New Apple MBP M1 + Resolve 17 - Is it worth the price?

But it's still not relevant to anybody buying a new machine today.

That was a 2019 release wasn't it?

Why not? You can buy a MacBook Pro 2019 right now. You can buy my CPU right now (or at least, you could until very recently - it's showing as out of stock now, not sure if that's temporary or permanent). You can definitely buy Hendrik's CPU right now.

Even if/when they're no longer in stock as new items, they'll certainly still be available to buy used/refurbished. And/or will be owned by people who might be considering buying an M1 as well/instead, and want to know how it will compare.

The recent point of this thread is to collect comparison figures in real-world Resolve tests, comparing the new M1 MBPs against non-M1 hardware. And any benchmarks of other systems people want to post, which are also interesting.

If you have a benchmark for a 12th gen Intel system + NVidia GPU running the Fusion spaceship comp, or Uli's GPU-effects render test, you should certainly post it. That would definitely be good to see.

I hope someone will test this on an MBP M1 Max, which should perform better. Might be interesting to see if any thermal throttling is going on and if the power mode does anything.

Hi Uli,

I've exported your project with a 14" M1 Max 32 core GPU / 64 Go RAM / 2 TB HDD.

With DVR 17.3.4 it took 19'15 and 12'20 with DVR 17.4.1

Have a nice day

Pascal

Tom,
Thanks for the detailed description on to do it.

Untitled.jpg

Here is the screenshot. With about 1:30 left I heard the fans spin up but not much and by the end they were silent again.

So you rendered in 12:35 with a 24-core M1 Max.

Considering most of the render is GPU-bound, I wonder if the 32-core would perform proportionally better, at roughly 1.3x the speed (or under 10 minutes). Just a guess but this still seems pretty great, particularly for a portable device.

I can't even render this project at all on my desktop (see my signature) and my GPU crashes on the last clip with optical flow and speed warp.

In my opinion, one of the greatest strengths of this new M1 architecture is the fact that you won't run out of VRAM in Resolve.

Thanks for posting your results.

So you rendered in 12:35 with a 24-core M1 Max.

Ah yes, I missed that he had the 24 core GPU. So yes, could well breach 10 minutes in a top-spec M1 Max.

In my opinion, one of the greatest strengths of this new M1 architecture is the fact that you won't run out of VRAM in Resolve.

To be fair, this is also true on non-M1 Macs.

I can count on the fingers of one finger the number of times I've had an out of VRAM message in Resolve, including when I had an 8GB GPU. I've never had it on my new 16GB GPU, even when testing at 8K (which I'm told needs 24+GB VRAM). Where it seems to be really common on NVidia with CUDA.

Roger said something earlier about NVidia GPUs not being able to use system RAM when out of VRAM? If that's correct, that would certainly explain why the issue seems so prevalent on Windows with NVidia vs macOS with Metal. (Though that does confuse me slightly, because I thought Windows itself reserved extra RAM for the GPU - I've seen in Task Manager screenshots that it shows the available VRAM as being real VRAM + half of RAM?)

I can count on the fingers of one finger the number of times I've had an out of VRAM message in Resolve, including when I had an 8GB GPU. I've never had it on my new 16GB GPU, even when testing at 8K (which I'm told needs 24+GB VRAM). Where it seems to be really common on NVidia with CUDA.

Really? It happens to me all the time with my P4000 under Windows 10 - could be a driver issue. The limitation of 8GB in GPU RAM is a show-stopper for me quite often where Resolve is concerned. On my older laptop with only 2GB of VRAM, Resolve is effectively unusable for anything but the simplest projects.

I was unable to load Uli's project and do anything without creating proxy media.

I'll go back to suspecting a bad driver and so now I will hunt a bit for another nVidia driver. I just tried the latest for the P4000 and it crashed resolve almost immediately as soon as I placed the timeline cursor over the last clip (with Speed Warp).

Add - I've just returned to 471.41 and it seems more stable. I'll try again later.

Really? It happens to me all the time with my P4000 under Windows 10 - could be a driver issue. The limitation of 8GB in GPU RAM is a show-stopper for me quite often where Resolve is concerned. On my older laptop with only 2GB of VRAM, Resolve is effectively unusable for anything but the simplest projects.

Yeah, that's what I'm saying - on Windows + NVidia, out-of-VRAM seems really common.

Where on macOS using AMD GPUs and the Metal API, it's extremely rare (at least in my own experience).

There appears to be some difference in how VRAM is handled. I don't know the technical details of what exactly is different - it could be NVidia vs AMD, it could be Windows vs macOS, it could be CUDA vs Metal.

Roger made a remark earlier indicating that it's not possible for NVidia GPUs to use system RAM to augment VRAM:

Has anyone managed to run Uli's test on an Nvidia GPU where it can't offload to system RAM when needed?

That would certainly explain the difference. (Though as mentioned I'm a little confused, because I thought Windows had a system for adding system RAM to the available VRAM; Task Manager seems to indicate it does, at least.)

Anyway my point in all this is that you were pointing out that M1 might be better than NVidia because it wouldn't have out-of-VRAM. I was just saying that yes that's true, but it's not specific to M1. Any macOS system would have that benefit, including Intel with AMD GPUs. And maybe a Windows machine with AMD GPU (though I'm not sure if the performance hit of using OpenCL instead of CUDA would be worth it.)

It is confusing and I don't understand the details of the Windows Task Manager combined memory display.

On some AMD GPUs with the High Bandwidth Cache Controller (HBCC) activated it can use system RAM as a cache for VRAM. And of course all integrated CPU+GPU units share memory. It works, only a lot slower.

Maybe the architecture or API differences between AMD and Nvidia is the reason they perform so differently when it comes to VRAM usage. Maybe shared memory doesn't work with the CUDA API by choice, for performance reasons. Believe it or not, we're not the target market.

The last clip does use the GPU for me, but much lower. Maybe this effect isn't as optimised or something.

On my PC, the GPU load on the last clip is 100%


Tested on my old PC: i7-8700K, 32GB, gtx 1080ti
Render time: 7:55

it plays everything in real time (incl. external video display via HDMI in UHD)
...
but also mean, you are happy that the export is so fast and then comes the last clip +g*

I only have playback from 7 to 15 fps (except the last clip of course)
I think SpeedWarp is the key here which renders much faster on Nvidia cards.

On Nvidia cards with tensor cores maybe. My Nvidia cards with the Pascal architecture do not like Speed Warp at all but work absolutely fine on everything that doesn't use the DaVinci Resolve Neural Engine.

Well, when I tried the test project for the first time, but with 12K sources, I didn't get an out-of-memory message on my iMac with 8 GB VRAM (AMD, of course), but it crashed when getting to the Speed Warp part. That's why I went for 8K footage.

But it's still not relevant to anybody buying a new machine today.

That was a 2019 release wasn't it?

Why not? You can buy a MacBook Pro 2019 right now. .

I don't think the fact Apple is selling 2019 machines today is a positive point.


I don't want to pick on your CPU but even before Alderlake your CPU wasn't the latest greatest.

So far in this thread we've had from one poster

In the video world, MacBook Pro is at the pinnacle.

From another poster the claim he's impressed the Mac is half the speed of your two year old system.
i9-10980XE

https://www.pugetsystems.com/benchmarks ... p?id=61804

10900K system

https://www.pugetsystems.com/benchmarks ... p?id=64640

12900k

https://www.pugetsystems.com/benchmarks ... p?id=64529

On the Puget test a 30% increase tenth gen I9 to 12th gen. I don't think very many people will be considering your CPU. Especially with the improvements in the media decoder in between.

But the big reason

https://www.pugetsystems.com/benchmarks ... p?id=62408

I5 12600k faster than a 10900k

While the MacPro is not really the most up-to-date technology and overpriced, this one is interesting too:

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

....
I only have playback from 7 to 15 fps (except the last clip of course)
I think SpeedWarp is the key here which renders much faster on Nvidia cards.

yes, without this addition on the last clip would have been just under 6:00

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

no one who knows apple can say with certainty that this will ever come.
and max yuryev is for me a loudmouth who is only out for clicks.

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

He is telling about combining 4 pcs M1 Max dies:

4x M1 Max dies with 8 Performance and 2 efficiency each. How do you expect it to Compare to a AMD Threadripper with 64 or in 2022 96 Cores?

4x M1 Max dies with a 32-Core GPU each. He want to use some form of Infinity Fabric to interconnect its unified memory?
But how will you compare it to 2x RTX 3090?

It may be an Alder Lake KILLER, bit not a AMD/nVidea Killer.

Regards Carsten.

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

no one who knows apple can say with certainty that this will ever come.
and max yuryev is for me a loudmouth who is only out for clicks.

That is why they are called predictions.

There is no need here for tossing around childish insults. Lets try to keep it civil.

It's all speculation. He also assumes having four video encoders would make exports four times faster on a single file. Disregarding the time taken to render any grading or compositing before the encode can even take place. No-one knows if the Apple API will be open to use them in parallel for a single video stream.

It's all speculation.

Exactly, and quite interesting speculation at that. The Mac Pro may turn out to be a beast of a machine.

He also assumes having four video encoders would make exports four times faster on a single file. Disregarding the time taken to render any grading or compositing before the encode can even take place. Also, no-one knows if the Apple API will be open to use them in parallel for a single video stream.

Only one thing is certain, and that is that the only thing slowing Apple down is getting the chips fast enough. The supply chain problems are costing the company billions. And Vadim and Max Yuryev are among the best out there for giving us countless tests and comparisons. They are the only ones I know of who preordered eight machines and who have performed dozens of tests, at the dazzling rate of one per day, including many video rendering and export time comparisons which many find valuable, myself included.

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

no one who knows apple can say with certainty that this will ever come.
and max yuryev is for me a loudmouth who is only out for clicks.

+1. He quite often has no clue what he is talking about, yet pretends to be an expert
His famous comment that half resolution decoding applies only to eg. width is one of the examples.
He spends time doing those tests, but this is his job for which he gets paid for.

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

no one who knows apple can say with certainty that this will ever come.
and max yuryev is for me a loudmouth who is only out for clicks.

+1. He quite often has no clue what he is talking about, yet pretends to be an expert
His famous comment that half resolution decoding applies only to eg. width is one of the examples.
He spends time doing those tests, but this is his job for which he gets paid in form of clicks.

There is enough hostility online already. Perhaps we can stick with the topic rather than attacking people.

Attacking?
These are facts- not attacks or speculations or predictions.
Many of these YT experts videos simply mislead people instead of guiding them.

Attacking?
These are facts- not attacks or speculations or predictions.
Many of these YT experts videos simply mislead people instead of guiding them.

Your comments are only derailing this thread. Vadim and Max are not misleading anyone.

And who started it ?

I'm personally hoping they don't limit themselves to a strict SOC architecture when it comes to the updated Mac Pro systems. One of the biggest mistakes with the so-called "trash can" Mac Pro which Apple themselves admitted to was that they painted themselves into a corner in such a way that expandability/upgradability was impacted.

I am hoping they take that into account and make not only the processors but also memory, GPUs and SSDs individually expandable, and keep expansion slots available as well, for the Mac Pro systems.

It is great to have the extreme performance they are achieving with the SOCs, but things change, needs grow, and at the very high end, pros often have different requirements from other users, so the need to support growth and internal expansion is critical for the Mac Pro, even if the rest of the systems might be designed for all-or-nothing replacement.

For memory, one way this could be accomplished using a variation of a NUMA architecture, by continuing to have some amount of RAM on the chip with the CPU (as with the current SOC systems) but also supporting (likely slower) DIMMs external to the CPU which could either be used for fast swap space or as additional memory for lower-priority tasks and the like, with more important things gradually migrating to the in-SOC memory and less heavily used things migrating to the DIMMs until reaching the point where they are finally swapped out to the SSD(s).

For GPUs, they could continue to have an in-SOC GPU, but also continue support for discrete GPUs on expansion cards or possibly re-implement support for eGPUs as an option for those who want them.

Upgrades for the SOC itself could be readily provided for simply by placing it in a socket or on some sort of daughter card.

In any case, I suspect the Mac Pro will probably be the last of the products to see a switch to Apple Silicon. Time will tell.

All this would turn it into a cpu with integrated gpu in a traditional architecture.

It will definitely be interesting what they come up with for workstation. Modularity/expandability are somewhat counter-productive for the whole SoC logic. And it will be even more interesting how AMD, Intel and nVidia counter these developments.

I can add to this discussion by saying I got the new MacBook Pro M1 Max about 3 days ago and it's been great so far. Hands-down the fastest laptop I've ever owned. Terrific keyboard, amazing sound (for a laptop), and the screen is fantastic... but not good enough for color correction. It's like a really nice $500 consumer TV: great, but uncalibrated. As a GUI screen, it's clear and sharp and detailed: the native resolution is 3456x2234, which takes some getting used to.

The setup for the new laptop was an absolute nightmare, but that's because Apple's Migration Assistant has a lot of bugs and flaws, so we wound up having to re-install a bunch of stuff by hand. Eh, clean installs never hurt anybody, and we have all the serial numbers we need.

Resolve 17.4.1 runs like a top, very fast, no lag, but I haven't tried anything serious with it yet -- say, 6K Raw + 10 corrections + diffusion + NR, or any Fusion titles. But it does "feel" demonstrably faster than my previous 2019 Intel MacBook Pro. But it's also heavier and a little "clunkier," and the notch at the top is an annoyance (but not a deal killer). Still, the performance upgrade is worth everything else.

I assume that’s the one with 32 GPU cores. Does it render my small test under 10 minutes?

Second that. You don't need to max out the internal SSD, it will never be enough for all your video anyway.
Even the RAM could be fine at 32 GB, as far as I can see from the tests by MaxTech.

BTW, I made a PC version: https://www.dropbox.com/s/afvx4zismgwes ... C.drp?dl=0

The only change is the output setting to DNxHR, since PCs don't encode ProRes in DR. Of course, the hardware encoding of ProRes in the new Macs can make a difference here.

5:43 from my PC. Interestingly, even though the GPU was maxed out most of the test, GPU BRAW decode was still faster than CPU BRAW decode, which added about 30 seconds.

the screen is fantastic... but not good enough for color correction. It's like a really nice $500 consumer TV: great, but uncalibrated.

Just to clarify, all Apple devices are calibrated to the highest industry standards at the factory and their mobile devices routinely receive higher marks than other displays. In no universe are Mac displays comparable to an uncalibrated $500 television set purchased at Walmart. The out of the box factory calibrated color accuracy of DCI-P3 for Apple's similarly named Pro Display XDR measured nearly 99%, which is an outstanding result. If you need to recalibrate the XDR display for a workflow tuned to a third-party spectroradiometer, here's one recommended by Apple. Also, while not in any way trying to diminish your own experience using migration assistant, I've been using Macs for over a decade and for the most part the experience has been seamless.

Screen Shot 2021-11-14 at 8.33.11 AM.jpg (176.12 KiB) Viewed 12585 times

While some will no doubt freak out at the price, it's a clear indication that Apple takes calibration seriously. I hired a calibration technician to work on my previous OLED TV and when he arrived with an X-Rite colorimeter, I immediately took him for an amateur. Of course, I use one myself; but it's expected that a professional earning their livelihood doing calibration for broadcast, post production houses and hospitals not only here but internationally (certified technicians are in short supply in these parts and must be prepared to travel) would be able to afford a device costing more than $269.00.

Second that. A professional here in Germany who’s a friend of mine wouldn’t travel without his Klein K-10.

In his lengthy and detailed calibration tutorial for the new MacBook Pro XDR displays, BenQ ambassador Art Suwansang says that prior to even calibrating the new MacBook Pro liquid retina XDR (miniLED) display, it was already almost equivalent to his own BenQ SW hardware calibrated reference display.

While performing calibration on the new MacBook Pro with miniLED display, BenQ ambassador Art Suwansang says, “A few things I wanted to share while this is flashing all these different cool amazing colors is that the way how the display behaves when you change into different reference modes is very similar to that of a hardware calibrated display where it’s loading the LUT data, that known data, somewhere. If you ask me, I really think that Apple have gone in to calibrate these panels really well from the factory and they store all these known values on the chip on a LUT somewhere inside the computer logic board and every time you go in and change all these modes it’s loading the data from that LUT. But what Apple has done here that’s different than external hardware calibrated displays is that because they built in the display, they built the laptop, they pretty much integrate everything from the ground up, they can go in and measure the LUT, not just only for the display that’s on here but on all the different reference modes. For example, in the gamma BT.1866 gamma, SRGB for instance, or in all the different color modes and also the different luminance. And this is really awesome because it’s really astonishing we can go in and customize a display brightness by just setting a value up pretty much right from the operating system. This pretty much will never happen with another external display and the reason why is because those external display companies have to constantly think about how this is going to integrate with PC, how this is going to integrate with Mac; because this display is already built in, I mean they can pretty much do everything. So what Apple have done here is really amazing; they have changed the conversation, how we talk about these displays in general and how we go through color calibration in general. My feeling is that these displays, what we’re seeing now, the way how we’re going to calibrate these are going to be the standard going forward – where you just go in to adjust the white points choose the color gamut you want to use and you can get good colors from your display right away. Let me also say that these displays - the color response on them and the accuracy on these laptop displays - are much better than any laptop that have come before from Apple, and that is really saying a lot. And don’t get me started about PCs because I don’t think they even live up to this. So essentially, what Apple have done here is create a closed-loop system for color management that if you don’t want to own these devices, if you’re just using these displays to do your color calibration workflow, you can. I still recommend that if you’re a pro and you need color critical work, getting the external hardware calibrated display is definitely going to be worth it. But what I can tell you right now is once we have done the calibration fine tuning it looks really close to what my BenQ SW display can produce – and that’s a hardware calibrated display – the differences between them are so small that I would attribute that to different display, different backlight technology and also different LCD panel being used”.

So, nothing could be further from the truth than to compare Apple's liquid retina XDR miniLED display to an uncalibrated $500 television.

Nevertheless, it might be premature to tackle calibrating your brand new M1 Max MacBook Pro. Art Suwansang explains, “Now it may not be the best time to do just yet just because the programs and the devices that we have do not necessarily yet understand the way how these miniLEDs are emitting color information and also light information and it doesn’t understand what data is gathering and how to interpolate that and create a good profile”.

If they do the same with a future iMac Pro that might be a very attractive solution.
On the current MBP it may be a viable approach to move your GUI to an external monitor and use the laptop for checking color.

If they do the same with a future iMac Pro that might be a very attractive solution.
On the current MBP it may be a viable approach to move your GUI to an external monitor and use the laptop for checking color.

Just imagine for a moment the brand ambassador for a camera manufacturer making a similarly bold statement to the effect that another brand's approach is the one that will forever change the way cameras operate! hehe

Still awaiting a true test of the HDR capabilities of the MacBook Pro displays, but given that currently, the least expensive 27" 4K HDR monitor runs $2,800, the display alone could justify the purchase of a Mac.

Many of these YT experts videos simply mislead people instead of guiding them.

I had enough of these Apple Fanboy YouTubers videos; many are just misleading and one more I'll send the big balloon orange Trumpy to their channel.

I don't want to pick on your CPU but even before Alderlake your CPU wasn't the latest greatest.
..
From another poster the claim he's impressed the Mac is half the speed of your two year old system.
i9-10980XE
...
On the Puget test a 30% increase tenth gen I9 to 12th gen. I don't think very many people will be considering your CPU. Especially with the improvements in the media decoder in between.

This is all true, but I think missing a big point: these are all desktop systems, compared to a laptop.

That was the point that people like Andrew were making, I believe: that it's quite impressive for the MBP Pro to be half as fast as a decently powerful workstation (and if you discount my system as 'powerful', remember Hendrik's 24-core TR performed the same as my system in that test.)

I don't think anyone in this thread is saying that the new MBPs are workstation beaters. I've said specifically that they're not, and the benchmarks prove that.

If you have a Windows laptop using an 11th or 12th gen chip, please do post benchmarks. If it shows that it's much faster than the MBP Pro or Max, then that would deflate the claims about the performance of the new MBPs.

But right now, they're looking pretty capable.. for a laptop.

One thing that skews the numbers somewhat is the actual utilization of hardware during these tests. For example in Fu spaceship comp my cpu was hovering st best at 50%, usually less. So if Fusion were doing better at hardware use it would go a lot faster. How big this cap is in M1, don’t know. In practice this is irrelevant ofcourse, for this specific comp, because this is how fast it goes in practice. But it isn’t transferrable to other use cases, for example 3d rendering where it can go full 100 easily.

If you have a Windows laptop using an 11th or 12th gen chip, please do post benchmarks. If it shows that it's much faster than the MBP Pro or Max, then that would deflate the claims about the performance of the new MBPs.

But right now, they're looking pretty capable.. for a laptop.

https://www.pugetsystems.com/benchmarks ... p?id=61658

11th gen I think sells for about $2k

BTW there are Apple results in the Puget benchmark datebase. Not that hard for people to find and compare

I can find only one for M1 under Premiere, nothing for Resolve.

BTW there are Apple results in the Puget benchmark datebase. Not that hard for people to find and compare

It would be useful if someone with these Windows laptops could run Uli's test, though, so we would have a better real-world comparison to something we can all benchmark.

Yeah, that's what I meant - Uli's test and the Fusion test, on an 11th/12th gen laptop.

There's no Puget benchmark available for Resolve on macOS, so we can't use Puget benchmarks to directly compare to the new M1 machines.

It would be useful if someone with these Windows laptops could run Uli's test, though, so we would have a better real-world comparison to something we can all benchmark .

Hi.

Both Apple and Intel have Launched new systems. Their priorities are a little different.

Apple have recently launched its M1 Pro and M1 Max Laptops, and we are waiting the its Next Gen Desktop, may be called iMac Pro 2022?
And properly produced in the TSMC 4 nm process node.

Intel have just launched it new Alder Lake Desktop, and we are waiting for the Alder Lake Laptops.
They will be produced in the i7 process node.

The only reason to ask for performance, for the Intel 10th/11th gen which was produced on the 14nm process node, is to compare it with the new M1 Max which is produced on the 5nm process node. Of cause the 14nm laptops are slower than the laptops produced on the 5nm. It is an ODD comparison.

In a few months will we be able to make much better comparisons. We have already seen the first Benchmarks of the Gigabyte's Aero 5 XE and HP's Omen 17 laptops with Intel's Core i7-12700H,
which are produced in the i7 process node. But with performance in the same range as the M1 Max, but at a much lower price.

But when will we be able to compare Intel Alder Lake Desktop to the replacement for the iMac Pro 2019?

Regards Carsten.

Max Tech just shared their predictions for the 2022 Mac Pro which they say will be the fastest consumer computer in the world.

He is telling about combining 4 pcs M1 Max dies:

4x M1 Max dies with 8 Performance and 2 efficiency each. How do you expect it to Compare to a AMD Threadripper with 64 or in 2022 96 Cores?

4x M1 Max dies with a 32-Core GPU each. He want to use some form of Infinity Fabric to interconnect its unified memory?
But how will you compare it to 2x RTX 3090?

It may be an Alder Lake KILLER, bit not a AMD/nVidea Killer.

Regards Carsten.

From MacRumors: Future Apple Silicon Macs Will Reportedly Use 3nm Chips With Up to 40 Cores. The Information's Wayne Ma this week shared alleged details about future Apple silicon chips that will succeed the M1, M1 Pro, and M1 Max family. Ma's report said Apple's third-generation of chips will be particularly impressive, as he said they will contain up to four dies that could allow for up to 40 cores for improved performance. For comparison, the M1 chip has an 8-core CPU and the M1 Pro and M1 Max chips have 10-core CPUs.

In a few months will we be able to make much better comparisons. We have already seen the first Benchmarks of the Gigabyte's Aero 5 XE and HP's Omen 17 laptops with Intel's Core i7-12700H,
which are produced in the i7 process node. But with performance in the same range as the M1 Max, but at a much lower price.

But when will we be able to compare Intel Alder Lake Desktop to the replacement for the iMac Pro 2019?

Regards Carsten.

Lets wait for real products. Lets see how much energy Intel is drawing, how well those laptops are performing real world long term tasks.
Atm. those are still half synthetic benchmarks.

Here are some observations I have but keep in mind I'm in over my head here among professionals.

Since I wanted to keep this machine clean and use it only for Resolve projects I thought I could eliminate bringing over the unwanted apps when using Migration Assistant by unchecking the boxes. For whatever reason it didn't work and brought over things I unchecked. I ended up reinstalling the OS and downloaded just the stuff I wanted from websites since I have all my licenses.

I've had my MBP for a few days and ran a small project through it yesterday. It is 1:22 long and took 8:30 apx to render. I used a power grade with 20 nodes or so which included a bit of NR. Once this was applied the playback was slow and choppy. I kept the timeline for export at the resolution I shot the footage: 5.7k 2.40:1 and the monitor viewing to HD 1080p in the project settings. Render times don't mean much to me but playback does.

I didn't use proxy media and I don't know if I should have activated render cache.
I really wanted an iMac with a large screen, the 16" is nice but still too small to have a meaningful viewer and the node tree all visible. I have to learn how to adjust it better. I did turn off some things to give me more space but it's still a 16" screen.

I'm going to post the video on YouTube later today...it'll be under creative110.

From MacRumors: Future Apple Silicon Macs Will Reportedly Use 3nm Chips With Up to 40 Cores.

Hi.

I have also seen the 3nm rumor on the internet. Here is what TSMC wrote in a Press Release on October 18, 2021:

Quote: ' N3 in 2023
TSMC's N3 technology will provide full node scaling compared to N5, so its adopters will get all performance
(10% - 15%), power (-25% ~ -30%), and area (1.7x higher for logic) enhancements'

'TSMC Roadmap Update: 3nm in Q1 2023'

From: https://www.anandtech.com/show/17013/ts ... nm-in-2025

So if Apple get the 3nm process node will they get better
Performance (10% - 15%), power (-25% ~ -30%), and area (1.7x) compared to its current 5nm process node.

I am sorry to tell, but personally I expect it to be to little to cach up.

As AMD has one CPU architecture group and two implementation groups, and have planed two new CPU generations next year.
One in Q1, where AMD will implement 3D-Vcache in the top Ryzen Zen 3 CPUs, increasing performance around 15%.
And introduce Zen 4 with 96 Cores CPU, DDR5 Ram and PCIe 5.0 in Q3.

Nvidea current RTX 3080 and RTX 3090 are manufactured at Samsung in its 8nm process node, which is a improved 10nm node. Next year will nVidea return to TSMC and get its RTX 4090 and RTX 4080 produced on a smaller process node with a huge performance upgrade.
Here is a rumor and a link:

Nvidia RTX 4070, RTX 4080, and RTX 4090 GPUs could arrive in July 2022

https://www.pcgamesn.com/nvidia/rtx-400 ... -july-2022

And Intel will introduce its gaming graphics card in Q1. Here is a link to a

'Updated: Intel Cans Xe-HP Server GPU Products, Shifts Focus To Xe-HPC and Xe-HPG'

From: https://www.anandtech.com/show/17041/in ... -and-xehpg

Apple told its customer about its Apple Silicon and its two years transition period. Until now have we
seen Apple Marketing Department convince us that Intel can't produce competitive CPU's as Alder Lake. But what about AMD and nVidea?

Regards Carsten.

@ Chris
You were obviously not aiming at a laptop and bought it for availability, right?
I'd get a larger screen for the GUI then and use its internal one for checking colors.

And then, why are you using a 5.7K timeline? It's not a distribution format and no CMOS sensor is resolving as much detail as the number of photocells suggests. It should give you a very good image in UHD.

I updated to Windows 11 and Resolve Studio 17.4.1, and one of those shaved off a little time to 5:20. I guess I'll have to save the buyer's remorse for the next release cycle.

From MacRumors: Future Apple Silicon Macs Will Reportedly Use 3nm Chips With Up to 40 Cores.

Until now have we
seen Apple Marketing Department convince us that Intel can't produce competitive CPU's as Alder Lake.

Regards Carsten.

Is that so? Please link to a single press release or advert where Apple specifically says "Intel can't produce competitive CPUs as Alder Lake", I seem to have missed that one. And while you’re scrambling to find that non-existent link, you mean to tell me that Apple’s rivals have just been sitting idly by, not promising bigger, more powerful chips or comparing themselves to Apple? Be that as it may, it is unsurprising that you, like many consumers, base your purchasing decisions solely according to what marketing departments would have you believe, rather than actual test results. Because if marketing isn’t the reason you buy goods, why bring it up in the first place? That'd be a little like bringing up astrology in the middle of a conversation about benchmark results. LOL What relevance does marketing have to credible sources passing along rumors about chips? Maybe you believe that Apple marketing executives dictate to Wayne Ma and Mark Gurman what to write! hehe Lastly, how can you be so cocksure that every single rival will build a better mousetrap when we’re talking at least a year out? You’re basing your hunch on what exactly? Astrology? In any case, as things stand right now, the folks over at Cupertino absolutely own the laptop space and that understandably has got many PC advocates in a bunch. Isn’t it weird how they squander all of their precious time leaving aberrant remarks in the comments sections of websites like MacRumors? Like one sideshow freak just the other day claiming Apple's marketing is being dishonest because they 'only' measure 1600 nits peak brightness of their XDR display across a 39% window! I'll refrain from giving them oxygen by repeating any more of their idiocies. You’d think they could make better use of their time. It must be exhausting! To return to marketing, Apple could dispense with advertisements altogether, as the performance of their state of the art products speaks for itself, as countless tests by unbiased, objective reviewers - not a few of whom are lifelong PC owners - amply demonstrate. Looking forward to your response.

For those who think the M1 Pro and M1 Max in the MacBook Pro are impressive, the new Mac Pro desktop is expected to come in at least two variations: 2X and 4X the number of CPU and GPU cores as the M1 Max. That’s up to 40 CPU cores and 128 GPU cores on the high-end.
— Mark Gurman (@markgurman) October 18, 2021

I updated to Windows 11 and Resolve Studio 17.4.1, and one of those shaved off a little time to 5:20. I guess I'll have to save the buyer's remorse for the next release cycle.

That's a very nice result! A whole minute faster than my 6900XT.

A few months ago Puget did their 6900XT benchmark and showed the 3090 being close to double the performance in GPU Effects. I had hoped at the time that some part of that would be accounted for by the fact that they tested on Windows, where AMD GPUs use OpenCL, and so comparing the 6900XT to the 3080/3090 was also comparing OpenCL vs CUDA. I hoped that running the 6900XT in macOS with Metal, as I do, might show higher performance.

I'm not sure if it does, or not. Certainly you didn't get twice as fast a result as I did, more like 20% faster. But there's likely some bottlenecks in there as well, especially given this was a mixed-effect render, including a Fusion title. Regardless, there clearly is a significant advantage to the 3090 vs the 6900XT. It'd be interesting to see how a 3080 performs in this test.

Back to the subject of GPU usage, I re-ran the benchmark today (getting 6:20, within one second of my first result). And this time I thought to check my GPU History, as recorded by iStat Menus:



Test period shown by the black square, and red line indicating the peak - around 80%. Once again we see lower GPU near the end, on that Speed Warp clip.