[2017-05-27] PrimoCache 3.0.0 Beta is available now!

[sinus]

I do not know whether there is a system defect here, or it is feature of the software. I have read that was written by users and what they have stated wishes. All users have own remarks, people are interested. I bought this software because it is interesting to me, and I apply it on the home computer. Perhaps, I in vain waste time?

[Davey126]

I do not know whether there is a system defect here, or it is feature of the software. I have read that was written by users and what they have stated wishes. All users have own remarks, people are interested. I bought this software because it is interesting to me, and I apply it on the home computer. Perhaps, I in vain waste time?

Perhaps - expectations differ among individuals. My assessment is that the various write commit modes are working as designed and documented. Both in this version and those immediately prior. To date there has been no community discussion that suggests otherwise.

You have shared observations that suggest write modes behave differently on your workstation/workflows. If you believe there is an undiscovered product defect I suggest raising an official bug report with Romex support.

[Jaga]

I have no problem with the write cache modes. I've used perhaps 3 in the past, settled on 1-2, and they always did what I thought they'd do. That isn't to say I don't have ideas on other modes that might be better for other circumstances, but that's a discussion for elsewhere.

[sinus]

To Davey126: I have already specified - I don't know whether there is a system defect here or it's software feature. I have described in detail what has noticed, and no more than that.
Jaga, Your PrimoCache never "urgent" flush onto disk?

[Jaga]

Nope, every time I look at PrimoCache stats, Urgent writes are zero. Even if I set the cache on an abnormally long timeout like 300 seconds. I'm not using a L2 cache however, just a very large L1 cache (~32 gig) on 2 separate Cache Tasks covering three volumes on two drives. It can take a week depending on what I'm doing to fill it entirely from clear. But when I do use the machine, I use it heavily. Still never seen any urgent writes.

[sinus]

I selected 8 GB, 12 GB and 16 GB for L1. Jaga, Your 32 GB it is very great cache, but even this big cache can full, isn't it? As in your opinion what will occur further? How you think what will occur further if new data continue filling of a cache quicker, than it will be exempted from deferred data?

[sinus]

Any assumptions in science and practice always either are confirmed experimentally, or are rejected. Without having results of experiments, there is no sense to argue because it is not soccer and not policy. The experiment must be clean and clear: established checkbox of "Enable Defer-Write", the "Latency" value is enable, in "Advanced Defer-Write Options" select Write Mode "Intelligent", "Idle-Flush" or "Buffer" (with condition of "Windows idle"). Some background task is carried out not to allow "Windows idle", and start of a procedure, for example, copy files on cached disk; (total size of files has to be more than Cache Size).

[idefix44]

To sinus:
You're right. My L1 cache is 512MB and I have the same problem each time I run Firefox ( several times per day of course ) when the OS ( Windows XP 32 ) need memory and began to swap.
My Defer-Write is 60s, mode Buffer.

[Jaga]

Then perhaps the issue is an OS that has too much of it's memory taken by Primocache. Or it's just a memory-starved system, so Primo isn't "fitting in" well for it. That's not Primo's fault - give the system more memory perhaps?

[Jaga]

Sidenote: I managed to force some Urgent writes today. I had to start a 25 gigabyte FTP download, as well as do some unzipping of large .7z archives, all on the same volume at the same time. Add to that, the volume only has a 8gb cache task for Primocache (my smallest).

So yes, it's possible to get the cache into a state where urgent writes are needed. But the amount of disk activity has to be absolutely gigantic (in my experience), or completely clog up the write cache which is what I managed to do today.

Still looks to me like it's working as designed. If you have that much write activity going on, you are going to fill up the cache one way or another, and urgent writes are probably going to happen.