Hmm, I guess this isn’t a surprise, but it is interesting all the Macs that don’t support discrete GPUs were the ones updated today, implying in the future they will either introduce a chip with a much better GPU or otherwise are still trying to work with AMD to maintain their graphics options in later models.
I think they will drop all 3rd party GPUs in everything except the MacPro and even the high-end machines will have the 4 power saving cores along with x number of high performance cores. In total I don't expect Apple to have more than 3 CPU models with varying numbers of cores. The M1 will obviously be used in the low-end iMacs and the high-end iMacs will get whatever goes into the 16" MBPs with extra cores for the iMacPro. MacPro will likely use its own chip since it will be tied to 3rd party cards.
Maybe it was just an embarrassment to Apple to support external GPUs that had slower speeds than their internal one.
Well one of the main concerns was for most MacBooks, anything close to high end AMD GPUs wasn’t an option, we’ll just have to see how these perform to see if an eGPU option is really necessary (again, most pro users would save the eGPUs for the MacBook Pro or iMac or something that’s not today’s introduced Macs, so it’s a low bar).
Given the (limited) data that we have now for performance, the integrated GPU in Apple Silicon with eight GPU cores is about the same as a RX 590.
Huh? Apple said the onboard GPU in M1 is good for 2.3 TFLOPs. Intel Iris Xe in Tiger Lake is 2.1 TFLOPs. Radeon RX 590 is 7.1 TFLOPs
Most people are ignorant of how powerful GPUs are.
The new Radeon RX 6800 for instance:
Apple keeps comparing the GPUs in their chips with the integrated Intel ones, which are well known to be crap. There has been no mention of GPUs external to the CPU yet. I don't think the CPU package has enough pins to support PCIe, in the same way it probably doesn't have enough to support external RAM. Metal scores on an A12X are around 10000, the same as the Intel Iris Plus, but my iMac with a now reasonably old Vega 48 GPU gets 50000.
Maybe it was just an embarrassment to Apple to support external GPUs that had slower speeds than their internal one.
Well one of the main concerns was for most MacBooks, anything close to high end AMD GPUs wasn’t an option, we’ll just have to see how these perform to see if an eGPU option is really necessary (again, most pro users would save the eGPUs for the MacBook Pro or iMac or something that’s not today’s introduced Macs, so it’s a low bar).
Given the (limited) data that we have now for performance, the integrated GPU in Apple Silicon with eight GPU cores is about the same as a RX 590.
Huh? Apple said the onboard GPU in M1 is good for 2.3 TFLOPs. Intel Iris Xe in Tiger Lake is 2.1 TFLOPs. Radeon RX 590 is 7.1 TFLOPs
Apple likes to make this simple, to our detriment sometimes. TFLOP is a weird measure, and you need to be sure that you're comparing like with like half, single, double-precision, and we have no idea what that was based on.
The 590 comparison is derived from other measures that they've used like "three times faster" and the like for other machines. We'll all see together.
I suspect the desktops will have a different CPU (M2? D1?) than the laptops. Presumably some iteration eventually will end up in the Mac Pro, with PCIe support, and with it PCIe GPUs. Either that or Apple will just abandon the iMac Pro and Mac Pro, I wouldn't be hugely surprised.
They abandon their pro lines they might as well stop selling Logic Pro X and Final Cut Pro X, because the entire creating markets will abandon them.
Well honestly I'm not sure they really care, because I don't think there *are* many creatives left using Macs. Apple is so ditzy when it comes to pros (and their software, FCP X for example, Shake, Aperture, etc) that plenty of businesses that rely on pro software have switched to Windows where they can be sure they aren't going to be abandoned at the drop of a hat. The BBC for example used to do most of their editing on Macs, but they switched to Windows and Avid a number of years ago. Not sure what they're using software-wise now.
Well, the folks in the BBC that we know use the iMac Pro, and Mac Pro plus Final Cut Pro, so. There may be fiefdoms, as there ever were.
I suspect the desktops will have a different CPU (M2? D1?) than the laptops. Presumably some iteration eventually will end up in the Mac Pro, with PCIe support, and with it PCIe GPUs. Either that or Apple will just abandon the iMac Pro and Mac Pro, I wouldn't be hugely surprised.
They abandon their pro lines they might as well stop selling Logic Pro X and Final Cut Pro X, because the entire creating markets will abandon them.
Well honestly I'm not sure they really care, because I don't think there *are* many creatives left using Macs. Apple is so ditzy when it comes to pros (and their software, FCP X for example, Shake, Aperture, etc) that plenty of businesses that rely on pro software have switched to Windows where they can be sure they aren't going to be abandoned at the drop of a hat. The BBC for example used to do most of their editing on Macs, but they switched to Windows and Avid a number of years ago. Not sure what they're using software-wise now.
Well, the folks in the BBC that we know use the iMac Pro, and Mac Pro plus Final Cut Pro, so. There may be fiefdoms, as there ever were.
Oh fair enough, there may well be. Been a while since I had anything to do with the TV/film industry now.
This makes you wonder if Apple Silicon Mac Pros will also prohibit eGPUs. Or even internal GPUs on expansion cards.
I expect not. Of course, I could be wrong, but I think the Pro desktops, especially the ones with add-in GPUs currently, will have eGPU support. I'd have thought they might skip the MBP altogether if that wasn't ready, but I expect they have better data on how the 13" is being used. Plus, they do still sell a fairly modern Intel-based MBP 13" which will support an eGPU.
I'd like to see it included when the 16" MBP transitions, but I'd also like to see more ports, especially on the 16"/ (By which I mean the 4 ports they currently have. Two ports is just a bit too far.)
I wish Apple would shit or get off the pot when it comes to supporting pros. Integrated graphics don’t cut it, period,especially since major pro software packages are built to take advantage of real GPUs. They aren’t going to rewrite their stuff for the small Mac market. It’s exhausting being strung along by Apple.
So you've used one of the new Silicon Macs enough to render a verdict on this? And which software packages are you referring to that are built for the new M1 chips?
Hmm, I guess this isn’t a surprise, but it is interesting all the Macs that don’t support discrete GPUs were the ones updated today, implying in the future they will either introduce a chip with a much better GPU or otherwise are still trying to work with AMD to maintain their graphics options in later models.
No one is buying a Mac mini, iPad Air, or 13" MacBook Pro for the GPU power. These are entry level machines for the average person.
Maybe it was just an embarrassment to Apple to support external GPUs that had slower speeds than their internal one.
Well one of the main concerns was for most MacBooks, anything close to high end AMD GPUs wasn’t an option, we’ll just have to see how these perform to see if an eGPU option is really necessary (again, most pro users would save the eGPUs for the MacBook Pro or iMac or something that’s not today’s introduced Macs, so it’s a low bar).
Given the (limited) data that we have now for performance, the integrated GPU in Apple Silicon with eight GPU cores is about the same as a RX 590.
Huh? Apple said the onboard GPU in M1 is good for 2.6 TFLOPs. Intel Iris Xe in Tiger Lake is 2.1 TFLOPs. Radeon RX 590 is 7.1 TFLOPs
Apple likes to make this simple, to our detriment sometimes. TFLOP is a weird measure, and you need to be sure that you're comparing like with like half, single, double-precision, and we have no idea what that was based on.
The 590 comparison is derived from other measures that they've used like "three times faster" and the like for other machines. We'll all see together.
The 3x measurement is derived from the previous Intel IGPs used in their Macs. It really has nothing to do with the Radeon RX 590 AFAICT. None of the machines announced today -- 13-inch MacBook Pro, MacBook Air, Mac mini -- have used discrete Radeon graphics. They've always used Intel's IGP, which is crap. So for Apple to claim that the GPU in M1 is 5x faster or 6x faster only shows how crap Intel's integrated graphics PREVIOUSLY was. It's a bit faster than the 12th generation Iris Xe found in Tiger Lake, which seems like a fair mark, but nothing earth shattering.
And Apple's 2.6 TFLOP figure is FP32, just liked the other quoted figures. A Radeon Vega 8 in a Ryzen 4800U APU is around 1.8 TFLOPs.
Hmm, I guess this isn’t a surprise, but it is interesting all the Macs that don’t support discrete GPUs were the ones updated today, implying in the future they will either introduce a chip with a much better GPU or otherwise are still trying to work with AMD to maintain their graphics options in later models.
No one is buying a Mac mini, iPad Air, or 13" MacBook Pro for the GPU power. These are entry level machines for the average person.
Much ado about nothing as always.
No, it’s about running off at the mouth before any real information is available. They think in terms of the Intel model and nothing else.
I’d say it’s becausw today we have seen the lower end consumee products. Means for ‘normal’ people who don’t need eGPU by default. Real pros should wait for next batch od launched products
I suspect the desktops will have a different CPU (M2? D1?) than the laptops. Presumably some iteration eventually will end up in the Mac Pro, with PCIe support, and with it PCIe GPUs. Either that or Apple will just abandon the iMac Pro and Mac Pro, I wouldn't be hugely surprised.
They abandon their pro lines they might as well stop selling Logic Pro X and Final Cut Pro X, because the entire creating markets will abandon them.
Well honestly I'm not sure they really care, because I don't think there *are* many creatives left using Macs. Apple is so ditzy when it comes to pros (and their software, FCP X for example, Shake, Aperture, etc) that plenty of businesses that rely on pro software have switched to Windows where they can be sure they aren't going to be abandoned at the drop of a hat. The BBC for example used to do most of their editing on Macs, but they switched to Windows and Avid a number of years ago. Not sure what they're using software-wise now.
The BBC uses Nuke, by Foundry. And yes, they now use Windows.
A friend who works there remembers the days when it was all G5 PowerMacs, ah the glory days.
Maybe it was just an embarrassment to Apple to support external GPUs that had slower speeds than their internal one.
Looking at how they benchmark their own GPU, it seems to land somewhere between the Radeon 5300M to 5500M range, slightly edging closer to the latter. And this is incredibly impressive if true. However, it doesn’t touch the speed of a high-end card, so buyers should be aware of it. Apple’s M1 speed to me is also a solid confirmation of their capabilities to deliver a higher-end version for the MacBook Pro 16, iMac and Mac Pro.
Well, the eGPU support has been a hellish crap for me ever since Catalina. And I use a new-ish 2019 13" MB Pro, in an Apple recommended combo (Sonnet box + RX580). The MacBook won't even boot if the GPU is connected, I have to switch it off first, then once it's on, things freeze relatively often. Before Catalina, everything was smooth, so it's not a hardware issue. Apple is good as long as you don't try any kind of hack, even if it's officially supported.
I'm dumbstruck that ALL of the M1 Macs have a maximum of 16 Gb of RAM - even the MacBook Pro! It appears that it is all in the Unified Memory of the M1 chip. How is this a good idea?
I'm dumbstruck that ALL of the M1 Macs have a maximum of 16 Gb of RAM - even the MacBook Pro! It appears that it is all in the Unified Memory of the M1 chip. How is this a good idea?
Fair question. I don't know either, so I ordered the 16 GB Mini. But it's entirely possible that programs compiled for Apple Silicon take less memory than those compiled for the x86 architecture which still has backwards compatible design dating back about 40 years. We are talking Apples (Apple Silicon) and Oranges (x86).
I'm dumbstruck that ALL of the M1 Macs have a maximum of 16 Gb of RAM - even the MacBook Pro! It appears that it is all in the Unified Memory of the M1 chip. How is this a good idea?
This is another thing that makes me think this is essentially a rehashed iPhone 12 CPU. iPhones use PoP (package on package) for the RAM, and more than 16GB physically won't fit. Therefore the CPU package would have to change to add extra pins for things like external memory and actual PCIe (not Thunderbolt), and that might have been too much work to get done for this event.
Comments
The 590 comparison is derived from other measures that they've used like "three times faster" and the like for other machines. We'll all see together.
Well, the folks in the BBC that we know use the iMac Pro, and Mac Pro plus Final Cut Pro, so. There may be fiefdoms, as there ever were.
I'd like to see it included when the 16" MBP transitions, but I'd also like to see more ports, especially on the 16"/ (By which I mean the 4 ports they currently have. Two ports is just a bit too far.)
No one is buying a Mac mini, iPad Air, or 13" MacBook Pro for the GPU power. These are entry level machines for the average person.
Much ado about nothing as always.
And Apple's 2.6 TFLOP figure is FP32, just liked the other quoted figures. A Radeon Vega 8 in a Ryzen 4800U APU is around 1.8 TFLOPs.
A friend who works there remembers the days when it was all G5 PowerMacs, ah the glory days.
However, it doesn’t touch the speed of a high-end card, so buyers should be aware of it.
Apple’s M1 speed to me is also a solid confirmation of their capabilities to deliver a higher-end version for the MacBook Pro 16, iMac and Mac Pro.
Why is that?
Ah! Wait, I was thinking of discrete graphics, not eGPUs.