2015 15" MacBook Pro cannot produce 5K over a single Thunderbolt 2 connection. What does the timing info in SwitchResX say, or the output of AGDCDiagnose command?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Pro Display XDR - Owners Thread
- Thread starter - rob -
- Start date
- Sort by reaction score
My personal machine is so lightly used these days, I figure I may as well trade it in for something small that I will actually use. I do need to figure out WTF is going to happen with containers though. Not sure how I can run containers targeted for x86 servers locally in Docker if its VM is ARM.
I'm more interested in how well the XDR will be driven, and how much it will make the (fanless) MBA throttle. If there's a bunch of tearing because the GPU is underpowered, that would be disappointing. And if just running the XDR makes it heat up and throttle, I might get the MBP instead since it's got a fan. Or maybe not do anything. Just interested in hearing some real world experiences before pulling the trigger.
Am I the only one that constantly feels like my XDR is slightly tilted?
I bought a tiny level to put the issue to rest in my head because I was fiddling with it so often.Yep - got the MBA on order for my non-work stuff (due Dec 1), can tell you more after that. If it struggles with the display it will go back and I’ll wait for next year. But it all sounds like it should be alright
Looking forward to hearing the review.
Do we have any consistent way to measure display performance against various machines?
I ordered a 13 inch M1 pro to run my XDR display, but now im thinking I might want to have the Air instead for the fanless design. The most intense thing I do is edit photos (XDR is amazing for that), but I also want to make sure the Air won't choke on the XDR. I'd be interested on feedback on how well the Air runs the display.
Same situation here for the most part. I spend all day on computers, mostly writing code in any number of languages and DSLs, but nothing I write needs compiling, and nothing I do produces sustained loads. I've realized that I only ever buy the top spec machines because I want apps to open and respond quickly. If Apple can deliver that now in their base models, it's gonna save me some cash in the future.
I managed to get a pickup order in for tomorrow (Australian timezone). Should be able to connect to the M1 MBA in ~24hrs. If anyone needs something specific tested, let me know ...
/System/Library/Extensions/AppleGraphicsControl.kext/Contents/MacOS/AGDCDiagnose -a > AGDCDiagnose_M1_MBA.txt 2>&1ioreg -filw0 > ioreg_M1_MBA.txt (remove your serial numbers if you like)system_profiler SPUSBDataType SPThunderboltDataType SPPCIDataType SPNVMeDataType > system_profile_M1_MBA.txtSwitchResX screenshot of timing info (does SwitchResX even work on M1? - it doesn't use kexts so it should),
Geekbench 5 iOS vs Geekbench 5 macOS,
AJA System Test Lite.app (is it Universal yet?),
Blackmagic Disk Speed Test.app (is it Universal yet?),
etc.
Lots of M1 reviews today, embargo must have lifted this morning. Panzarino's review mentions a test with the XDR, but specifically mentions that he did not get to test it with two XDRs.
We have at least one dual XDR user here, any chance one of you folks can get your hands on an M1 to test driving two?
We have at least one dual XDR user here, any chance one of you folks can get your hands on an M1 to test driving two?
Reporting back with a M1 13 inch MacBook Pro, it runs the XDR well! Much better than previous 13 inch MacBooks and almost as good as the 16 inch MacBook Pro. The only dropped frames I've seen was going into LaunchPad. The real benefit is that it runs it SILENTLY! Which is a huge benefit over the 16 inch.
That's good to hear. Fan noise running the XDR on MBPs have come up a fair amount ITT.
I'm wondering what the future holds for discrete graphics with Apple Silicon. I really think they were on to something with the BlackMagic eGPUs--they work so well.
I wonder if Apple might design their own fanless Apple Silicon eGPU that is half the size, uses a fraction of the power and delivers absolutely insane GPU benefit.
You want a battery powered eGPU? Otherwise, there's nothing wrong with a 400W GPU if it's 20 times faster than a 20W GPU.
Just connected my MBA (16GB/8C) and everything seems to be good. Played a bit in the UI and watched some 4K videos in Safari (doesn't give me more I think?). No slowdowns I could see
@joevt see attached files for the debug output. It seems that diagnose script didn't work and MacRumors doesn't like the other file (too big with 5MB?). Any part I should copy/paste out?
@joevt see attached files for the debug output. It seems that diagnose script didn't work and MacRumors doesn't like the other file (too big with 5MB?). Any part I should copy/paste out?
Thanks for the output file. With that AGDCDiagnose output, it seems the M1 Macs do graphics much differently (or at least it's not using Apple Graphics Device Policy or not using it in the usual way). My Intel Mac mini 2018 has same version of AGDCDiagnose and has 3 AGDC nodes - so it produces all the expected output in Big Sur as it does in Catalina.
For large txt files, it is best to zip them in the Finder (right click and select compress). Then you should have no problem attaching them to MacRumors post.
ah nice ... I edited my post above with the zip. Either way, really positively surprised how flawless everything works so far while getting set up
Tried this out on a new M1 Air, same result, so much so that I am actually going to step down to the Air. It's crazy how fast the M1 is.
I saw in one of the reviews that the Air doesn't start thermal throttling until 8-9 minutes of running at full pin. It's hard not to LOL about how fast these base models are. Imagine what the larger mobile and desktop parts could do.
Thanks.
There's a lot of interesting stuff in the ioreg.
Accelerometer - Is this a known feature - do iPad apps that use the accelerometer work on the M1 Macs? Also: Gyro and Compass.
Apple's two Thunderbolt/USB4 controllers have class "AppleARMIODevice" (not PCIe devices) and compatible name "acio". The driver is AppleThunderboltHALType5.
Apple's two Thunderbolt/USB4 ports have class "AppleARMIODevice" (not PCIe devices) and compatible name "apciec". The driver is IOPCIBridge:AppleEmbeddedPCIE:AppleT8103PCIeC. This is for the PCIe upstream bridge for attached Thunderbolt PCIe devices. The bridge has PCI vendor and device id of 106b:1010 (Apple).
We're up to Thunderbolt type 5 now: AppleThunderboltHALType5/AppleThunderboltNHIType5/IOThunderboltControllerType5
type 4 was Ice Lake Macs.
Type 3 is Thunderbolt 3: Alpine Ridge and Titan Ridge.
Type 2 is Thunderbolt 2.
Type 1 is Thunderbolt 1.
Silent, fanless and small is what I would want. For people rendering on location, it would be good if it was powered over the TB3 from a laptop.
The black magic eGPUs are getting listed on eBay sort of fast and furious right now.
I’ve been watching the vega pricing for months. Curious why. I had thought Apple might release an AS-based egpu, but that it would be unlikely to support x86.
I use the 580 with my XDR and won’t replace my mini until it gets the M2 / docker support. So it’s a bit tempting to look at the upgrade.
I’ve been watching the vega pricing for months. Curious why. I had thought Apple might release an AS-based egpu, but that it would be unlikely to support x86.
I use the 580 with my XDR and won’t replace my mini until it gets the M2 / docker support. So it’s a bit tempting to look at the upgrade.
Last edited:
Mine’s due next week. The Docker thing is a hinderance for sure. I really hope there’s a way to run x86 containers eventually.
I assume mini users were the biggest consumers of eGPU. Maybe they’re moving on since the new mini is cheap and capable? It’s crazy to think they could probably cut the Mini’s size in half, put a chip in there with a little higher power consumption, and dog walk the current MacPro.
I assume mini users were the biggest consumers of eGPU. Maybe they’re moving on since the new mini is cheap and capable? It’s crazy to think they could probably cut the Mini’s size in half, put a chip in there with a little higher power consumption, and dog walk the current MacPro.
So I have a 16" MBP with the 5600M, 64GB RAM, 2TB SSD. My XDR runs pretty well on it, but I can for sure sprint it quickly into the 90's for temp if I want to. Lid open on my MBP also, and as I type this post I sit in the 50s for temp.
However, I can't help but get caught up in the commotion of the M1 and the thought of an Air with no fan running this XDR more efficient than my 5600M.
I am sure some of you have crossed this road as well and were not sure if you should offload and bump down to an m1 or to hold fast. I also worry about resale once the 16's do come out with Apple chips.
Any insight appreciated.
However, I can't help but get caught up in the commotion of the M1 and the thought of an Air with no fan running this XDR more efficient than my 5600M.
I am sure some of you have crossed this road as well and were not sure if you should offload and bump down to an m1 or to hold fast. I also worry about resale once the 16's do come out with Apple chips.
Any insight appreciated.
I'd wait for someone else to take the leap first. The XDR is pushing a LOT of pixels and there's always been a gap between iGPU and dGPU macOS desktop performance. It isn't clear yet to me how much of a difference the new M1 will make here, or if it will be enough.