Effect on Boot Times? Topic is solved

[BikeHelmet]

Hello there,

My computer loads a fair bit of stuff when it starts. I've got it set to open my web browser (Firefox), load a bunch of background apps, then start up Steam. This normally takes about 85 seconds. Keep in mind that POSTing and booting is much of that time - my resolution flips to my desktop resolution at around the 40 second mark.

I was trying out SuperCache, and it dropped the time to load everything from 85 seconds to 65 seconds. Supercache absorbs all the writes, allowing my drive to do nothing but read. This seemed to drop my program start time in half. (I have a visible timer bar that counts upward until Steam finishes loading, so it really did) If not its instability (crashes) and volume caching (rather than disk caching), I'd consider calling it a winner due to those results alone. But instability is bad, so I went looking for alternatives.

Now I'm trying out FancyCache, and surprisingly, it doesn't affect boot times nearly as much. It only dropped them to around 80 seconds. I've got three theories.

The first is that FancyCache doesn't kick in as fast. (unlikely)
The second is that I have some setting set wrong. (quite likely)
The third is that the Deferred Write algorithm does something different. (possible)

I think it's probably #2 or #3. In the case of #3 - if I set it to 30 seconds, what are the odds FancyCache will completely avoid writes for at least 25-30 seconds? I wonder if it would it be more useful to split deferred writes into two or three separate values, to allow fine-tuning the algorithms to have different behaviour?

[Support]

Hi,

First, please make sure that the SuperCache is totally uninstalled before doing the comparison testing. If it is not uninstalled, it might have the filter driver working, even you didn't enable caching, thus affect the results.

The settings might affect the performance greatly. Can you upload a setting screenshot?

FancyCache now has two main Defer-Write algorithms. For your reference, please see
viewtopic.php?f=26&t=735

[BikeHelmet]

Hello,

I have ensured that Supercache is totally gone.

Interesting info on those Defer-Write algorithms!

Here's a screenshot. I've tried it both with and without Write Averaging ticked.

[Support]

Hi,

I think you shall use FancyCache For Volume Edition if you try to make a comparison between SuperCache and FancyCache. Because the Disk Edition caches all disk's data while the Volume Edition caches one volume's data. And of course, you shall setup the same Level-1 cache size while doing a comparison. (You might disable Level-2 caching)

And it is recommended that to match the block size with the cluster size, please refer to
viewtopic.php?f=25&t=706

[BikeHelmet]

Wow, you were right about those block sizes. The lower I go, the faster it is!

But.. I can't seem to set it below 16KB? I get this error message. Maybe I don't have the cache size set large enough? (Well, I can't go much higher on 32bit Windows)

[Support]

Smaller block size requires more additional memory, especially when the target volume/disk size is quite large.
This is another reason to use FancyCache Volume edition to cache specified volume instead of the whole disk.

[BikeHelmet]

Well, now that I've tweaked settings a bit better, it's a lot closer. It's down to between 70-75 seconds. Probably about ~72, which is reasonably close. The difference might be from slightly different algorithms, or it might be from testing varience. (I did test them weeks apart - software like Steam does change over time)

I'm happy. Now all I need is more RAM!