Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Still having problem of installing RX480

M

mavericks7913

Suspended
Original poster
May 17, 2014
812
281
I followed the instruction base on the thread in here but Mac Pro does not even boot up with RX480 but previous graphic card.

I tried this several times up to adding kext text and then reboot but didn't boot at all.

update1: I manage to tried to follow all instructions but it still not working and the login window is grey background instead of Sierra background.

update2: It's finally working but the GPU performance is not that good that FPS was quite low when too many units on the map from SC2. Video RAM is 8gb and fully working but idk why. This is what it said "AMD Radeon HD Baffin Unknown Prototype Compute Engine". I ran Geekbench and it scored much better than HD5870 tho.
 
Last edited:
mavericks7913 said:
I followed the instruction base on the thread in here but Mac Pro does not even boot up with RX480 but previous graphic card.

I tried this several times up to adding kext text and then reboot but didn't boot at all.

update1: I manage to tried to follow all instructions but it still not working and the login window is grey background instead of Sierra background.

update2: It's finally working but the GPU performance is not that good that FPS was quite low when too many units on the map from SC2. Video RAM is 8gb and fully working but idk why. This is what it said "AMD Radeon HD Baffin Unknown Prototype Compute Engine". I ran Geekbench and it scored much better than HD5870 tho.
Click to expand...

When lots of moving object on the screen and FPS drop. It may be a sign of CPU limiting. Especailly if the situation didn't improve by lowering the resolution or settings.
 
h9826790 said:
When lots of moving object on the screen and FPS drop. It may be a sign of CPU limiting. Especailly if the situation didn't improve by lowering the resolution or settings.
Click to expand...

SoyCapitanSoyCapitan said:
Blame it on poor old APIs. Not cut for gaming.
Click to expand...

I totally disagree.


I have dual X5677 with 32gb of RAM. Other people playing those games with high FPS. I see that guy playing SC2 with better FPS. So I assume that I have a problem with my RX480.
 
mavericks7913 said:
Still the performance should be way better with RX480. Few people playing SC2 on Mac Pro with better performance.
Click to expand...

So, what's the performance of the others Mac Pro (in MacOS I mean)? How about yours?

Do you know their settings? Did you try the exact same settings?
 
h9826790 said:
So, what's the performance of the others Mac Pro (in MacOS I mean)? How about yours?

Do you know their settings? Did you try the exact same settings?
Click to expand...

I dont know their set up but they run games for Max so much faster than I
 
mavericks7913 said:
I dont know their set up but they run games for Max so much faster than I
Click to expand...

Sorry, your info doesn't help to diagnosis the problem. If they are your friend, please ask them. If you don't know them, how do you know that they are running at max setting, and how do you know their performance?

You need to know their setting. If you 100% sure everything is maxed (including vsync, AA, etc), then you need to know the resolution. And I assume all cases are running in full screen mode.

Since your 5677 is the fastest CPU already, (no matter single or dual, it won't make any difference), I assume it's not CPU speed causing the difference.

32GB RAM is more than enough for any games, so, memory is also not the limiting factor.

GPU can be the suspect. However settings, resolution, or even difference OS can be the suspect as well.

If you want to know the truth, you can't jump to the conclusion that GPU is the problem source and only pick the factors that favour for this outcome. We should stay objective, and examine all possible factors one by one.

The RX480 can't perform in MacOS as good as under Windows is a fact, because macOS can only utilise 16 CU out of 36. However, according to the others report. It can still performance more or less like the 7970.

To have much better FPS then a 7970 in OSX gaming, there are at least 2 requirements.

1) the game is very GPU limiting.
2) the others use a much faster GPU (e.g. TitanX)

If either one of these requirement doesn't exist. Then the performance difference can hardly because of the poor performing GPU (assume the RX480 is not particularly buggy in SC2).
 
h9826790 said:
Sorry, your info doesn't help to diagnosis the problem. If they are your friend, please ask them. If you don't know them, how do you know that they are running at max setting, and how do you know their performance?

You need to know their setting. If you 100% sure everything is maxed (including vsync, AA, etc), then you need to know the resolution. And I assume all cases are running in full screen mode.

Since your 5677 is the fastest CPU already, (no matter single or dual, it won't make any difference), I assume it's not CPU speed causing the difference.

32GB RAM is more than enough for any games, so, memory is also not the limiting factor.

GPU can be the suspect. However settings, resolution, or even difference OS can be the suspect as well.

If you want to know the truth, you can't jump to the conclusion that GPU is the problem source and only pick the factors that favour for this outcome. We should stay objective, and examine all possible factors one by one.

The RX480 can't perform in MacOS as good as under Windows is a fact, because macOS can only utilise 16 CU out of 36. However, according to the others report. It can still performance more or less like the 7970.

To have much better FPS then a 7970 in OSX gaming, there are at least 2 requirements.

1) the game is very GPU limiting.
2) the others use a much faster GPU (e.g. TitanX)

If either one of these requirement doesn't exist. Then the performance difference can hardly because of the poor performing GPU (assume the RX480 is not particularly buggy in SC2).
Click to expand...

They are not friends but I know them. Are there any apps that I can run some tests for graphic card?
 
mavericks7913 said:
They are not friends but I know them. Are there any apps that I can run some tests for graphic card?
Click to expand...

Is it possible to ask them what setting, setup, and resolution they use? gamers usually very willing to help the others to get good gaming exp. The friendship can be start from there :D
 
I just benchmarked RX480 with LuxMark and I got this score: 1737. I think something is wrong. A thread starter got 11455 from LuxMarker.
[doublepost=1478463482][/doublepost]Base on LuxMarks, I got

Hotel: 1756
Neumann: 6257
Luxball: 11056

Is it a problem or wait for another updates?
 

Attachments

  • Screen Shot 2016-11-06 at 2.56.07 PM.png
    Screen Shot 2016-11-06 at 2.56.07 PM.png
    125 KB · Views: 102
You're running your GPU on hacked beta drivers meant for a different GPU, you can't expect a lot from that. Additionally, OS X APIs/drivers have a quite high CPU overhead, so you might be limited there.
 
Fl0r!an said:
You're running your GPU on hacked beta drivers meant for a different GPU, you can't expect a lot from that. Additionally, OS X APIs/drivers have a quite high CPU overhead, so you might be limited there.
Click to expand...

But that doesn't explain why am I getting low benchmark score compare to other RX 480 owners.
 
h9826790 said:
11455 vs 11056, nothing wrong here.

p.s. You can't compare Hotel score to the Luxball score.
Click to expand...

It should've be around 40 FPS but it's not.
 

Attachments

  • Screen Shot 2016-11-06 at 5.53.15 PM copy.jpg
    Screen Shot 2016-11-06 at 5.53.15 PM copy.jpg
    436.4 KB · Views: 94
  • Screen Shot 2016-11-06 at 4.40.47 PM copy.jpg
    Screen Shot 2016-11-06 at 4.40.47 PM copy.jpg
    456.3 KB · Views: 117
mavericks7913 said:
It should've be around 40 FPS but it's not.
Click to expand...

I can't 100% rule out that your GPU is not performing as good as the others in OpenGL (OpenCL is definitely fine by comparing the Luxmark score). However, 2560x1440 is NOT comparable to the 1920x1080 result.

Your score may be normal because you were NOT using the standard preset. You can make it run at 8K resolution and have extreme low score if you wish.

In order to make it comparable, please choose one of the standard preset and do not alter anything.
 
ps SC2 is massively CPU dependent, even on windows people complain about this.
do not use SC2 as a benchmark
 
Can you repeat the benchmark with Heaven / Extreme preset or Valley / Extreme HD preset? That's what most people use.
 
  • Like
Reactions: orph
Fl0r!an said:
You're running your GPU on hacked beta drivers meant for a different GPU, you can't expect a lot from that. Additionally, OS X APIs/drivers have a quite high CPU overhead, so you might be limited there.
Click to expand...

The Polaris base code should be the same across GPUs. They aren't in beta anymore. But you are right that macOS drivers have higher CPU overhead. This goes for other APIs on the Mac too.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back