[jmiddel]

My first gen iPad is beyond sluggish on WiFi. Just now I did a test: my iPhone 4S is sitting next to the iPad. Both show my network enabled in settings, and have the max of 3 bars. Using speedtest.net, the iPhone clocked in at 24.4 Mbps. The iPad is still looking for the nearest server, 15 minutes later. It is connected, I just switched to Safari and it opened a page that it was previously stuck on for a while.

Have folks noticed a difference between performance on 1st and 2nd gen iPads? Anyone else have these problems?

[ghporter]

There's something wrong, and it isn't your iPad's WiFi system. My wife has a first gen iPad and I have an iPad 2; we have relatively comparable performance in downloading pages. This is with both 2.4GHz and 5GHz WiFi.

In the iPad's Settings, turn off WiFi and then restart the iPad. Once it's restarted, try the test again. And probably more importantly, make sure you aren't running 427 apps at the same time on the iPad-it's easier to get a lot of them running and forget about them than with an iPhone. My iPhone 4 (not S) will get absolutely sluggish when too many apps are eating up too much memory, just like my iPad 2, but I'm usually more aware of what's running on the phone.

[Spheric Harlot]

Originally Posted by ghporter 

And probably more importantly, make sure you aren't running 427 apps at the same time on the iPad-it's easier to get a lot of them running and forget about them than with an iPhone. My iPhone 4 (not S) will get absolutely sluggish when too many apps are eating up too much memory, just like my iPad 2, but I'm usually more aware of what's running on the phone.

That's rubbish, if I may be so rude.

Fraser Speirs - Blog - Misconceptions About iOS�Multitasking

And the follow-up:
Fraser Speirs - Blog - iOS Multitasking in�Detail

If I can summarise my point: killing apps manually is fine as a troubleshooting step but it shouldn't be part of your daily routine.

[Big Mac]

Nope, the original iPad is not generally awful in any respect, although I don't know how good it is at running iOS 5 since we haven't upgraded ours.

[ghporter]

Originally Posted by Spheric Harlot 

That's rubbish, if I may be so rude.

Fraser Speirs - Blog - Misconceptions About iOS�Multitasking

And the follow-up:
Fraser Speirs - Blog - iOS Multitasking in�Detail

My experience has been otherwise. Past some number of apps using some amount of resources, both my phone and my iPad have been sluggish. I've gotten to the point with my phone that apps won't launch, and closing the dozens of open apps clears that right up. Maybe my surmise about the cause is faulty, but that has indeed been what I've seen.

Also, one thing Speirs doesn't address is the amount of resources required to save the execution state of inactive apps, which appears to be variable among different apps. So whether it is because I have dozens of apps "running" or merely that those dozens of apps are inactive but hogging memory with saved execution states, it still seems that an iOS device has a finite limit of such apps before it starts to have problems.

[andi*pandi]

my ipad is fine with wireless for most things. Watching streaming tv sometimes it hangs, but that's likely my ISP being a jackass.

[Cold Warrior]

jmiddel,
Try changing your wifi channel. I was having horrible wifi problems with my 2nd gen apple tv; all other devices were working as normal. I changed my wifi channel and the appletv wifi returned to normal.

[simonjames]

My first gen iPad is fine over wifi though it is hungry on the battery

[msuper69]

Originally Posted by ghporter 

My experience has been otherwise. Past some number of apps using some amount of resources, both my phone and my iPad have been sluggish. I've gotten to the point with my phone that apps won't launch, and closing the dozens of open apps clears that right up. Maybe my surmise about the cause is faulty, but that has indeed been what I've seen.

Also, one thing Speirs doesn't address is the amount of resources required to save the execution state of inactive apps, which appears to be variable among different apps. So whether it is because I have dozens of apps "running" or merely that those dozens of apps are inactive but hogging memory with saved execution states, it still seems that an iOS device has a finite limit of such apps before it starts to have problems.

Apps save their execution state to the iPad's SSD (assuming they have been written correctly, unlike the first version of The Daily, which did not save it's state). When they are not running (i.e. in the background and are not multi-tasking aware), they don't use up RAM.

[Spheric Harlot]

Originally Posted by ghporter 

Also, one thing Speirs doesn't address is the amount of resources required to save the execution state of inactive apps, which appears to be variable among different apps. So whether it is because I have dozens of apps "running" or merely that those dozens of apps are inactive but hogging memory with saved execution states, it still seems that an iOS device has a finite limit of such apps before it starts to have problems.

That's irrelevant, because, as Speirs demonstrates in the video, once inactive, they are immediately and completely kicked out of RAM as soon as resources are required.

[Big Mac]

There is no question my 3GS feels more sluggish when a lot of apps are open. Perhaps it's all placebo, but if I have more than say, six open, I can feel the foreground app perform more poorly.

[ghporter]

I don't care if they are out of RAM, the OS still must keep track of where the state is; more apps, whether inactive or not, MUST use more resources than fewer apps. And as Big Mac notes, I'm not the only one who sees my iOS devices behaving differently when a lot of apps are in the bar.

At least I'm glad to see that the "obvious stuff" has been handled. But I have other thoughts too-does the OS poll the currently inactive apps? Does it need to check states on any of them? Does it already know which ones need attention even when inactive, or does it have to check each one? A multitasking OS, even with the limitations iOS is built with, still needs to do more housekeeping when more processes are in play (and maybe even extra housekeeping when some processes may be inactive and others active).

[Spheric Harlot]

You haven't watched the video at the second link.

"Inactive" means, literally, "NOT active". There is no "polling", and no need for "attention", because the app is not running. At all. The state is saved to the flash drive, and it is completely irrelevant to the use of system resources whether it's still resident in memory or not.

You're mistaking "inactive" for "background", which allows for certain activity, within very tightly stipulated limits, or only for a very limited time (10 minutes, maximum). The messages you see from things like WhatsApp, eBay, or Facebook, are push messages sent through the mothership, and generally have absolutely nought to do with whether an application is running or not (stock phone, voicemail, email, and SMS excluded).

[ghporter]

Sorry I didn't clarify that my additional questions were more rhetorical than I did. My OS structure background is quite out of date, though the principles are still the same. I probably should have left off those questions.

However, I'm not mistaking anything in my point: the OS MUST have some way of keeping track of the apps that have been opened, no matter where their state data is stored. This takes some amount of resources. The more of these pointers/addresses/aliases/scribbled notes, the more resources are needed. Such resources are finite in any real device, so having more things to keep track of requires more resources and eventually the functioning of the OS can be impacted, even if only slightly, as in "feeling laggy" or (as I have experienced) abruptly exiting an app on launch.

Back to the OP's issue, that his iPad wasn't very good at WiFi. I'd like to hear if the base station channel change fixed it, or if it is still giving him fits. On reflection, even "feeling laggy" shouldn't have much of an effect on the throughput of the WiFi system, just (perhaps) how quickly it connects.

[Spheric Harlot]

Originally Posted by ghporter 

However, I'm not mistaking anything in my point: the OS MUST have some way of keeping track of the apps that have been opened, no matter where their state data is stored. This takes some amount of resources. The more of these pointers/addresses/aliases/scribbled notes, the more resources are needed. Such resources are finite in any real device, so having more things to keep track of requires more resources and eventually the functioning of the OS can be impacted, even if only slightly, as in "feeling laggy" or (as I have experienced) abruptly exiting an app on launch.

Um, I'm not too deeply involved with the internals beyond what Freirs shows in his video and explains on the other link (have you watched the video?), but how many "resources" do you think a .plist text file listing recently used apps consumes, and why should it slow down the OS? As for address space in RAM: We're talking about devices with power equivalent to G3 iMacs or G4 Powerbooks. Keeping a list of allocated RAM shouldn't really take more than a kilobyte or so of system memory, no?

It's certainly possible that an app with nasty bugs or a botched state-reporting implementation might be able to hog system resources, but the points you bring up strike me as being quite absurd even to a (relative) layman.

IANAD.

[jmiddel]

Cold Warrior, thank you. I did change the channel, and things are better. Just wanted to give it some time to verify, since the problem was intermittent. The channel was Automatic, I changed it to 4, so why is auto not be better? Isn't it supposed to select the best channel? Are there some channels that are least used, although automatic which comes as default should use them all equally?

[Cold Warrior]

I don't know what logic goes into the automatic selection, but I've never known it to be perfect. With wifi problems, it can be useful to view nearby channels and switch to one farthest away from any others.

[Big Mac]

Originally Posted by ghporter 

However, I'm not mistaking anything in my point: the OS MUST have some way of keeping track of the apps that have been opened, no matter where their state data is stored. This takes some amount of resources.

You should know that the point SH is driving at is that iOS does not give multitasking time to background apps generally speaking. There are exceptions for audio and push notifications, but most of the time iOS background apps are asleep. You can see this by opening an app that frequently updates its display. When you put it in the background and then go back to it later, you'll often see that for a second or more the screen hasn't updated and looks just as you left it. The app was suspended. That's a big reason why iOS devices enjoy superior battery life as compared to Android.

But I agree that background apps take some sort of system resources. I've seen lag in games when I have a lot of background apps open, and when I close them the game stops lagging. Again, this is on my 3GS. I'd be very surprised to find out I'm mistaken in this area.

[Spheric Harlot]

Originally Posted by Big Mac 

You should know that the point SH is driving at is that iOS does not give multitasking time to background apps generally speaking. There are exceptions for audio and push notifications, but most of the time iOS background apps are asleep. You can see this by opening an app that frequently updates its display. When you put it in the background and then go back to it later, you'll often see that for a second or more the screen hasn't updated and looks just as you left it. The app was suspended. That's a big reason why iOS devices enjoy superior battery life as compared to Android.

Thank you. (Even push notifications, AFAIK, do nothing to the status of a backgrounded app. They're just messages pushed from Apple's servers; AFAIK, they are not able to re-launch inactive apps on their own.)

Allow me to repeat, again:

WATCH THE FRASER SPEIRS VIDEO.

Here:
Fraser Speirs - Blog - iOS Multitasking in�Detail

[ghporter]

I will set aside time later today to watch the video and re-read the text in detail (I'm getting ready for work at the moment). I'm not referring to battery power or clock cycles, but something, even if it is only .plist text, is there to let the OS track what's suspended. Using the .plist analogy, the more apps in play (suspended, active, background, etc.) the longer the .plist file, and the more work the OS has to do to keep up with everything. And when an app is invoked, the OS needs to see whether it's already suspended or if it needs to be loaded, so it has to look through that .plist to find that out-which is simply more overhead processing. That's all. Whether that's enough to cause what I and others have experienced or not, it IS a valid (potential) explanation for the effects being seen.

[Big Mac]

If it's just a plist tracking recent apps, that's not going to cause any user-perceptible time penalty to deal with, Glenn. But I'm not convinced that's all that's involved. I think finite RAM is also very much involved. The background apps go to sleep and don't take any processor time, but they have to be shuffled in and out of RAM.

[ghporter]

...until something is invoked, when the OS would have to run through it to find out whether the requested app is already there. Not a big drain, but it's something. And it would explain why we see slow downs when an app is launched (which is the only time I see them).

[Spheric Harlot]

Originally Posted by Big Mac 

If it's just a plist tracking recent apps, that's not going to cause any user-perceptible time penalty to deal with, Glenn. But I'm not convinced that's all that's involved. I think finite RAM is also very much involved. The background apps go to sleep and don't take any processor time, but they have to be shuffled in and out of RAM.

THAT, I think, is the key.

Keeping and parsing a list of available apps amounts to looking through a couple hundred lines of text. If that's gonna impose a performance penalty, well, welcome to 1976.

But paging out the active app to flash memory, and loading the saved-state application back into RAM from flash when switching apps: That is going to cause a delay.

[Spheric Harlot]

Originally Posted by ghporter 

we see slow downs when an app is launched (which is the only time I see them).

You didn't mention that earlier. You were talking about general performance issues when many apps are loaded (which they aren't; they're just shown in the switcher).

That, of course, makes perfect sense. But actually going into the switcher and "quitting" all those apps makes not one iota of difference.

[jmiddel]

Back to my original pots: Cold Warrior suggested changing channel, which worked for 2 days. An hour ago, I was streaming a 50 minute lecture (off iTunes U), and WiFi just stopped working twice. I had to turn it off and back on to get it going again. Very frustrating. I'll change channels again and post back. AarrgghhhH!

[ghporter]

Originally Posted by Spheric Harlot 

You didn't mention that earlier. You were talking about general performance issues when many apps are loaded (which they aren't; they're just shown in the switcher).

That, of course, makes perfect sense. But actually going into the switcher and "quitting" all those apps makes not one iota of difference.

I went back and checked-I completely failed to mention that point.

However, when it happens, if I retry there is no change in behavior, but if I go to Switcher and quit a bunch of apps, the device DOES respond more quickly. Typically this is with a dozen or more suspended apps.

[Big Mac]

Originally Posted by Spheric Harlot 

THAT, I think, is the key.

Keeping and parsing a list of available apps amounts to looking through a couple hundred lines of text. If that's gonna impose a performance penalty, well, welcome to 1976.

But paging out the active app to flash memory, and loading the saved-state application back into RAM from flash when switching apps: That is going to cause a delay.

Thank you, SH. It also makes complete sense that RAM is a huge factor when you look at an app that displays iOS RAM usage. I'm amazed my 3GS can get out from under its own weight given how little free RAM is left to apps, and many have commented that the original iPad was also under-RAMed (but again I think the original iPad's performance, at least under iOS 4, is great).