What have you done with a PowerPC today?

[davisdelo]

I found out today that you can control a ROKU via their developer site. So if you have video capture, like an AJA, you can control your streaming device locally from your Mac. Feels so much more native, definitely the best way to stream HD youtube/netflix/etc. I like using it with BTV Pro or just plain old QuickTime. One of the best features is you can search via the keyboard input and not have to peck around for letters like you do with a regular remote.

 

[mmphosis]

I got busy today and yesterday. Today, I plugged my G4 Mac Mini into my 3rd Monitor as a single monitor connected to the PowerPC Mac. It's the one on the right with the tree frog.

I normally run the Mac headless and using Screen Sharing and Remmina from Linux. I've had an issue where the AppleVNCServer process goes to 100% if I move the mouse away from Remmina for a bit. I unplugged the 3rd monitor from my Linux PC and I will continue using the PowerPC Mac with a monitor and a mouse. There are some old Applications that just don't cut it using other platforms.

 

[Certificate of Excellence]

The long and the short of it is that I screwed something up on my quicksilver & MDD that broke TenFiveTube. As we all know YT is a fatboy and sluggish in a box. We have awesome apps like PPCMC but that still leaves having to source the vid url. To source said urls, I was using iteroni.com so ended up building a Fluid instance of Invidious. Pretty snappy and I can quickly either watch or grab urls for PPCMC. It's not YT but patches the issue until I can figure out (if I ever figure it out LOL) what I did to cause tenfivetube to bork on this box :0/

Dual 867mhz 1mb L3 per cpu MDD
2gb DDR
64gb SSD
Radeon 9000 Pro
Airport
USB2/FW400 PCI
Leopard 10.5.8

I used MUA:

Mozilla/5.0 (iPhone; CPU iPhone OS 10_3_1 like Mac OS X) AppleWebKit/603.1.30 (KHTML, like Gecko) Version/10.0 Mobile/14E304 Safari/602.1

 

[MarkC426]

I got busy today and yesterday. Today, I plugged my G4 Mac Mini into my 3rd Monitor as a single monitor connected to the PowerPC Mac. It's the one on the right with the tree frog.
View attachment 2144844

I normally run the Mac headless and using Screen Sharing and Remmina from Linux. I've had an issue where the AppleVNCServer process goes to 100% if I move the mouse away from Remmina for a bit. I unplugged the 3rd monitor from my Linux PC and I will continue using the PowerPC Mac with a monitor and a mouse. There are some old Applications that just don't cut it using other platforms.

How the heck do you get a laptop to display in portrait.....

 

[mmphosis]

How the heck do you get a laptop to display in portrait.....

Linux. I do see a Rotate option in Mac OS X as well. Linux has a lot of flexibility although I find it aggravating to use. When I have all 3 monitors going on the laptop: laptop + 4K + 1280x1024, I set the 3rd monitor to 2x scaling which matches it up to the 4K monitor. I've also rotated the camera using software using a lot of ffmpeg incantations from the command line to add a rotated camera.

 

[MarkC426]

Linux. I do see a Rotate option in Mac OS X as well.

I am aware of this option, didn't realise it was possible on a laptop...

 

[mmphosis]

It's software

I am aware of this option, didn't realise it was possible on a laptop...

Anything is possible using software. I was recently surprised to see Scale which is not available on other platforms. Using scale x2, I can turn that old 1280x1024 monitor into 2560x2048 even if it doesn't look very good.

 

[Amethyst1]

How the heck do you get a laptop to display in portrait.....

There’s an option for that in Tiger and up. Works with internal LCDs too. Been there, done that

I was recently surprised to see Scale which is not available on other platforms. Using scale x2, I can turn that old 1280x1024 monitor into 2560x2048 even if it doesn't look very good.

You can do this on macOS as well. See e.g. here. This feature was introduced for “Retina Displays”. The larger-than-native HiDPI modes make use of it.

 

[MarkC426]

There’s an option for that in Tiger and up. Works with internal LCDs too. Been there, done that

Would this work on my G4 PB.....?
Not that I would have a use....😜

 

[ToniCH]

Today I installed 16GB of new RAM into my G5 2.0GHz DC 2005 + 250GB mSata drive with a DeLock no name chinese mSata to SATA -adapter for boot and sw disk. Then I cloned the Sorbet Leopard -system from the spinner to it and installed some software including Adobe CS2, benchmarks and cloner software. I also installed a wifi card with a PCIe -adapter but for some reason that does not work.

I ran a speedtest, can't remember my old results but it seems promising. But I guess I need to compare it to others. At least Adobe stuff seems to start quickly now.

 

[Amethyst1]

Would this work on my G4 PB.....?
Not that I would have a use....😜

If it’s a 12” or 17” with an NVIDIA GPU, I don’t think so. If it’s one with an ATI, it should work. Try holding [Alt] while opening System Preferences > Displays and see if the Rotation option shows up. If that doesn’t work, use the ATI Displays tool to rotate the screen.

But I guess I need to compare it to others.

You also need to use something more reliable than Xbench

 

[mmphosis]

There’s an option for that in Tiger and up. Works with internal LCDs too. Been there, done that


You can do this on macOS as well. See e.g. here. This feature was introduced for “Retina Displays”. The larger-than-native HiDPI modes make use of it.

Oh yeah, the newer macOS's with Retina displays and regular displays use the term "Scaling." The Mac nicely says: UI Looks like 2304x1440 at 2304x1440 resolution. Can you set the display resolution beyond the maximum resolution of the connected monitor? On the Linux PC, I can set the UI beyond the size of the actual monitor resolution. I just tried cranking it up to 3x the resolution of the 4K monitor, so if Linux could state it the same way: UI Looks like 11520x6480 at 3840x2160 resolution.

BTW, I tried 4x and it didn't work, and even 3x was getting sluggish.

---

Today, I've been attempting to add a Linux PC as Bluetooth Audio to the PowerPC Mac. I hate Bluetooth!

In my opinion, Bluetooth is a horrible protocol. Anyways, when it does manage to connect, the Linux PC acts as an Audio mixer and I can hear the PowerPC Mac through louder speakers.

 

[Amethyst1]

Can you set the display resolution beyond the maximum resolution of the connected monitor?

Yes. On my system, using 3840×2560 monitors:

The Mac nicely says: UI Looks like 2304x1440 at 2304x1440 resolution.

Which isn’t very useful because it doesn’t say anything about the monitor’s actual resolution as soon as scaled or HiDPI modes come into play. On my system, it says "Resolution: 5120×3414" and "UI looks like 2560×1707". Very well, but... no mention of the fact my monitors run at 3840×2560 anywhere.

 

[ToniCH]

You also need to use something more reliable than Xbench

Like what? Is it a secret?

 

[Amethyst1]

Like what? Is it a secret?

For CPU and RAM: Geekbench 2
For storage: QuickBench 4.0

 

[ToniCH]

For CPU and RAM: Geekbench 2
For storage: QuickBench 4.0

Ok, I recognize the QuickBench screen, I think I have that that one. Where do I find data for comparison in case of the G5? Or what is the baseline for stock spinner or generic SSD in a G5?

 

[Amethyst1]

Where do I find data for comparison in case of the G5?

Here are a few.

Or what is the baseline for stock spinner or generic SSD in a G5?

There is no baseline in QuickBench.

 

[ToniCH]

Here are a few.

And the thread is mostly Xbench. Are the Xbench results not comparable to each other? Totally random or otherwise unreliable?

By baseline I meant what is the normal result in Quickbench when using the stock spinner or some generic SSD? Something for me to compare against.

 

[Amethyst1]

Are the Xbench results not comparable to each other? Totally random or otherwise unreliable?

Unreliable. As in, one of the results in this thread is 102 MB/s... over ATA-100 which has a theoretical limit of 100 MB/s. That's just plain impossible.

 

[ToniCH]

Unreliable. As in, one of the results in this thread is 102 MB/s... over ATA-100 which has a theoretical limit of 100 MB/s. That's just plain impossible.

Theoretical vs empirical

So the 100 MB/s is an absolute limit? The engineers somehow managed to design a bus cabable to transfer excatly 100MB/s, no possibility to go even slightly more or less? That would be a first. And even if they did a 2% error makes the sw totally unreliable? 🧐

 

[Amethyst1]

Theoretical vs empirical

You're absolutely right, but due to overhead, the empirical results will be lower than the theoretical ones. So you won't even reach 100 MB/s over ATA-100, making 102 MB/s even more unrealistic.

By baseline I meant what is the normal result in Quickbench when using the stock spinner or some generic SSD? Something for me to compare against.

#26 is one I found.

 

[TheShortTimer]

How the heck do you get a laptop to display in portrait.....

I am aware of this option, didn't realise it was possible on a laptop...

I've done this with my 2010 MBA and an external display so that I could enjoy arcade games with vertical monitors as they'd been intended. First in clamshell mode and later by just switching the white menu bar.

There’s an option for that in Tiger and up.

At some point soon I'm aiming to try this out in Tiger with my iBook G4 using the Screen Spanning Doctor hack to enable the extended desktop mode - which is available in the GPU but Apple very helpfully disabled it on the cheaper machines in order to push consumers towards their more expensive range.

 

[ToniCH]

You're absolutely right, but due to overhead, the empirical results will be lower than the theoretical ones. So you won't even reach 100 MB/s over ATA-100, making 102 MB/s even more unrealistic.


#26 is one I found.

But, then again, sometimes naming conventions are that they simply just round up the number to the closest meaningful round number - up or down. If they in testing managed to get max speed of 105.78MB/s they surely would not name it ATA105.78 but something simpler. Don't know how they did it in this case though. You are most likely right.

Oh, well - this is not important really, benchmarks are just fun side show. Just wondering how the coder of the Xbench got his/hers calculation so wrong, it should be pretty simple math.

Thanks for the tips anyway. I'll give the another benchmark a test. BTW. how it was determined that its results are accurate?

 

[ToniCH]

Tried to investigate and make the wifi-cards work but no success. The BCM94322MC (or its adapter) and the MT USB dongle Wifi are visible in system profiler but no luck to get them to show in the Network devices... It seems that in some machines the 322MC works right away with Leopard but in some others it doesn't. Need to find the 321MC which should be easier to get to work.

I then run some benchmarks over and over and took averages which I then compared. Xbench and Quickbench results. are mostly identical but some result are wildly different. Don't know which is correct. Quickbench results vary more between tests but seem more logical. Xbench produces more closely same results in every run but the 4k writes are suspect. Don't know if the test methodology is different or if there is a error in the result calculation?

I think next I will install a spinner so I can see what advantage going to mSata -drive had.

BTW. mSata in the internal SATA-bus with an adapter beats easily a OWC 6G 1TB SSD in OWC Firewire box and FW800 bus. The OWC disk pretty much tops at less than 75MB/s averages being in the 36-53 MB/s region. I had not before benchmarked SSD in FW bus and was quite surpriced by this. Expected it to be a bit better. But, its quite obvious the FW bus limits the top speed.

 

[Amethyst1]

I think next I will install a spinner so I can see what advantage going to mSata -drive had.

QuickBench/Xbench don't measure random access time, which is where SSDs totally obliterate any spinner. So they're not painting the full picture.

The OWC disk pretty much tops at less than 75MB/s averages being in the 36-53 MB/s region. I had not before benchmarked SSD in FW bus and was quite surpriced by this. Expected it to be a bit better. But, its quite obvious the FW bus limits the top speed.

FireWire 800 is 786 Mbit/s max (so they're rounding up here) which translates to 78 MB/s due to 8b/10b encoding. So 75 MB/s max is quite good, but the averages are quite low for an SSD.