Episode 1: FFS vs PFS3 on YT

[CmdrAmiga]

Hello folks!

I'd like to inform you about a new YouTube channel where my colleague and I will be publishing some interesting Amiga hardware and software tests. Like many of you, we also have some questions regarding proper and the best configuration in order to enjoy our great machine. List of episodes:

EPISODE 1: [ Show youtube player ].

EPISODE 2: [ Show youtube player ]

EPISODE 3: [ Show youtube player ]

EPISODE 4: [ Show youtube player ]

EPISODE 5: [ Show youtube player ]

EPISODE 6: [ Show youtube player ]

[mrprawns]

good content

[Neil79]

[Radertified]

It's one thing to see a graph, but to see it in action is great. I absolutely loved this and am looking forward to what you compare in future.

If I had a YouTube account, I would like and subscribe

[gulliver]

And the multitasking test?

[meynaf]

If you just took your original FFS partition and copied it to PFS3, then i'm afraid the test can be biased due to file fragmentation on the source. Both must be totally empty prior to any test. Is it the case ?

[Thomas Richter]

Originally Posted by CmdrAmiga

In our videos some myths will be busted and you won't have to rely on biased opinions anymore.

Before you "bust" something, it becomes necessary to understand the problem. This is not covered by "how much data is transfered if I write a file", because the limitation of that is easily computed: It is the transfer rate of the harddisk that limits the speed, in either file system. Things become interesting if you ask "how much CPU power does it take to adminstrate files" or "how long does it take to locate a file" or "how long does it take to list a directory". All of these are different questions, and the answers are very different depending on the file system.

PFS in general has higher CPU requirements because it has more complex data structues to administrate. FFS takes longer to list directories because it has "one administration block per file" rather than a classical directory structure with inodes. FFS is very fast to locate files because it uses hashing to find the right file header (one indirection less).

So, it is not that obvious. For lower powered machines, I would still say that FFS has an advantage due to its simplicity. For flash-based media, seek times are absent, so the directory seeking problem does not exist.

[nikosidis]

Good video!! PFS3 is so much faster. I never use anything other than PFS3. I have Blizzard 030/50 on my A1200 and Furia 020/37 on my A600. Never had a single problem either.

[CmdrAmiga]

@Tomas, no offence but exactly for that reason this video has been made. I am enough of technical data about how wonderfully some piece of software has been made. I'm an end user and I don't really care if the dev used the latest compiler or wrote it in assembly, or the structure of his program beats other products on paper. What I care is practical efficiency and that can bee seen on the video very clearly with overwhelming results (I didn't expect such a difference to be honest so I'm double happy that I made it). It wasn't meant against FFS. I was (and still am with some exceptions) a long time user of this file system. But I wanted to see how much the difference can be. I shared my work with the community but I leave the conclusions to draw for others - I have my answer.

@Gulliver, what do you mean? Specify please how such a test should be done and if it's practical and helpful to the user then I'll carry it on.

@Meynaf, both partitions have been made from scratch for this test (no other data was on them). So there was no fragmentation and other than that it's a flash type memory, there's no mechanical moving parts that could affect the speed. I hope that answers your question :-)

[klx300r]

I tried PFS3 a few times and both times I had problems after few months of usage so went back to trusty FFS and no matter which you use the only way to save anything thats important is ALWAYS back up and then make another back up

[gulliver]

@CmdrAmiga

Multitasking is a feature of AmigaOS since inception. To make it simpler for you, I will put it this way: it enables you do more than one thing at a time.

Example: copying files and listening to music or formatting a floppy. Pick two task or more and do them at the same time. There are thousands of different multitasking operations you can perform. You could have started multiple different copy instances instead of just one.

AmigaOS can multitask since day 1 but your test do not show how is that multitasking being affected with each option you chose.

[CmdrAmiga]

Dear Gulliver, I don't know whether you did not understand my question or you're just trying to ridicule me? I perfectly know what Amiga multitasking is. I asked you about a specific method to carry on such a test. What program(s) should I use and what actions to perform.

You could have started multiple different copy instances instead of just one.

Personally it is not practical test to me. I use a file manager and always perform one copy at a time. But if more users asks for it I see no problem in doing it.

[utri007]

I use FFS, not because it is better, but because it hasn't ever let me down. Generally FFS has problems often, but they those are always solvable. PFS has problems rarely, but I couldn't solve them.

[gulliver]

@CmdrAmiga

You made a question and it seems you did not understand the concept, that is why I explained it to you. I also don't know you and have no reason to mock you.

If you don't want to do more tests it is fine. But then the conclusions that can be driven from the tests are severely limited and far from practical or accurate.

[CmdrAmiga]

@Gulliver, no problem mate, all is clear now. It's not that I don't want to do more tests, I simply choose those that are as practical as possible because doing them takes time. Then it takes even more to video edit them.
Anyways, I have an idea how to run the test you suggested. But it will have to wait a bit. I'm about to finish editing another test. I think many will like it very much since it compares two 060 on two different turbo cards ;-)

[nikosidis]

Originally Posted by gulliver

@CmdrAmiga

Multitasking is a feature of AmigaOS since inception. To make it simpler for you, I will put it this way: it enables you do more than one thing at a time.

Example: copying files and listening to music or formatting a floppy. Pick two task or more and do them at the same time. There are thousands of different multitasking operations you can perform. You could have started multiple different copy instances instead of just one.

AmigaOS can multitask since day 1 but your test do not show how is that multitasking being affected with each option you chose.

That is funny cause I made a video about multitasking, Amiga and of cause I use PFS3 as always.

https://www.youtube.com/watch?v=FqcVltXeT0M&t=96s

[nikosidis]

CmdrAmiga: I agree 100% with you. Who would copy different directories or unpack different archives at the same time. I do not even like to do that on Linux or Windows. What matters for 98% of people is that the file system is realible and fast.
You can search for "10 min. retro cast" and the video where he compare different disks. DOM, Compact flash etc. He only use FFS. Only the DOM disk give you some CPU time. Other disks used 100% CPU time doing operations.
Booting the computer with FFS is like double the time of PFS3. I had lots of problems with FFS and never 1 problem with PFS3.
I for sure don't turn off my computer while copying something to disk. Most of the time I copy, unpack to RAM and test before I save to disk.

[fxgogo]

I enjoyed that video. Thanks for making it. I do not know PFS at all really as I only recent got back into my Amiga stuff. I would love a follow up video where you give us a bit of history and technical info on both file systems.

[walkero]

@CmdrAmiga
Congrats for the fine video and your intention to make more. I really like that and I am looking forward to see more of them.

From this video I missed the depth of the tests. As others said, measuring filesystem performance is has to be done with many different aspects.

Let mention some of them:
- Copy files, as you did, but it is better these to be done from RAM. I don't know were this SRC: mount is, if this is an other partition on the same SD
- When you archive/dearchive a folder, it would be nice to have a measure the same process on Ram, which will show the potential of the machine (CPU, Ram speed etc.)
- I would like to see comparison on opening a folder on WB with, let's say, 100 files. How much time will each FS to show them in text mode (no precached)
- Also, I missed CPU measures on simple tasks, i.e. while you copy from Ram to each FS partition. You can provide an average usage.
- Multitasking, as Gulliver mentioned, is necessary to be measured. For example, while you do the copy of the files, try to open a jpg image with multiview and get the time needed. You can also use an app to convert a aiff file to mp3, or something similar. This should be done on Ram, so you won't get wrong measurements because of the IDE bandwidth.

And something that is totally my opinion and feel free to ignore it if you want. I would like to listen you giving some extra information about each FS, how it works and the differences between them. Then talk about how did you get those measures, what you did to be sure that they are accurate, I mean what methodology did you follow, what are the machine specifications. And finally show some graphs based on your measurements, comparing the two FS in depth.

[CmdrAmiga]

I see many positive feedback and that's good, thanks guys. This channel and tests are done for the community so I'd like them to be as informative as possible. However, I can't do all you ask for the simple reason: I have limited time. I know it'd be nice to present all the characteristics of the file systems but it really takes a lot of time I don't have. But I don't say no. Maybe I'll find some time after I do the tests I have in plan.

@Walkero, you mentioned a lot of interesting stuff. I'll consider it in my future tests (and my colleague so maybe we split our effort and he does something I can't). But some of your questions have been already answered. For instance, the specifications of software and hardware used are given in the YT description. I know not all read them but please do so to get more info on the tests.

Another video is being uploaded to YT. Will be available in minutes...