Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I wish we had some benchmarks for 5700. In pc benchmarks 5700 XT is 10-15% faster but the iMac upgrade from 5700 to 5700 XT is 67% more expensive. The price certainly doesn't match the extra 10-15% performance. Everybody seems to buy XT but I wonder if people will even really need it.

Ta daaaaa
 

Attachments

  • Screenshot 2020-08-16 at 13.34.43.png
    Screenshot 2020-08-16 at 13.34.43.png
    234.1 KB · Views: 257
  • Screenshot 2020-08-16 at 13.34.34.png
    Screenshot 2020-08-16 at 13.34.34.png
    693.5 KB · Views: 188
I won't pretend to be an expert, but from what little I understand is that Metal is better about leveraging Vega's compute abilities than other graphics benchmarks. Somehow Vega was a number-crunching monster while being weak in traditional graphics tasks.

Metal isn't a benchmark, it is an API. That said the Unigine benchmark uses OpenGL not Metal unlike some of the other benchmarks so that could account for some differences.
 
Which configuration did you get, Azrael? You've probably announced it multiple times, but I must have missed it.. and are you ordering you own RAM (if yes, which one?)


Thanks!

10 core. 5700XT 16 Gigs of Vram. 8 gigs. (Crucial 2x16 gigs. On its way. £128. Amazon.)

1TB SSD.

Standard ethernet.

Azrael.
 
  • Like
Reactions: macbro08
Anybody worried about gaming performance.

Any of the RDNA1s should be decent.

But that 5700XT 16 Gig..? Seriously impressed me.

Get it if gaming on the Mac or Bootcamp is your thing. Worth the extra.

Azrael.
 
10 core. 5700XT 16 Gigs of Vram. 8 gigs. (Crucial 2x16 gigs. On its way. £128. Amazon.)

1TB SSD.

Standard ethernet.

Azrael.
Thanks Azrael. Could I ask you to share the link or model number to the memory with me? I'm getting the same config as you, but am a bit unsure which RAM to get.

Thanks in advance!
 
Thanks a lot @MacRS4 for the review and for telling us about the special nanoscreen cloth (I specifically asked about it in one of my previous posts):
 
Last edited:
  • Like
Reactions: MacRS4
Mine is going back. Over the course of 3 days of ownership, I had 2 kernel panic errors and the machine froze multiple times. I guess I'm doing to wait until an ARM version comes out to upgrade my old 21.5".
 
  • Angry
Reactions: mechapreneur
If it takes a full 2 years to transition the top of the line systems, does anyone think we will see another refresh on these iMacs for Big Navi coming out soon? AMD really shocked the world with Ryzen, I am hoping they do the same with Big Navi and get to at least be on par with NVIDIA if not give them a good shakeup.
 
He talked about using SwitchResX to turn off the internal display but that's just under macOS. What about under windows? I've got a 1440p/144hz display I'd like to use when I'm gaming.

Actually I do not agree with his statement. I have a 2019 i9 iMac with a 2560x1440 display at 144hz and I can game at 144FPS and have my monitor at 144Hz without any additional tools. I also have a 4K 60hz third display connected. I get these in both macOS and Bootcamp where I can use 144Hz without issue with just macOS and Windows 10 without any additional tools and keeping the iMac screen active.
 
Oh, I completely agree with that. I don't think we'll exceed 8GB VRAM in games before other advances give us plenty of compelling reasons to upgrade. So I really don't find 8GB vs 16GB VRAM interesting in terms of future proofing.

There's also the fact that the iMac has a shared power limit between all of its components and a someone gimpy cooling system, so it's pretty unclear how much a game can push everything at once before the whole system gets throttled. I don't feel reviewers have addressed this thoroughly yet and in general the issue makes me suspect of trying to get all the highest spec and power components.

It also depends on Big Navi and the upcoming consoles and how they will treat VRAM. This is the first time consoles are getting brand new hardware before PC, because Big Navi is coming out for them first.
 
Mine is going back. Over the course of 3 days of ownership, I had 2 kernel panic errors and the machine froze multiple times. I guess I'm doing to wait until an ARM version comes out to upgrade my old 21.5".

I think this is just Catalina. I have had several kernel panics on my 2019 i9 iMac using Catalina. But 0 when running Mojave. In fact, I just had a kernel panic today while working in Final Cut Pro X. But it can go weeks without any issues. Let's hope Big Sur resolved the issue once and for all.
 
Without knowing the cause of the kernel panics, such reports are meaningless. In every case, the question must be asked: What does Apple say is the cause.

Tech Support calls must be made and crash reports must be forwared to engineers on your behalf. Reporting to Apple does not do this. Going to Support and asking for a call back does.

I've been working with Apple on these for a few months now. So far, every one we've found was caused by something old — removing it solved the problem. Sometimes, it takes Apple Engineering awhile to find the exact cause, however. This support is free even if no longer under AppleCare as long as the OS is currently supported — through October, that means High Sierra and newer. In November, 2020, OS 10.13.x will be dropped from the list.

The one that happened to me over Mojave with the March 2020 Security Update took six weeks to resolve. The culprits were an old Soundflower .kxt and another for NIGuitarRigII — both had been installed to ensure compatibility with OS 10.6/7. BTW, those same .kxt files were found on my Catalina machine but it never crashed (I still removed them, of course). Obviously, the only change upon removal was no more kernel panics and crashing. With my clients, I have a list of over 30 other causes including HP drivers from 2005 and Finale 2009 Help files...

Contrary to the many threads on the subject, the T2 chip/BridgeOS is not the cause. It's the symptom and the first thing seen in the crash report. Think of it as the canary in the coal mine.

It would be nice if the BridgeOS was more tolerant of old, obsolete and poorly written code — but it's not. Perhaps with Big Sur — beta has been promising.

Since my spouse has crap on her 2011 reaching back to the Stone Age (ok, her Performa 6200), I expect to be going through this all over again on hers. Apple will own any hardware issues for the next three years but they'll own OS issues for as long as they're still supported. I will hold them to it as I have been doing since the Mac Plus.
 
  • Like
Reactions: kazmac
Same! I tried every config of two 8GB OWC sticks and the default two 4GB sticks Apple includes, could not get anything above 2133. I tried every config. Another two 8gb sticks from OWC overnighted via Amazon has me at a cool 32GB and 2667 thankfully. A shame the Apple RAM is so finicky.
So it didn't work according to this? If you have 2x4 in Slot 1-2 and 2x8 in Slot 3-4?

 
  • Like
Reactions: mikehalloran
Well said.

That 5700XT 16 gigger?

Kicks az.

Bomb proof in terms of future gaming that needs Vram.

When it's sitting near a 2080Ti in some instances...that's just mind blowing.

Some voodoo magic to get an under clocked version out performing a higher clocked PC version in 4k gaming.

Azrael.

If they've got time, I'd like John90976 to run the 5 minute in-built benchmark of Red Dead Redemption 2 which will reveal the true framerate that his iMac runs the game at and also give us a screenshot of his resolution and graphics settings used to run the benchmark.

No doubt this GPU option gives us an amazing GPU option finally. But I can't just take his account, that the TDP limited 5700 XT in the iMac is nearly at the performance level of an RTX 2080 Ti, at face value. One person's subjective experience of how it is performing doesn't make it actually so. That this iMac 5700 XT is somehow defying all logic to beat non-power limited RTX 2070 Supers and nearly equal the RTX 2080 Ti would truly be 'voodoo magic' if the benchmarks show it
 
Last edited:
If they've got time, I'd like John90976 to run the 5 minute in-built benchmark of Red Dead Redemption 2 which will reveal the true framerate that his iMac runs the game at and also give us a screenshot of his resolution and graphics settings used to run the benchmark.

No doubt this GPU option gives us an amazing GPU option finally. But I can't just take his account that the TDP limited 5700 XT in the iMac is nearly at the performance level of an RTX 2080 Ti at face value. One person's subjective experience of how it is performing doesn't make it actually so that this iMac 5700 XT is somehow defying all logic to beat non-power limited RTX 2070 Supers and nearly equal the RTX 2080 Ti
I mean why bother, he’s obviously got the numbers wrong. Just ignore them.
 
Without knowing the cause of the kernel panics, such reports are meaningless. In every case, the question must be asked: What does Apple say is the cause.

Tech Support calls must be made and crash reports must be forwared to engineers on your behalf. Reporting to Apple does not do this. Going to Support and asking for a call back does.

I've been working with Apple on these for a few months now. So far, every one we've found was caused by something old — removing it solved the problem. Sometimes, it takes Apple Engineering awhile to find the exact cause, however. This support is free even if no longer under AppleCare as long as the OS is currently supported — through October, that means High Sierra and newer. In November, 2020, OS 10.13.x will be dropped from the list.

The one that happened to me over Mojave with the March 2020 Security Update took six weeks to resolve. The culprits were an old Soundflower .kxt and another for NIGuitarRigII — both had been installed to ensure compatibility with OS 10.6/7. BTW, those same .kxt files were found on my Catalina machine but it never crashed (I still removed them, of course). Obviously, the only change upon removal was no more kernel panics and crashing. With my clients, I have a list of over 30 other causes including HP drivers from 2005 and Finale 2009 Help files...

Contrary to the many threads on the subject, the T2 chip/BridgeOS is not the cause. It's the symptom and the first thing seen in the crash report. Think of it as the canary in the coal mine.

It would be nice if the BridgeOS was more tolerant of old, obsolete and poorly written code — but it's not. Perhaps with Big Sur — beta has been promising.

Since my spouse has crap on her 2011 reaching back to the Stone Age (ok, her Performa 6200), I expect to be going through this all over again on hers. Apple will own any hardware issues for the next three years but they'll own OS issues for as long as they're still supported. I will hold them to it as I have been doing since the Mac Plus.

I have called Apple Support when it was happening daily. They escalated it to senior engineering to fix the issue. They had stated it was an issue with Catalina as many others were reporting the same problem. It has certainly gotten better with the latest patches, but I still get one every few weeks.

When Catalina first came out, I couldn't even let my computer go to sleep without it causing kernel panics. Now I can let my system run for weeks at a time, sleeping a lot and not receive a single kernel panic for weeks.

And it is related to WindowServer and the GPU drivers. I got an eGPU specifically for preventing Kernel Panics, I had to have my eGPU always connected to my iMac. Now I haven't used my eGPU in a few months.
 
  • Like
Reactions: mikehalloran
If they've got time, I'd like John90976 to run the 5 minute in-built benchmark of Red Dead Redemption 2 which will reveal the true framerate that his iMac runs the game at and also give us a screenshot of his resolution and graphics settings used to run the benchmark.

No doubt this GPU option gives us an amazing GPU option finally. But I can't just take his account, that the TDP limited 5700 XT in the iMac is nearly at the performance level of an RTX 2080 Ti, at face value. One person's subjective experience of how it is performing doesn't make it actually so. That this iMac 5700 XT is somehow defying all logic to beat non-power limited RTX 2070 Supers and nearly equal the RTX 2080 Ti would truly be 'voodoo magic' if the benchmarks show it

I'll be happy to give you screenshots tomorrow, I got those results directly from the internal benchmarks of both RDR2 and GTA V. I would genuinely like to know as well, my house mate with a 2070 Super is quite jealous after months of telling me to jump over to a platform I'd never be happy with. Now the iMac is on top, and I'll prove it. Don't know how it's possible though, I agree it's something weird.
 
  • Like
Reactions: Azrael9
Well here is Red Dead 2, at 4K with .5 resolution scaling and Ultra texture quality. It’s running 5ps lower on the minimum FPS than my initial report (I was running native 4K with high textures instead of using scaling) but averages the same 59 FPS. In my quick research I saw that a 2080 TI got 57fps on high. They probably could’ve optimized their tests better and beaten this 5700XT PRO 16GB, but I don’t care to go test that. I have this new card that plays 4K ~60 and here is the proof. Even better, you can go get it and find out if you’re so in disbelief. I do not deny any of the facts about this card, actually, they are against it, and the power envelope is ridiculously small, and yet here are my results. Read ‘em and weep. And enjoy the iPhone shots, I hate windows and I’m not gonna go hunting around the horrid interface to simulate a windows+prntscrn key so I can take a screenshot for a forum post. This is as good as it gets.
 

Attachments

  • 52406F92-A55B-45C1-870F-3B06E5F1CE86.jpeg
    52406F92-A55B-45C1-870F-3B06E5F1CE86.jpeg
    98.3 KB · Views: 185
  • 42E454AB-5586-4076-BC90-E201CD894EE1.jpeg
    42E454AB-5586-4076-BC90-E201CD894EE1.jpeg
    443.2 KB · Views: 136
  • 6D66521F-2EB7-4C53-9899-5C0BE5666EDD.jpeg
    6D66521F-2EB7-4C53-9899-5C0BE5666EDD.jpeg
    436.6 KB · Views: 133
  • 86827956-35A0-48DB-A977-50418CB19083.jpeg
    86827956-35A0-48DB-A977-50418CB19083.jpeg
    427.5 KB · Views: 127
  • 4DDD0BBE-63DD-4110-A657-0F55347238A3.jpeg
    4DDD0BBE-63DD-4110-A657-0F55347238A3.jpeg
    431.2 KB · Views: 170
  • 1E8A9DA5-E497-406B-8AFA-EAA96AB90B33.jpeg
    1E8A9DA5-E497-406B-8AFA-EAA96AB90B33.jpeg
    425.7 KB · Views: 137
  • 3F054C20-8157-4B37-A4E5-5CA667E7FD06.jpeg
    3F054C20-8157-4B37-A4E5-5CA667E7FD06.jpeg
    427.3 KB · Views: 140
I mean why bother, he’s obviously got the numbers wrong. Just ignore them.
Interesting, claiming someone has the numbers wrong when you’ve presented none yourself at all. See above, I hope you’re as happy about it as I am!
 
  • Like
Reactions: Azrael9
Well here is Red Dead 2, at 4K with .5 resolution scaling and Ultra texture quality. It’s running 5ps lower on the minimum FPS than my initial report (I was running native 4K with high textures instead of using scaling) but averages the same 59 FPS. In my quick research I saw that a 2080 TI got 57fps on high. They probably could’ve optimized their tests better and beaten this 5700XT PRO 16GB, but I don’t care to go test that. I have this new card that plays 4K ~60 and here is the proof. Even better, you can go get it and find out if you’re so in disbelief. I do not deny any of the facts about this card, actually, they are against it, and the power envelope is ridiculously small, and yet here are my results. Read ‘em and weep. And enjoy the iPhone shots, I hate windows and I’m not gonna go hunting around the horrid interface to simulate a windows+prntscrn key so I can take a screenshot for a forum post. This is as good as it gets.

Thanks for running the benchmark and showing your settings, which make sense!

You've rightly set textures to ultra as you've got plenty of VRAM, set the lighting and reflection details to high, kept all the other graphics settings to Medium/Low/Off, and used half resolution scaling to reach around 60fps at 4K.

I think the half resolution scaling setting is probably the key to reaching that almost 60fps framerate on the iMac (renders the game at half resolution then upscales it to 4K), with the mix of High/Medium/Low settings also helping. The game should still look great at 4K ~60fps with that mix of settings.

I do think your friend should have a closer look at their GPU and game settings as a RTX 2070 Super should easily be able to perform similarly with the optimised mix of settings you use at 4K to hit 60fps.
 
Last edited:
Interesting, claiming someone has the numbers wrong when you’ve presented none yourself at all. See above, I hope you’re as happy about it as I am!
Lol yah there’s the proof alright. Resolution scale is 0.5 of 4K, hence you are running at 1080p. 2080 Ti my ass.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.