How is Primo Cache better than Superfetch?

[Davey126]

Since you are using an EVO for the system, read+write caching might be a good idea to have those TLC chips to last longer, I would never use a TLC based SSD as system drive and especially not for important data.

SSD write endurance for consumer grade drives in typical usage scenerios is proving to be a non-issue based on user reports and more structured testing such as that being conducted by The Tech Report which has writen over 500TB to a handful of popular consumer drives with no catastrophic failures. SSDs still suffer far too many failures due to other causes (often traced to firmware glitches) but there does not appear to be a correlation to write activity.

[MaXimus]

I requested a refund as I didn't see any difference between primocache and SUPERFETCH. not worth my moneh.

maybe because I have an SSD and 32 GB RAM so it was fast as heck anyway, I don't know what I was expecting but on white paper primocache sounds great, in reality, it didn't do anything for t3h m3h

[Violator]

Since you are using an EVO for the system, read+write caching might be a good idea to have those TLC chips to last longer, I would never use a TLC based SSD as system drive and especially not for important data.

SSD write endurance for consumer grade drives in typical usage scenerios is proving to be a non-issue based on user reports and more structured testing such as that being conducted by The Tech Report which has writen over 500TB to a handful of popular consumer drives with no catastrophic failures. SSDs still suffer far too many failures due to other causes (often traced to firmware glitches) but there does not appear to be a correlation to write activity.

I was not even in on enterprise SLC / MLC.
TLC (1000 to 5000 write/erase cycles) has less lifetime compared to MLC (3000 to 15000 write/erase cycles), only talking consumer drives here.
TLC was cheaper but with todays prices you can get MLC for close to the same price, just look at performance and price of a 500GB Crucial M500 compared to a Samsung EVO of the same size.
I bet Tech Report did test with consumer MLC SSD's?

[Davey126]

I bet Tech Report did test with consumer MLC SSD's?

Yup - all consumer grade SSDs including one Sammy EVO. That drive was the third to fail after nearly a petabyte of data had been written to it (first two were from Intel and Kingston; both MLC). That is way beyond the rated life cycle and ridiculously more writes than a consumer drive will typically see during its lifetime. NAND technology is only one of several factors I consider when purchasing a SSD; I certainly don't worry about drive failures due to excessive writes in a consumer (SOHO) application.

Here's the latest update from The Tech Report: http://techreport.com/review/26523/the- ... a-petabyte

[Violator]

Pro not the Evo, although I would take the testing method with a gran of salt since it doesn't cover real usage.
But even with MLC's labeled lifespan that still makes it last 18 years when you modify/delete/write 20GB of data per day @ 3x write amplification.

This might be a better way to test: http://ssdendurancetest.com/

[Davey126]

Pro not the Evo, although I would take the testing method with a gran of salt since it doesn't cover real usage.
But even with MLC's labeled lifespan that still makes it last 18 years when you modify/delete/write 20GB of data per day @ 3x write amplification.

This might be a better way to test: http://ssdendurancetest.com/

Both PRO and EVO. The testing methodology is well documented and attempts to include some real world elements such as limited power cycles, use of compressed/uncompressed data for some samples and regularly shelving drives for 7 day intervals to insure they retain data after heavy abuse. But in the end you can not simulate real world work loads and reach the failure point within a reasonable timeframe without some serious acceleration. I would argue (respectfully) the methodology used by Tech Report is more realistic than drawing conclusions from SMART data which can be tweaked by drive manufactures both in implementation and response to reported results; some drives brick themselves while others continue on until data loss occurs. Both testing approaches have their place and SMART data is discussed in each installment of findings published by TR. In most cases SMART did predict immanent failure which is reassuring. That said, few consumers pay attention to SMART and many BIOS/UEFI implementations have it disabled by default.

"Our test subjects include six SSDs designed for consumer desktops and notebooks: the Corsair Neutron GTX 240GB, Intel 335 Series 240GB, Kingston HyperX 3K 240GB, and Samsung 840 Pro 256GB, which are all based on two-bit MLC flash, and the Samsung 840 Series 250GB, which uses three-bit TLC NAND."

[Davey126]

FYI - summary of Tech Report findings to date on SSD longevity (quick read): http://www.extremetech.com/computing/18 ... oud-expect

[InquiringMind]

...most mainstream users should not enable deferred writes on the system partition if they do not have a UPS or solid backup strategy...

Since a UPS wouldn't help in the event of a bluescreen (and these would be more likely than a power failure for most), I'd hesitate to make it a recommendation except for those with an unreliable power supply.

A multi-layered backup strategy though is excellent advice, whatever the situation. A combination of regular disk image backups, ERUNT for registry snapshots and a file versioning utility (which copies files to another folder or disk whenever they are modified) for user data (I prefer Aphar Backup but AutoVer, Yadis Backup or FileHamster are viable alternatives) should suffice for most people.

[Dhruv96]

Hello guys....
I have a issue here

My specification is
-Intel core i7 4940MX quad mobile extreme core
-32GB of kingston hyperX laptop cl11 RAM oc'ed to 2400Mhz
-1TB hybrid Drive 5400rpm with 16GB cache(not accessible)


My issue is primocache slowed my computer by a lot(boot up time and shutdown times, game stuttering)......
My config here is
-10GB for cache
-Tried all sort of block size(using 16KB as it was default)
- defer writing also tried all and custom to(using 10sec right now) and tried read only

Using eboostr is somehow improving/tackling primocache slow boot up and shutting down....

I fixed all DCOM, activeX and superfetch problem manually.....fixed means total fixed, and no custom value

Windows fully stable......

Help me with slower boot up and shutdown issue

When i switch off primocache....it runs fine

And game always hang/stutter when defer time is done(the more hang at more defer rates, it is like if 10sec then the game will hang less but at interval of 10sec.....but if 4min, then it will hang after 4min but hang time woulf be a lot)


No IO issue with HDD too

Paging is done by windows

And dont flush data is unchecked in device manager


Primocache even when copying data.....goes insane fast for sec and then comes to even lower speeds than normal


All benchmarks are near to fake.....
pls suggest me a setting as i have seen many people positive review....i want to get the same

No l2 cache on lappy

I tried dataram ramdisk as l2 but it slowed the pc even more

App takes longer to load after primocache

[InquiringMind]

Welcome to the forums Dhruv96,

I presume you've thoroughly tested your memory OC (using the likes of Prime95 or a similar benchmark run over a couple of hours or so) but it may still be worth resetting your memory to its original speed to see if that makes a difference.

It might also be worth trying a non-hybrid hard disk if you have one (use an imaging utility to copy your setup onto it and boot from that) to rule out any conflict there.

Finally, which version of Windows are you using and are you running any other memory-caching or disk-related utilities? You mention trying Dataram so it might be worth running a simple benchmark (like CrystalDiskMark) on it to see if it provides the speed expected (I'd expect in excess of 10GB/s sequential).