Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Why? Are you telling people not to buy the new MacBook Pro? :D
The card itself seems to be also about 5% slower than the W7000, if you want to compare.

It would also be somewhat faster than the PS4 gfx.

$100 for 2 GiB, $130 for 4 GiB
 
Last edited:
This thread is very deep and confusing. Is the RX 480 support by OSX with acceleration?

AFAIK, it's supported by macOS (but not any other previous OSX) natively. However, acceleration need kext mod. And even after that, it seems only 16 Compute Unit will be utilised out of 36.
 
AFAIK, it's supported by macOS (but not any other previous OSX) natively. However, acceleration need kext mod. And even after that, it seems only 16 Compute Unit will be utilised out of 36.

Thanks for info, I'm really surprise there is no support for rx 480 without hack.
 
AFAIK, it's supported by macOS (but not any other previous OSX) natively. However, acceleration need kext mod. And even after that, it seems only 16 Compute Unit will be utilised out of 36.

What matches perfectly non-released RX 480M specs (16 CUs). 67FF (which I presume it will get) is present in AMDX4100 kext and AMD9500Controller.
 
  • Like
Reactions: h9826790
Yep.
http://wccftech.com/amd-launches-radeon-pro-wx-7100-5100-4100/
http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-newsArticle&ID=2220116

I could see the WX7100 as D510, the WX5100 seems weak to be the D310 though. More like 470, but gets too close to the 7100 though. And they can still stretch the power envelope.
Only D710 missing now.

Or Apple will use these 3 tiers for nnMp, but that would be a shot in the foot. Nah...

Note the 5k HDR DP1.4 in AMD's announcement - Apple will want to target that.
 
Yep.
http://wccftech.com/amd-launches-radeon-pro-wx-7100-5100-4100/
http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-newsArticle&ID=2220116

I could see the WX7100 as D510, the WX5100 seems weak to be the D310 though. More like 470, but gets too close to the 7100 though. And they can still stretch the power envelope.
Only D710 missing now.

Or Apple will use these 3 tiers for nnMp, but that would be a shot in the foot. Nah...
If dual D510 or whatever it will be called will use full Polaris 10, with 2304 GCN cores, and 8 GB of RAM, at 400$ price uplift from base model - that will be good proposition.

If it will be D710, with 1000$ uplift from base model(Or higher price...) - Apple can f*** themselves.
 
If dual D510 or whatever it will be called will use full Polaris 10, with 2304 GCN cores, and 8 GB of RAM, at 400$ price uplift from base model - that will be good proposition.

If it will be D710, with 1000$ uplift from base model(Or higher price...) - Apple can f*** themselves.
Phil, "It's for experience."
I can see 710, 510, and whatever the low end is called happening soon....and they mark that up after that MacBook amateur 2016 release.
 
So...

1024 GCN core chip, with 1.17 GHz core clock, and 2.4 TFLOPs has 50W TBP(power consumption for the board).
896 GCN core chip, with 1.206 MHz(36 MHz difference, and lower core count) has 75W TDP and around that or higher power consumption.

Genuinely confusing it can be, or Polaris architecture was not designed for high frequencies but low power.

I would love to be able to buy, as a consumer the WX5100 chip at "normal" consumer price, and without 6 pin connector.
 
I am liking the WX 5100, but it is expensive.
I hope this is the pro version of the RX 465. And that we will not have to wait long for that.

Also that Apple will enhance the drivers for such things.
[doublepost=1478566767][/doublepost]
So...

1024 GCN core chip, with 1.17 GHz core clock, and 2.4 TFLOPs has 50W TBP(power consumption for the board).
896 GCN core chip, with 1.206 MHz(36 MHz difference, and lower core count) has 75W TDP and around that or higher power consumption.

Genuinely confusing it can be, or Polaris architecture was not designed for high frequencies but low power.

I would love to be able to buy, as a consumer the WX5100 chip at "normal" consumer price, and without 6 pin connector.
https://en.wikipedia.org/wiki/AMD_Radeon_Pro
 
1080Ti spotted at Zauba.
10GB? Well, seems like another 970 all over again if this is true. If it is, will they ever learn? Unless they're not using all mem controllers, it seems they disabled one and not use the full 384b width. Let's wait and see.
 
I was thinking today, about my yesterdays conversation with one game developer who had already got his hands on complete GPU from Vega architecture, however, do not get excited. He didn't said anything about the performance of the GPU, nor how it behaved, I could not pull from him absolutely anything meaningful in the context of performance compared with other GPUs. However he have said more technical stuff in both specific and general terms, of the hardware.

And would you have thought, about the synchronicity...

http://lpgpu.org/wp/wp-content/uploads/2014/09/lpgpu_scalarization_169_Jan_Lucas.pdf

This has appeared in Beyond3D Forum in the thread about Vega GPUs and in the context of Scalarization of the GPU. It is directly related to the patent that has appeared in the internet many many months ago.

What he have said is that Vega will be Scalar architecture and it is pretty much reset of GCN and Radeon as a brand. Both(!) engineering teams were working on Vega architecture for past 2.5 years, to this point, because it had to be synced with Zen release, because the GPUs will be used in Zen APUs. He did not wanted to say how big difference the scalarization of the GPU might have make, because that way he could tell too much specifics. He also said that he knows how Zen will behave and perform, but it is the same story as with Vega - nothing specific I could take out from him. It was great conversation about the gaming industry, technology, poltics in the gaming industry(well, who would have thought...). He also have said that Polaris was mostly made for Apple, Sony, and GloFo request(WSA required buying wafers, what would be better way to do this?).


So in the end, how big difference in terms of performance we could get with Scalar GPU architecture? I see that from theoretical models the improvement is around 30% Register File space saved by this, and around the same amount of power efficiency. But this is theoretical model, based on very old architecture.
 
I was thinking today, about my yesterdays conversation with one game developer who had already got his hands on complete GPU from Vega architecture, however, do not get excited. He didn't said anything about the performance of the GPU, nor how it behaved, I could not pull from him absolutely anything meaningful in the context of performance compared with other GPUs. However he have said more technical stuff in both specific and general terms, of the hardware.

And would you have thought, about the synchronicity...

http://lpgpu.org/wp/wp-content/uploads/2014/09/lpgpu_scalarization_169_Jan_Lucas.pdf

This has appeared in Beyond3D Forum in the thread about Vega GPUs and in the context of Scalarization of the GPU. It is directly related to the patent that has appeared in the internet many many months ago.

What he have said is that Vega will be Scalar architecture and it is pretty much reset of GCN and Radeon as a brand. Both(!) engineering teams were working on Vega architecture for past 2.5 years, to this point, because it had to be synced with Zen release, because the GPUs will be used in Zen APUs. He did not wanted to say how big difference the scalarization of the GPU might have make, because that way he could tell too much specifics. He also said that he knows how Zen will behave and perform, but it is the same story as with Vega - nothing specific I could take out from him. It was great conversation about the gaming industry, technology, poltics in the gaming industry(well, who would have thought...). He also have said that Polaris was mostly made for Apple, Sony, and GloFo request(WSA required buying wafers, what would be better way to do this?).


So in the end, how big difference in terms of performance we could get with Scalar GPU architecture? I see that from theoretical models the improvement is around 30% Register File space saved by this, and around the same amount of power efficiency. But this is theoretical model, based on very old architecture.

It seems that AMD generally made the same bet with Zen and Vega: milk the old architecture to get by until the next stage in microarchitectures is actually ready. Reminiscent of Apple...

I do hope Vega makes the leap. Many consider their GPU architecture to be superior, especially for compute, and the new process nodes help too. Apple's in bed with AMD at this point when it comes to GPUs, so let's see where it goes next year. I do hope they don't just stick with Polaris stuff though, it'd be tragic to release another Mac Pro with a previous-gen GPU, at this point it's the only thing that actually affects performance.
 
It seems that AMD generally made the same bet with Zen and Vega: milk the old architecture to get by until the next stage in microarchitectures is actually ready. Reminiscent of Apple...

I do hope Vega makes the leap. Many consider their GPU architecture to be superior, especially for compute, and the new process nodes help too. Apple's in bed with AMD at this point when it comes to GPUs, so let's see where it goes next year. I do hope they don't just stick with Polaris stuff though, it'd be tragic to release another Mac Pro with a previous-gen GPU, at this point it's the only thing that actually affects performance.
But what "leap" we are talking about? All I got is that Vega will be Scalar architecture. And that Zen+Vega is just restart for AMD, as a brand. First steps has been taken: Radeon, Radeon Pro brands, software initiatives, a lot of open Source movement happening, but nothing in terms of performance, so far.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.