.3.2, the same instability which caused me to disable it previously has reemerged.

mattschnaidt, maybe you could try as I did. Since reducing cache to only 16GB stability problems seem to have disappeared.

To make it so, I’ve run diskpart on cmd admin:

diskpart (opens diskpart)

list disk (list drives on system, take note of #)

select disk # (to select the sandisk SSD listed with 29GB)

clean (clears partitioning, make sure you have the correct #, else you’ll lose some other drive’s data)

exit (exits diskpart)

eccmd -partition # 16384 (creates a 16GB cache partition for use)

eccmd -format (makes that partition ready for readycache use)

After that, launch Readycache, skip the error message, and size changes should work immediately. A reboot should be advisable anyway.

To revert to a full drive in use, just repeat the procedure but instead do “eccmd -partition #” without specifing any size. No need to reinstall software.

To make sure this is a readycache related bsod, you could run bluescreen view and analyse the dump to check the BSOD code and most important files on use at that moment.

AlleyViper, I’m noticing the boot delay and what my feeble memory says are cache operation slowdown as I approach 28Gb.   I’d like to try reducing my cache partition size as you show above, but I want to make sure I get the syntax right as I’ve gotten used to GUI partitioning utilities.

For comparison, here is a method slotmonsta posted in the “ReadyCache ssd hangs on startup” thread to make an 8k partition:

  1. From the command line: ECCmd -format (this will clear the information out of the cache) 
  2. Delete the partition (this is done from the Disk Management pain by right clicking on the drive and selecting “delete” the partition 
  3. From the command line: ECCmd -partition (drive number) 8192 (this will create a partition of 8GB in size) 
  4. From the command line: ECCmd -format (this will format the new partition and make it ready for EC/RC to utilize)

So:   eccmd -partition requires a drive number, but eccmd -format does not?

Thnaks for helping someone who started with 8" floppys but hasn’t used the command line in years and doesn’t want to partition or format their hard drive!.

Hi,

Doing slotmonsta’s procedure should end up the same (except for exact cache size). I just find it more practical to use diskpart as it will list the correct drive number to use next with eccmd, that is also CMD based. It’s just a matter of doing less clicks, but I also skip the first format instruction because the partition will be deleted anyway. Just do what you’re more confortable with!

eccmd -partition command can be used with an optional drive ID, which I find safer to specify (as slotmonsta first instructed), but eccmd -format doesn’t allow other commands/switches, so I assume it’ll only work when it finds a valid partion on a valid sandisk sdd. If you just run eccmd these available commands are described (except for unsuported but working: -exclude Driveletter; -clearexclusions; -preload Filename [usagecount]).

Btw, I chose to use 16GB (after having good results with only 8GB, that is a bit short for caching, for a few days) because most problems on my system seem to start when there’s about >20GB filled cache. Either a delayed boot, or some random hard crash when cache is nearly filled.

Thanks very much AlleyViper!   Your explanations give me the confidence to give this a try.

Unexpectedly, when I examind my original cache partition it showed the starting offseet as 2048.   I’d always heard to start at 4096 (or multiples) to get a 4k alignment  ?? 

I was researching SSD partitioning and ran into the following: http://www.tomshardware.com/forum/292105-32-best-format-partition-performance-wear-leveling#6100306 .  One thing they say is that trying to use more than 80% of the allocated space on an SSD will result in performance issues.   Well, 0.8 x 29 GB = 23.2 GB – that number looks familiar!     Maybe these full cache slowdowns are just symptoms of a normal ssd issue of needing at least 20% slack space, and reducing the cache partition size is the actual fix and not just a work-around…

1024/2048/4096KB is also 4K aligned, so don’t worry. With diskpart you can check current partition offset with “list partition” after “select disk #”. Currently my 16GB readycache partition (created with eccmd) has a valid 1024KB offset.

What you pointed is the reason why I’ve already asked if this isn’t simply some sandforce SSD firmware issue (in case this is an old SF based SDD). The amount of needed freespace for consistent performance can vary alot with the controller/firmware, some older drives might even need half of the size free to make up for lacking garbage collecting, wear leveling, trim etc.

But such things *should* cause a noticeable IOps drop, not a total halt for some seconds like we experience on boot, so caching software is also a big suspect.

Anyway, I’m using only 16GB because I’ve seen delays with less than 24GB cached. Otherwise, I’d set a 24GB limit if software was working at 100% to ensure better wear leveling for such small drive. Strangely, my 32GB SSD came with 1 realocated sector from factory, which put health to 95%.

well up to now my 32gb readycache has been working great , however a few niggles ( but not complaints ) are 

  1. during windows 7 boot up the logo will pause ( like its crashed ) however around 5 seconds later it will continue and windows will load as per normal 

  2. the network takes much longer to load …  so much so i am now manually starting my steam client as this loads very quickly … ive asked about a solution for this on this forum ( http://forums.sandisk.com/t5/SanDisk-ReadyCache-SSD/network-taking-a-long-time-to-connect-since-installing/td-p/320394 )

3)the readycache is so quick i know dont see the windows welcome screen … it skips this and the desktop loads … this isnt a complaint,however it would be nice if we could fine tune the readycache …!

         apart from these im extremely happy with my readycache ssd …  and its well passed my expectations !

AlleyViper - thanks so much for your response.  Just followed your instructions (which were very easy to follow!), and I’m hoping for a good result!  Will report back to let you know how it went.  Thanks again!

Ok so capped my readycache partition at 16Gb using the below instructions:

  1. From the command line: ECCmd -format (this will clear the information out of the cache)
  2. Delete the partition (this is done from the Disk Management pain by right clicking on the drive and selecting “delete” the partition
  3. From the command line: ECCmd -partition (drive number) 16384 (this will create a partition of 8GB in size)
  4. From the command line: ECCmd -format (this will format the new partition and make it ready for EC/RC to utilize)

Here is a shot from ExpressCache - https://www.dropbox.com/s/ipdhp1fnzr9i6tl/cache16gbpartition.JPG

Here is my sub 1 minute boot to desktop - https://www.youtube.com/watch?v=GEJDe0XFOZ4

Here is my very long boot just to login screen before partitioning to 16Gb - https://www.youtube.com/watch?v=40FpvBixpiI&list=UUh6FciVZH5s5E3TNHhjV9LQ

So far so good!!! I will update if things change.  Sandisk needs to fix this problem!!!

rlewandowski23 that’s strange, using -partition # 16384 gives me an exact 16GB total on the GUI. About 12-14GB will be filled then.

Good luck to anyone who tried and can post her/his findings. I’ll be away from my test machine for a couple of weeks.

I used eccmd -partition (drive # shown in diskpart, in my case 1) 16384 , and ExpressCache splash screen, eccmd -info, and MiniTool Partition Wizard all show a 16.0 GB partition.   Maybe you typo’d 18384 – that’s very close to the 18432 of an 18 GB partition.

I’d leave it be at 18 GB – my memory is that AlleyViper was the only person obsrving delays as low as 17 GB, other reports were 23 GB and up.   If you can, let us know what happens as you fill past 17 GB.