SSD upgrade for iBook G4

[Cox Orange]

My wishlist of test to be run:

(of course, I don't expect someone to do this)

1. Could you for me copy a big file or a big folder (like 8-20GB) once with the HDD and once with the SSD installed?

2. Can you also open Terminal and watch page-outs, while opening several applications at the same time and copying a file?

3. I guess you won't have a Wattmeter or something alike, but if you have I would also be interested in power consumption change.

4. Then use iMovie to convert a video with h.264. This is only to do something that takes long (stop the encode after 1 hour). Would you write down the temperatures every 10 minutes for 1h for both the HDD and the SSD?

5. Also start Firefox and open 15 tabs with pressing cmd+T and secondly do the same but also always hit a bookmark in your browser for every tab, so that it loads something. Compare your HDD and SSD.

6. If something comes to your mind, that you do very often, please compare that too.

Of course other people than me wwould write a script for all that, but I don't because I can't and if I could, I was to suspicious, because in real world you do not use a script but do these things manually.

I don't trust benchmarks, sorry. I also think we all may have different results, because of the adapter and the SSD used. But then again I argue, that you can't trust a benchmark that tells you your SSD gets you 120MB/s on an SATA-PCI card in a PowerMac G4, where the PCI-X bus at best provides 65MB/s (Result taken from Japamacs interface comparison benchmark). Another user here in the forum used xbench and had 700MB/s, if I interpret that right. Can someone explain to me how that is possible on an ATA-100 bus (or even with a SATA-III bus)! I can only asume I read the numbers wrong.
Here, last post https://forums.macrumors.com/threads/1722496/

As a side note about the G5 (I know that is not what we discuss here, but...):
Altemose said, that in his G5 the SSD did poorly on the SATA-I and that with a SATA-II or -III card it will be better. There is a guy called flyproductions over at cubeuser.de that tried it with an SATA-III PCIe card (as I remember it first of involves some work getting the card itself work), he did get marginal better access times and throuput than with his 1TB Samsung HDD. With SATA-I and II it was even worse than the Samsung HDD.

My own SSD story:
I tried to do the best, when choosing my SSD, but it didn't work out perfectly, it seems.

a) choose SSD with best Garbage collection, because TRIM doesn't work on PowerPC Macs. Hm, my subjective feeling the more I was using this, was, as I already told, that it is getting slower and slower. GB not working as good as expected for this particular dirve? Or is it like some other user here (see some posts above) that just to many things from the OS 10.4 and 10.5 come into account.
b) save energy. When I blackend out the screen, the SSD had a 10W higher consumption than the 160GB HDD. All, the 256GB SSD, the 40GB 4200rpm stock HDD and the 160GB 5400rpm HDD ahd similar power consumptions (differing in 1W, which I would call measurement inaccurancy), despite the scenario with the blackened screen, where the SSD was worse.
c) heat. Well, it was hotter than the HDDs.
d) silence. The only reason why I still think of it, whether I should make this a permanent swap for my main ibook (tested in another ibook).
e) faster. Not that I felt it. Yes, folders and apps open quicker. Boot time was not noticablly faster, when I didn't use a stopwatch.

Can someone tell me in what scenarios you really notice that it is faster, other than boot time and opening apps or folders? Please, really interested, not asking this to make a point or even offend someone! Even if it would be theoretical for me, since I don't use Photoshop really - if someone would tell me, he works a lot with PS and it got better and that it took x seconds to load the pictures or templates over x seconds before - I would still be interested in it.

---
The connector/adapter:
As I said, I have the second adapter you posted. In another SSD thread (I guess about a PowerBook, someone said, that there is a Marvell Manhattan chipset that itself is SATA-II-to-IDE and makes better use of the SSD and is more stable and that my adapter and the one Intell recommends is SATA-I-to-IDE and that these two would cause problems from time to time.
I did only find this one (post #13 onwards) https://forums.macrumors.com/threads/1722496/ (I am not sure anymore, if that was the one. As it seems there are three different chips used).

Mine (OP's second link) has the chipset JMicron JM20330
I have an SATA-IDE adapter with the Marvell 88SA8040 (40 seems to be the 3,5" sign) chipset, similar to what is recommended in the thread I linked above and that works great in an PowerMac G4 AGP

 

[reco2011]

Can someone tell me in what scenarios you really notice that it is faster, other than boot time and opening apps or folders? Please, really interested, not asking this to make a point or even offend someone! Even if it would be theoretical for me, since I don't use Photoshop really - if someone would tell me, he works a lot with PS and it got better and that it took x seconds to load the pictures or templates over x seconds before - I would still be interested in it.

They tend to really shine when it comes to random disk I/O. Booting entails a lot of random disk I/O. So does loading an application. That's why you see improvements in these numbers. And, given they're things almost everyone does, that's why you see people refer to them (typically with statements such as "feels faster"). Most anything that performs random disk I/O is a good candidate for an SSD. I have a Sun x4600 storage server and the log files utilize SSDs (of only 20GB in size). Databases can also benefit from SSDs. Compiling software can too.

Sequential I/O also benefits but not to the same degree as random disk I/O as mechanical disks have decent throughput for sequential I/O. Not as good as SSDs but the difference is not as vast as that for random I/O.

With that said your typical users isn't likely to be performing anything with a lot of random I/O. As stated earlier startup / shutdown / app loads are the most likely things they'll be doing. That's why you continually see these referenced. The same can be said of sequential I/O...the typical user isn't doing anything that requires high sequential I/O either. Copying files may be it (and may be turned into random I/O if many files are being copied at once).

Given all of this, for the typical user, I have a difficult time recommending the installation in an older system given a new one could be had for slightly more than a decent capacity SSD. But with prices having fallen a decent capacity SSD can be had for a reasonable price (I just bought a Corsair CSSD-F180GBGT) for $55. 180GB is a reasonable size for the typical user (just a few years ago 60GB SSDs couldn't be had for $55 and, IMO, 60GB is too small even for the typical user).

In the end everyone has to do what they think is best for them. If installing an SSD in an old system is what they think is right so be it. I personally don't feel the improved boot / shutdown / app load times are worth it given the age of the system. But then, despite liking PPC systems, I don't feel using most PPC systems makes sense for most people.

----------

About 10 people who have more experience on this told you that you quite obviously don't. Perhaps it's time to limit your internet time and get some fresh air. You continue to miss all the points that people make, and flip them upside down. The thread doesn't need to go anywhere else, the question was answered, and I my SSD should come in today. Just for you I'll make a video comparing speeds, benchmarks and all to show how noticeable the improvement is over the stock 4200 RPM HDD. They say seeing is believing, huh?

No one has told me anything I don't already know. I've been down this path I don't know how many times. I've installed a number of SSDs in older systems for the very reason you're considering doing so: Because it's supposed to make them faster. And you know what? It makes them boot faster (IMO not worthy of optimization given the infrequency it is done), it makes them shut down faster (IMO not worthy of optimization given the infrequency it is done), and it makes applications launch a little faster (IMO not worthy of optimization considering most applications take a few seconds to start and launching applications isn't a frequent thing). There's no doubt these times improve. But when it is all said and done you still have a slow machine which now boots a little faster, shuts down a little faster, and opens applications a little faster. For all the people I've installed an SSD they were initially impressed with the "feel" of the system. And then you know what happened? After a period of time they came back and said "My computer is still slow". Do you want to guess at what the next step was to fix it? They bought a new one.

As to your specific situation I remind you once again: I did not enter this conversation in response to you. I responded to someone else about my thoughts on their advice. A perfectly reasonable thing to do. Ironically people say I'm off topic because my experience isn't directly related to an iBook or PowerBook. But the very experience I do have is with a system almost identical to the person I responded to. Somehow his advice is "on topic" and accurate despite it being based off of a Youtube video of a system almost identical to one which I have direct experience with (photos attached). Want to know how fast my T40 system takes to boot with a mechanical hard disk? 30 seconds from the time the BIOS screen disappears until a useable desktop. To me that's not a time worthy of improvement. Will an SSD make it boot faster? Certainly. Does it make sense to spend money to reduce 30 seconds down to 10 seconds given how infrequently a system is booted? Not in my opinion.

As for benchmarks benchmarks aren't real world. Take GeekBench for example. The 64-bit multiprocessor score for my quad core 2010 Mac Pro @ 2.8GHz is 8410. The same score for my 2012 rMBP @ 2.7GHz is 12649. The rMBP is 50% faster in this particular GeekBench score. In real world usage with Handbrake transcoding a ~ 4GB file the Mac Pro did so at a rate of 36.49 fps (21.4 minutes) and the rMBP did so at a rate of 28.28 fps (27.7 minutes). According to Geekbench the rMBP should have performed this task in 50% less time instead of performing the task 22% slower. This real world test shows just how misleading benchmarks can be.

And finally no where did I say you, or anyone else, shouldn't install an SSD in such an old system. I merely expressed my opinion, based on experience, it's not worth doing so. Furthermore I never said one wouldn't see any gains. To the contrary I said they'd be present but, for the typical user, they're insufficient and focus on infrequently performed tasks which already don't take a significant amount of time.

 

[jrsx]

Hmmm reco2011...I don't see the name iBook on that computer. OH, MY BAD. It isn't. It's a T40. Very different than an iBook. You're still missing every point I've made. I don't know if it's purposeful to annoy me, or if you seriously can't see what I'm saying. That isn't an iBook or PB. Thus the results will be different. So you mean to tell me that a Lamborghini is the same as a Ferrari? Or an F-22 is the same as an F-35? Nope, sorry. Not buying it.

 

[reco2011]

Hmmm reco2011...I don't see the name iBook on that computer. OH, MY BAD. It isn't. It's a T40. Very different than an iBook. You're still missing every point I've made. I don't know if it's purposeful to annoy me, or if you seriously can't see what I'm saying. That isn't an iBook or PB. Thus the results will be different. So you mean to tell me that a Lamborghini is the same as a Ferrari? Or an F-22 is the same as an F-35? Nope, sorry. Not buying it.

But it's almost identical to the system linked to by GraniteTheWolf (see post #46). It was his post which I responded to when I entered the discussion, not yours.

Why don't you just bow out? You got the answer to your question. There's no more reason for you to participate in the discussion unless you want to talk about SSD performance which was raised by mseth in post #14. If you want to discussion SSD performance then stop insisting it has to apply to your situation.

 

[jrsx]

Except this is my thread. In case you didn't notice.
Screw firewars

 

[reco2011]

Except this is my thread. In case you didn't notice.
Screw firewars

It ceased being "your thread" at post #13. It then became a general discussion in post #14 when mseth wrote:

"How much of a speed difference does it really make?"

Given you've already made your decision and posted as much in post #13 everything past that post would be considered off topic. However there are people who wish to continue the discussion started in post #14. I fail to see that as a problem. If you have an issue with that and, given you've already received answers to your question and made a decision, why not just ignore those who want to carry on this discussion?

 

[jrsx]

Ok, if you say this isn't my thread anymore... every post since 35 reported as it is an unnecessary firewar that does not need to be carried on and that does not entirely pertain to the thread, as you yourself stated.

 

[maflynn]

[MOD NOTE]
This thread is off the rails - closing.