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

AMD Confirms RX 480 At $199 USD, Other APU & Polaris Announcements

ManuelGomes said:
I'm inclined to believe it will be a dual Polaris really.
Stacc, indeed the info there is not very reliable since they seem to only copy and paste from somewhere else and add some extra sauce, most of the times with no sense at all. But the rumors are coming to this.
Where did you read about the GPU for the nMP expected by May?
Click to expand...
Previously in the linked article by you there was information that AMD is supposedly readying Vega GPU for Mac Pro, and that is supposed to go around may 2017.

Later, they have edited the article and cut out whole part about MP.
 
Asgorath said:
https://forums.macrumors.com/thread...os-suffer-from-major-graphics-issues.2019261/

Where are all the people complaining about AMD's terrible driver quality? If this was happening on NVIDIA you'd have to expect folks like SCSC would be foaming at the mouth as they rage about how bad the NVIDIA drivers are.
Click to expand...
Nvidia was providing MacOS drivers, because all of drivers on this OS, were written by Nvidia. Both system, and web drivers. There is no third party AMD or Intel drivers, because it is Apple who writes the drivers for Mac OS, and requests drivers for other systems, installed on Apple hardware(thats why Bootcamp drivers for AMD GPUs are stuck with very old versions of drivers).

And funinest part, in case you missed it: GPU glitches affect BOTH 13 inch, and 15 inch MBP. The ones with AMD and Intel only GPUs.
 
Last edited:
koyoot said:
Nvidia was providing MacOS drivers, because all of drivers on this OS, were written by Nvidia. Both system, and web drivers. There is no third party AMD or Intel drivers, because it is Apple who writes the drivers for Mac OS, and requests drivers for other systems, installed on Apple hardware(thats why Bootcamp drivers for AMD GPUs are stuck with very old versions of drivers).

And funinest part, in case you missed it: GPU glitches affect BOTH 13 inch, and 15 inch MBP. The ones with AMD and Intel only GPUs.
Click to expand...
S*** is hitting AMD's fan, but I suspect it's the Intel HD5xx series... I've seen similar glitches even on one of my clients Lenovo with Intel HD520. And also on Macbook 2016 with Intel HD515. But only when external display is connected to the machine.

This might have something to do with Intel Skylake iGPU.
 
koyoot said:
Nvidia was providing MacOS drivers, because all of drivers on this OS, were written by Nvidia. Both system, and web drivers. There is no third party AMD or Intel drivers, because it is Apple who writes the drivers for Mac OS, and requests drivers for other systems, installed on Apple hardware(thats why Bootcamp drivers for AMD GPUs are stuck with very old versions of drivers).

And funinest part, in case you missed it: GPU glitches affect BOTH 13 inch, and 15 inch MBP. The ones with AMD and Intel only GPUs.
Click to expand...

Apple does not write the drivers for Intel and AMD GPUs. All 3 vendors provide drivers for Apple to include in their OS. NVIDIA also provides a driver for their add-in cards, which provides unofficial/beta support for Maxwell GPUs.

So, given that these problems happen on both Intel and AMD, where are the rabid complaints about their driver quality? Or is that only reserved for NVIDIA? That's all I'm saying.

Having said that, the one consistent part of all these issues is Apple's hardware design. Maybe I don't follow other laptops as closely as the Apple ones, but I'm not aware of persistent issues with thermal design from any other company. Apple continues to push the limits on making their laptops thinner and thinner, and lo and behold, they keep having (major?) issues no matter what GPU they use.
 
  • Like
Reactions: Synchro3
Asgorath said:
Apple does not write the drivers for Intel and AMD GPUs. All 3 vendors provide drivers for Apple to include in their OS. NVIDIA also provides a driver for their add-in cards, which provides unofficial/beta support for Maxwell GPUs.

So, given that these problems happen on both Intel and AMD, where are the rabid complaints about their driver quality? Or is that only reserved for NVIDIA? That's all I'm saying.

Having said that, the one consistent part of all these issues is Apple's hardware design. Maybe I don't follow other laptops as closely as the Apple ones, but I'm not aware of persistent issues with thermal design from any other company. Apple continues to push the limits on making their laptops thinner and thinner, and lo and behold, they keep having (major?) issues no matter what GPU they use.
Click to expand...
HP Pavilion, few years ago. First thing that popped to mind. GPUs fried both from AMD and Nvidia in those laptops.

And I mean: fried.

As a Side note to the thread: http://videocardz.com/64475/are-those-radeon-rx-490-vega10-benchmarks-results

687F:C1 DeviceID. Time to search the macOS. If the traces are about Vega 10, we should see the DeviceID in the kexts. Yes, this GPU is most likely RX 490. Second part: No, it is most likely not Vega 10 GPU.
 
Last edited:
jeanlain said:
I thought they did for intel GPUs, based on specification given by intel.
Click to expand...
They do. It was Nvidia which was providing drivers, and Apple took them, or requested modifications.
They do write their own drivers for AMD and Intel hardware.
 
koyoot said:
They do. It was Nvidia which was providing drivers, and Apple took them, or requested modifications.
They do write their own drivers for AMD and Intel hardware.
Click to expand...

Again, completely false. Go and search LinkedIn and you'll trivially find people who work at both Intel and AMD on their Mac drivers.

Apple owns the OpenGL, OpenCL and Metal frameworks, but the actual hardware drivers are written by the hardware companies themselves.
 
Asgorath said:
Again, completely false. Go and search LinkedIn and you'll trivially find people who work at both Intel and AMD on their Mac drivers.

Apple owns the OpenGL, OpenCL and Metal frameworks, but the actual hardware drivers are written by the hardware companies themselves.
Click to expand...
Are you talking about drivers or firmware?
 
Asgorath said:
Again, completely false. Go and search LinkedIn and you'll trivially find people who work at both Intel and AMD on their Mac drivers.

Apple owns the OpenGL, OpenCL and Metal frameworks, but the actual hardware drivers are written by the hardware companies themselves.
Click to expand...
Well obviously, they have. We both are talking about the same thing, but from two different points of view. Which causes the confusion.
 
koyoot said:
Well obviously, they have. We both are talking about the same thing, but from two different points of view. Which causes the confusion.
Click to expand...

You just said "Apple writes their own drivers for AMD and Intel hardware". I'm telling you this is not correct.
 
koyoot, the issue had been more or less identified as a sloppy software implementation, true or not we'll see.
I guess it was particularly visible with Adobe stuff. And happened regardless of model, AMD or Intel iGPU. Worse in AMD eGPU probably due to higher load.

That one sounds like the dual Polaris 10 490, if it really is so.
 
Fl0r!an said:
Did anyone hear about "Polaris 10 XT2" yet? Sounds like a refreshed P10 to me, e.g. refined process (like they did with Tahiti XT2 back then). Might end up in a RX 485 or something like that...?

https://www.tonymacx86.com/threads/...d-radeon-drivers.197273/page-118#post-1386344
Click to expand...
That is interesting. There is also Polaris 12 traces in the kexts, which is more surprising.

XT2 name has appeared before in 7970 GHz Edition, and in R9 280X. So it might be new, improved process. Maybe after all there is some truth to rumors...

Fl0rian, are you able to deduct the DeviceID for Vega 10?
 
I don't have any beta builds installed at the moment, but from what I've heard the list of device ID's didn't change lately.

But the framebuffer kexts don't map device ID's to specific chipsets anyway, it's just a list of ID's per family (and by the looks Vega might end up in the same family as all Polaris variants - they could also make a own branch later though).
 
Well I am right connecting the dots, and Polaris12 kext is more surprising than I thought before.

It might be the RX 490 GPU. That would be more logical if Vega is coming before the 2 half of next year.

I have to say, I have been caught by surprise by this GPU, but overall my information may be actually accurate.
 
cube said:
What about RX 465?
Click to expand...
Two tumors are floating around the retail line:
RX 465 is 1024 GCN core GPU, with 4 GB of RAM, 128 bit bus and priced at 120-130$.
RX 465 is 1792 GCN core GPU, with 4 GB of RAM, on 256 Bit memory bus, and WITHOUT 6 pin connector, priced at 130-140$.
 
koyoot said:
Two tumors are floating around the retail line:
RX 465 is 1024 GCN core GPU, with 4 GB of RAM, 128 bit bus and priced at 120-130$.
RX 465 is 1792 GCN core GPU, with 4 GB of RAM, on 256 Bit memory bus, and WITHOUT 6 pin connector, priced at 130-140$.
Click to expand...
I think the last one. I was expecting it to be the consumer version of WX 5100.

But the question was if this would be Polaris 12.
 
Polaris 12 might be different chip. I will not be surprised if it will be RX 490. I have quite a lot to believe that is the case.
Fl0r!an said:
I don't have any beta builds installed at the moment, but from what I've heard the list of device ID's didn't change lately.

But the framebuffer kexts don't map device ID's to specific chipsets anyway, it's just a list of ID's per family (and by the looks Vega might end up in the same family as all Polaris variants - they could also make a own branch later though).
Click to expand...
Then I have different question. Is the 9500 framebuffer linked to trash can design for Mac Pro? If I remeber 666's statement correctly, it is.
 
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