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

Looking forward to your response.

Hi.

I can't se when Mark Gurman expect the new Mac Pro desktop to arrive?

So I tried to search for your Mark Gurman link, and found it at twitter with a timestamp of 12:54 AM · Oct 19, 2021. But still no launch date. Do he expect it to arrive in January og at the end of 2022?

As I se it, don't I believe any more, that Apple next year will be able to cach up on the performance of
the coming 96 core Zen 4 CPU from AMD, or the coming 5nm RTX 4090 GPU from nVidea.

Regards Carsten.

Apple doesn't have to do it. Their Pro machine doesn't have to match enterprise solutions and will still have plenty buyers. There is a market for every need.
Atm. none of those mentioned options exist. Maybe components shortcuts will be so high that no one will release anything ( not much less possible than your speculations ).

Today's fastest supercomputer is ARM based (opposite to what was in many last years).

Apple doesn't have to do it. Their Pro machine doesn't have to match enterprise solutions and will still have plenty buyers. There is a market for every need.

Well said!

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.

We're on very different platforms, so it's not as clear as it could be. I tried to test the speed running OpenCL just out of curiosity, but the render immediately failed

Uli's test package is, IMO, a good set for us to have some kind of level base as to what the Apple M1 silicons can do. It will be helpful and appreciative if those of you who have the original M1 (be in Air, MBP, or Mac Mini) with 16Gb of RAM, the M1 Pro (potentially configured to the max memory and cores), and the M1 Max (also configured to the max memory and cores) and run Uli's test, each rendering out to Pro Res, and report your render times here. It will be nice to use as a comparison with some of our older and current laptops (or even desktop) and see if it's worth buying one. I hope there's a few good men/women here who will share such info.

Uli's test package is, IMO, a good set for us to have some kind of level base as to what the Apple M1 silicons can do. It will be helpful and appreciative if those of you who have the original M1 (be in Air, MBP, or Mac Mini) with 16Gb of RAM, the M1 Pro (potentially configured to the max memory and cores), and the M1 Max (also configured to the max memory and cores) and run Uli's test, each rendering out to Pro Res, and report your render times here. It will be nice to use as a comparison with some of our older and current laptops (or even desktop) and see if it's worth buying one. I hope there's a few good men/women here who will share such info.

Reading previous post, here's some numbers for the M1 Max test from Pascal (M1 Max) and Uli (M1 Pro). Can someone do a test on the Original M1 with 16Gb RAM, preferably on an MBP or Mac Mini?

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
Pascal

... to render them into ProRes 422 HQ. My mid-level MacBook M1 Pro needs 25:09. Interestingly, the GPUs in the MBP are running full speed all of the time...

1:59:56 on the 8GB M1. Nothing to write home about, but points for finishing.

Here are my test numbers on my trusty old PC Desktop (spec on signature below).

Export to H.265 took 01:09:08.

test_render_time.JPG (36.18 KiB) Viewed 6015 times

Export to DHXHR 12-bit took 43:55.

test_render_time_dnxhr.JPG (40.56 KiB) Viewed 6015 times

[Tony Rivera]

As has been said in various other threads; Please remain on topic and not resort to bashing or name-calling. This is unacceptable behavior and will lead to further people being removed from the forums if it continues.

Bump. Any test using Uli's package with the original MBP M1 or Mac Mini M1, particularly one with 16Gb?

59min M1 Air Pro vs. 42min 32min for fairly good workstation. If I were to say what impressed me more than it has to be M1 (with its weak GPU). Of course this is just for this specific project and fact that GPU is not used much gives explanation for those render times. Sounds okish.

I think we've left reality if a two generation out of date consumer desktop (Yes an I7 is a consumer chip) is considered a fairly good workstation.

I'm sure I was talking about:

With AMD Threadripper 3960X, 2080Ti and 128GB RAM it took 32 minutes under Win10

which is a decent workstation.

I'm a bit lost with the different projects and such, but the basic conclusion is that a 3090 is still twice as fast with Resolve 17 when compared to the M1 max, right? Doesn't bode well for the new 27 inch macs next year.

Probably around that assuming 3090 is in 'fast enough' machine. It's not easy to say exactly as Resolve needs GPU and CPU and depending on your project and export option things may swing in one or another way (this also assumes you can actually buy 3090).

Based on result from:
Ryzen 3900X 4.4Ghz
32GB DDR4-3600
RTX 3090
it should give you about 2x faster render times (still probably not always).

When it comes to mobile solution with all thermals/weight/battery life etc. limitations, I don't think there is really a match for M1 now.

Bump. Any test using Uli's package with the original MBP M1 or Mac Mini M1, particularly one with 16Gb?

I'll try it with the 8GB M1 on my Macbook later. Frankly, I don't expect it to finish.

When it comes to mobile solution with all thermals/weight/battery life etc. limitations, I don't think there is really a match for M1 now.

I have to agree. It's the best laptop in the world, unless you have some extremely specific need like mobile CUDA.

The two test files stress the machines in very different ways.
The Spaceship is very much CPU bound with only small bursts of GPU usage.
The small DR project is using my GPU cores to the limit, while only the two efficiency cores and the first two performance cores are doing a bit of work, the rest is idling along.

The Spaceship will take forever on a M1 Air or MBP with only 8 GB RAM.
The DR project should get through.

1:59:56 on the 8GB M1. Nothing to write home about, but points for finishing.

That's really slow. But then, my iMac with two AMD Radeon 580s needed close to an hour.

1:59:56 on the 8GB M1. Nothing to write home about, but points for finishing.

Thanks Jack. I'll add it to the above post. Was this on DVR 7.1.4 and renders to Prores? H.265 renders much slower that Prores or DNxHR from what I can tell when I did my PC test.

At least now we have some baselines when comparing the various M1 models within itself and other laptops (probably as well as desktops). I wonder what impact does more RAM (16Gb vs. 8Gb.) have on the OG M1.

That's really slow. But then, my iMac with two AMD Radeon 580s needed close to an hour.

My I7 PC, 32Gb Ram, with two AMD Radeon R9 390X (with 8Gb VRAM each) needed close to 44 minutes rendering to DNxHR 12-bit, and a little over an hour rendering to H.265. So I would say those desktop numbers are pretty much around the ballpark as to what to expect. I've been using this configuration for years working on HD and 4K timeline as long as 140 minutes project, with primary/secondary color adjustments, NR, and some filters (mostly grains, transitions, and blurs) at 23.976, 24, and 30 fps clips, and occasionally short 60 fps with optical flow and playback is effectively good and render times could be as long as an hour for a 140 min length film.

The OG M1 w/ 16Gb has the capability to do decent HD timeline work, ingesting even up to 4K clips and for some of us, maybe is enough. The M1 Pro and Max can definitely do a lot more and I think could be a decent workhorse. My only hope is that Apple will have an Mac Mini M1 Pro or Max in their pipeline. If the price is right, that can definitely replace my old PC.

1:59:56 on the 8GB M1. Nothing to write home about, but points for finishing.

Thanks Jack. I'll add it to the above post. Was this on DVR 7.1.4 and renders to Prores?

Yes, tested on Resolve Studio 17.4.1 and rendering to ProRes 422.

My only hope is that Apple will have an Mac Mini M1 Pro or Max in their pipeline. If the price is right, that can definitely replace my old PC.

Well, from what we have learned about the new screens and their color management, the screen in the laptops might be good enough for anything but the most demanding environments to check color too.
So, what about getting a laptop, using a large screen (as you'd need for the mini) for GUI and the laptops screen for clean feed?

Only makes sense, of course, if you don't yet have an I/O device and a calibrated screen.

My only hope is that Apple will have an Mac Mini M1 Pro or Max in their pipeline. If the price is right, that can definitely replace my old PC.

Well, from what we have learned about the new screens and their color management, the screen in the laptops might be good enough for anything but the most demanding environments to check color too.
So, what about getting a laptop, using a large screen (as you'd need for the mini) for GUI and the laptops screen for clean feed?

Only makes sense, of course, if you don't yet have an I/O device and a calibrated screen.

I already have calibrated displays for GUI and a calibrated Flanders fed by DeckLink 4K for clean feed reference. I also have the LG CX 65 for client monitoring. HDR is not something of a demand for my work in the near term. Agree if one doesn’t have the display, the MBP screen can be use for clean feed too.

Understood.

We had been using the 16GB M1 Mac mini with 10GBit networking and the Intel 16GB MBP with the biggest Radeon card for our projects for about the last 6 months.

We got the one-up from base model M1 Pro laptop two weeks ago (16GB 10 core $2499) and the render times for our projects, 6K DPX in, 4 out to ProRes4444XQ jumped from 1.5fps on the M1 mini to 23fps on the M1 Pro.
The Intel gets us about 5-6fps. The pro does this silently and is the same performance with or without being plugged into the mains power.

There is no comparison between how much more instantaneous everything has been on the M1 Pro vs our Minis.

We do mostly restoration work from film scans, we have AMD 5850 PCs with 2080Ti and for the plugins we use, the MBP is generally around the same performance, faster in some things, slower in others, but of course we need the Macs for ProRes output.

Very happy with the midrange MBP right now, the M1 Max model should be coming next week.

If you can buy from an Apple Store, you get two weeks where you can take it back for a full refund, so I'd try one out and see if it suits your business, then take it back if it doesn't or take it back and get a bigger model if required.

Looking forward to your response.

Hi.

I can't se when Mark Gurman expect the new Mac Pro desktop to arrive?

So I tried to search for your Mark Gurman link, and found it at twitter with a timestamp of 12:54 AM · Oct 19, 2021. But still no launch date. Do he expect it to arrive in January og at the end of 2022?

As I se it, don't I believe any more, that Apple next year will be able to cach up on the performance of
the coming 96 core Zen 4 CPU from AMD, or the coming 5nm RTX 4090 GPU from nVidea.

Regards Carsten.

Unless you have a crystal ball, no one knows if there will be a 96 core Zen4, how well it will run, if resolve will use anything past the first 16 cores efficiently, how much it will cost, and if it will be available or if shortages will continue to be a problem. Same with an imagined new GPU from Nvidia, it could launch at $4000 just for the GPU for all we know, and not necessarily be a big leap with Resolve.

We also have no idea what Apple can do in the desktop space where cooling and power draw can be so different to the requirements for a laptop.

Putting the Current M1 Pro/Max against imagined performance of imagined future products is kind of pointless.

Testing a M1 Pro/Max with your own specific workflows against the computer configurations you can buy today makes sense. For our workflows, the new midrange laptop gives us a massive leap over our existing M1 and Intel machines, both Mac and PC. The midrange M1 Pro cost us less than a 3090 GPU alone. Your workflow might not suit these and a PC might be better. Test it, find out and buy what works for you.

It will be helpful and appreciative if those of you who have the original M1 (be in Air, MBP, or Mac Mini) with 16Gb of RAM, the M1 Pro (potentially configured to the max memory and cores), and the M1 Max (also configured to the max memory and cores) and run Uli's test, each rendering out to Pro Res, and report your render times here.

2020 M1 Mac mini here with 16GB RAM. Uli's test rendered to Prores went like this:

Screenshot 2021-11-18 at 23.17.34.png (24.65 KiB) Viewed 5418 times

His earlier Fusion comp wouldn't even load and hung Resolve. It appears then that the 2020 M1 16GB RAM version is considerably faster for rendering this project than the 8GB RAM version which Jack Fairley reported as taking 1:59:56 on his. I find this very surprising as I thought RAM would make no difference to rendering. However I should state that the project is more-or-less unworkable on my M1 mini even with timeline proxy to 1/4. It stutters badly and is barely usable.

So for laughs I tried the same thing on my 2015 16GB RAM PC fitted with a 1080Ti GPU expecting it to smoke the test but it immediately crashed on loading the project with Your GPU memory is full. Try reducing the number of correctors. This again was surprising as the old PC usually runs my 4K BRAW and 4k H265 drone and GoPro stuff (playback, scrubbing etc.) without needing timeline proxy (thanks to the 11GB GPU) and renders about 3 times faster than the M1. But Resolve crashes on it all the time, so I use the much more reliable mini M1 even though it is slower than the old PC for most Resolve tasks.

I should say also that the M1 mini is much faster than the old PC for everything else other than Resolve. Everything opens quickly. Zip files open instantly. For general computing they are a joy to use, but are at best average for Resolve. It's OK for up to 4K footage provided that you use 1/4 res timeline proxy for BRAW and H265. Higher res stuff will struggle. It also stutters badly on Fusion titles but catches up OK after caching. NR, Magic Mask and Optical Flow also run very slowly indeed. I tried doing some Fusion stuff recently (a simple animated sine curve) and it actually crashed - the first crash I have seen in months of using it.

So it seems to me that the suitability or otherwise of all these machines being discussed is all about what you want to do in Resolve.

• Work only in HD or 4K, use no Fusion, and very little noise reduction or FX? Then the 2020 M1's (in 1/4 timeline proxy) are fine, even with 8GB RAM as Alex Jordan argues in his videos.
• Need 6K plus? Get the 2021 M1 Pro or Max devices. Though it's not clear yet how well they work with NR, Fusion etc.
• Need to do it all? Get a big desktop PC or Mac Pro with 128GB RAM and the biggest GPU (or multiple GPUs) you can fit in there.

So many of these Youtube vids concentrate on just one aspect of Resolve and are therefore borderline misleading. Alex Jordan (who I greatly respect for his training videos) has put out several now claiming the 2020 M1s are as good as the new M1 Pro and Max models for multicam viewing provided you use 1/4 timeline proxy. But that's only one small part of Resolve and proves nothing. The better GPUs on the new 2021 M1 Pro & Max models are clearly much better than the 2020 M1s for all-round performance as the test results here show. Same for many of the other videos out there. The creators seem to start from a premise they wish to prove then cherry-pick tests to "prove" their point. I think the Zebra Zone video

first mentioned further back in this thread is probably the most well-balanced test out there so far (as well as being bonkers and brilliantly made).

Maddie's advice is the most sane one I've seen on the subject. It all depends on your needs and your workflow. These are very capable machines, but no be all, end all.

But those with the dreaded memory messages from Nvidia cards should take a long, hard look at them. I've yet to see such a problem on an M1 with 16 GB or more, what ever you throw at them. Slow? Can be. Swapping? Yes, with really heavy Fusion stuff. For those with such a message on a PC, try to replace the sources with 4K BRAW and use the same functions and effects as I did.

Putting the Current M1 Pro/Max against imagined performance of imagined future products is kind of pointless.

Hi.

Both Mark Gurman and I are looking into the next generation.

Quote: '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'

We don't know how well the M1 Max will scale, but I expect the 2x or 4x M1 Max to arrive some time in 2022?
What I am comparing the future M1 Max x2 or x4 with is not the current, but also what we expect some time in 2022. Do you find that wrong?


AMD's Lisa Su hold last week a Data Center Keynote. At the end she talked about the coming Zen 4 CPU. It will come with up to 96 Cores in 2022 and 128 cores in 2023. Here is a link to her keynote.
The Zen 4 part begin at 32:30

We do not all work with 8K REDCode RAW. But many of us want more cores in the future

Here is a link to AMD CES 2020 keynote. It is 51 min long:

From 46:02 to 48:02 do RED staff tell how amazing the 64 core Threadripper 3990X is for 8K REDCode Raw.


And 'Nvidia Lovelace GPUs booked in as TSMC 5nm parts' according to this and many other links:

https://hexus.net/tech/news/industry/14 ... 5nm-parts/

The current top RTX 3090 is produced at Samsung on a 8nm process node. What actually is a improved 10nm node.

From 10nm to TSMC's 5nm is two full process nodes. I don't know what kind of improvement you will expect. But the rumors expect a huge increase in performance. If you don't need that kind of performance, can you always buy a cheaper RTX 4080. And then you will have the possibility to upgrade if needed.

Regards Carsten.

If anything please move these threads to other section.
They are not directly related to Resolve at all. Those are tech news- cool to read but not necessarily here.

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.
.

I got 14 inch and speakers are definitely quite impressive. Keyboard is good, hinge is bit on the weak side (compared to few years old 13inch model), thicker body is not a problem. Notch is bit annoying, but not a big deal either.
Screen is very nice (no idea how close to P3 or PQ curve as can't measure). Fact that there are few refresh rates and predefined presets which lock screen to standard values is nice- good direction.
It's not true that SDR mode is locked to 500nits, which whole internet keeps repeating.
You simply copy 500nits preset and change SDR value to eg. 800 (975 works where 980 seems to not anymore) and you get crazy bright SDR mode (brightness will be locked though).
I actually watched an episode of The Witcher on 2 laptops simultaneously- one SDR Rec.709 and 2nd Dolby Vision on this new one. 2 home laptops, 2 grades yet they looked so close to each other. Absolutely close enough for home devices. Quite impressive. Some scenes shown nice HDR advantage, but overall I agree (after going through few movies) that HDR grades are quite dark by average. This been main complain among Blu-ray users. Many (maybe early?) HDR/DV grades are darker on average scenes than SDR grade. This makes no sense. Of course HDR is not about making whole movie at 800 nits by average, but no idea why average needs to be lower than on SDR grade (we have plenty room to play on HDR and tiny range on SDR)?
In the same time many scenes on SDR had this typical/annoying blue tint in shadows, which was not present on Dolby Vision grade- nice improvement. Overall DV titles on Netflix look stunning one that screen and I don't think home users needs anything more. It's a joy to watch.
Overall first experience with laptop is great. Smooth switch from old with much better keyboard/speakers/screen as main difference.
There seems to be problems with YT and HDR in Safari, but on Chrome all works well (including 8K AV1 hardware decode) and things like Peru HDR demo looks stunning.
So far never heard fan at all, which is so different to Intel Mac 16. M1 is simply different league when it comes to power/heat management.
Playing UHD 60p ProRes and seeing that CPU does nothing is cool. Exporting HD in Resolve from ProRes to ProRes with simple adjustments at near 1000fps is also cool
WiFi seems to be not very fast. Maybe early issues. Constantly 20% worse than on older Macs.

Did not try Resolve much- just a bit and it was smooth experience (again no fan on short UHD ProRes or h265 exports).

Some interesting review:
https://www.dpreview.com/reviews/laptop ... -vengeance
it just shows that theoretical speed of GTX 3080 is one thing and its performance inside some real machine is another.

Some measurements of new Mac screen:
https://forums.macrumors.com/threads/ap ... t-30551191
looks good enough for many projects.

Seems to fit Art Suwansang's position, who has always been a kind of BenQ evangelist after all:

Here's a text review of the 16" M1 Max 32GB, from RedShark, including the reviewer's experience of Resolve 17.4.2 and FCPX.

In the main I'm extremely happy with the M1 Max MacBook Pro. Despite some of my software based grumbles above, it has made me enjoy editing again rather than coming to a session wondering if I'll be able to hear myself think. It's made doing VO easier as well due to its quietness.

The reviewer throws some fun shade at many of the YouTube videos of these machines:

Has it made me sit at my desk with my mouth open wide and my hands cupping my head like a Youtuber's Macaulay Culkin Home Alone poster lookalike pastiche with a big yellow arrow pointing at nothing in particular, and exclamation marks all over the place? Erm, no. But it is very fast for most tasks and it has made my video editing life a lot more pleasant.

Touché!

Rendered in 4:44.

Didn't even bother to shutdown 2 Hyper-V VMs, VST Racks in Cantabile and a full system image backup was running to drive B:\ concurrently.

This is a 5900x & RTX 3090 with 128 GB slow ECC memory and all overclock disabled. Even AMDs native PBO is off.

I scrolled through this thread, but couldn't find how long it takes on the M1 Max?

Whatever the number for the M1 Max is, this benchmark will make it look artificially good as rendering this project virtually creates no CPU usage.

So half of the average Resolve workstation remains unused, giving the M1 Max a pretty unfair advantage.

Which functions would you like to add? Resolve is doing most of the complex tasks in GPU.
And, for sure, a high-end PC desktop will be faster.

Alright. Thanks for creating the benchmark as a base line. What is the time it takes a top end M1 Max and which export codec am I supposed to set?

I'd use ProRes on the Mac. Admitted, a bit unfair

DNxHR on a PC.

So the fastest M1 Max number from this thread seems to be 12:20 and that is using on chip ProRes encoding.

So a 3090 is 2.8 times faster at least. My build was around 5000 Euro, without pandemic inflation probably 3200 Euro.

Considering that a 3090 is only slightly faster than a 3080, I have no idea how Apple marketing comes up with 3080-like speed.

Seems like a nice machine overall, but also extremely over hyped and much worse performance for the Euro / Dollar than a Windows desktop.

So, as I wrote, a nice laptop but no match for a PC desktop.
The low energy consumption is good for the environment, though. But I'm afraid if we consider the energy consumed in production and for worldwide computing infrastructure, it doesn't count much. Not even to mention bitcoin mining with all those GPUs we can't buy for Resolve.

So the fastest M1 Max number from this thread seems to be 12:20 and that is using on chip ProRes encoding.

So a 3090 is 2.8 times faster at least. My build was around 5000 Euro, without pandemic inflation probably 3200 Euro.

Considering that a 3090 is only slightly faster than a 3080, I have no idea how Apple marketing comes up with 3080-like speed.

Seems like a nice machine overall, but also extremely over hyped and much worse performance for the Euro / Dollar than a Windows desktop.

Do you also believe Apple’s pr team?
There is most likely single task where Apple’s GPU may be as fast as 3080. For pr this is enough. It’s advertising at the end. Was Apple talking about desktop 3080?

This test is very GPU orientated. Overall your desktop advantage will be probably 2x which is still a lot. Just don’t forget you are comparing desktop to a relatively small laptop.
Price is not the biggest ‘problem’ here and it’s hard to do a fair comparison. There will be cases where M1 makes way more sense. For those who need portability sometimes, for those who will use this laptop as home machine as well etc. No idea why there is such a huge tendency to make things black or white. World is not B&W and there are many scenarios and you always should carefully evaluate your case.

I’m sure you can design a test case where your desktop will actually loose against M1 Max. Feed this to public and you will create even more wars. People believe in what they read waaay to easily, but if they can’t be bothered to check things properly then they can only blame themselves.

M1s are probably first Apple product which are worth their price.

I have a desktop Threadripper 3970X CPU /3090 GPU /64GB RAM and an M1 Macbook Pro with 16gb and am hoping to get my M1 Max 32core/64GB delivered later this week.

I do tons and tons of not so much editing, but colour grading and rendering (with some Temporal NR) with just HD files, so will be interesting to see how the 3 compare.
The M1 MBP doesn't do so bad in rendering time compared to the big, power hungry desktop (which kicks out tons of heat when rendering), but as soon as NR is added, then the M1 slows right down in rendering.

I've not seen any real tests on NR/rendering with the new M1 32core Max yet, so depending how it performs will determine whether its a keeper or not.
We need a second machine and the M1 sounds perfect as its not big and won't generate the heat like the desktop does.
Having x2 desktops working at once would be a killer in the Summer.

This test is very GPU orientated. Overall your desktop advantage will be probably 2x which is still a lot.

I'd assume the reverse to happen, as CPU performance is the weak part of the M1 and using both GPU and CPU cores also likely makes the SoC throttle.

This is why I wrote earlier I'd like to see some CPU usage generated by the benchmark. This absence is likely very favorable for the M1 SoC.

There is quite a bit of NR in my little benchmark. But I have no full-blown M1 Max to test.

I'd assume the reverse to happen, as CPU performance is the weak part of the M1 and using both GPU and CPU cores also likely makes the SoC throttle.

This is why I wrote earlier I'd like to see some CPU usage generated by the benchmark. This absence is likely very favorable for the M1 SoC.

Not so sure. M1 doesn't throttle much and that quickly.
Your GPU should be relatively stronger compared to Max than your CPU camped to those 10 cores. M1 also has advantage of unified memory, where data doesn't need to travel between GPU and CPU. Add Resolve OS differences and results are not that obvious.

Why don't you compare it to your XPS.

I'm hoping I'll be getting my 16" M1 Max 10 CPU/32 GPU/64GB RAM tomorrow, so will run the test once that's all set up and report back.

While I'm waiting on my M1 Max, here is what my desktop PC did it in.
Maybe I'm going to be disappointed with the M1 now seeing a sub 5 min render time

Ryzen Threadripper 3970X 32-Core Processor 3.70 GHz
Nvidia 3090
64GB
Win 10
DVR 17.4.2 Build 9

You mean 5min.
You have very strong desktop machine, so don't expect M1 Max to be better.
This test is all abut GPU performance and yours is top of the line and costs itself 65% of M1 price.
Card itself draw 3x more power than whole M1 machine and it's about the size of M1 machine Be realistic.

Your card won't do much by itself- you need all other components + some monitor. It's really up to you to decide if you need any portability and other advantages which M1 offers. There is no universal answer, yet so many (according to current trend) try to put down everything into 1 simple answer.


Btw. this Polish shop claims they have 3090 and 3080 in stock:
https://www.komputronik.pl/category/109 ... tiveOnly=1
this is fairly well know shop, so should not be any scam.
Looks like the only place which has them. MSI one is about 3700$ though, so sounds very expensive.

You mean 5min.
You have very strong desktop machine, so don't expect M1 Max to be better.
This test is all abut GPU performance and yours is top of the line and costs itself 65% of M1 price.
Card itself draw 3x more power than whole M1 machine and it's about the size of M1 machine Be realistic.

Your card wont do much by itself- you need all other components + some monitor. It's really up to you to decide if you need any portability and other bits which M1 offers. There is no simple universal answer, yet so many (according to current trend) try to put down everything into 1 simple answer.

oops yes, I meant minutes not seconds and edited original
Yeah, I know the M1 won't compete, was just curious to see what time the PC could achieve.

The thing is, I just went back into my Amazon/Overclockers account added up what the PC cost me - it was £4992 GBP (minus monitor/keyboard/mouse and I built it myself) and the new M1 is costing me £4099 so thats not a huge amount in it.
The PC also has x2 fast SSD drives with a 6GB total and a 10GB network connection where the M1 will only have 2GB.
Yes, you get a great screen with the M1 but I'll use it 99% of the time with a big external monitor that I already have. Ideally, I'd much prefer a non laptop but not yet available from Apple and cannot wait.

So why I dont i buy/build another similar PC... aside from the heat/power draw and size, Almost every component I bought (aside from the SSD drives which have gone down in price) a year ago to the day almost has gone up in price.
I got lucky with the 3090, I waited just 3 weeks and paid £1500 for it, they are now getting on for double so a new build with a like for like spec would cost over £2k GBP more now.
Kind of weird as you'd always expect a loss after a year on tech/PC stuff.

If you don't need any portability or OSX features then no point buying M1 for Resolve itself (specially if you use RAW source, NR etc.).

I just went back into my Amazon/Overclockers account added up what the PC cost me

Hi.

I wonder why you deceided to use your Amazon/Overclockers account?

I know HP OMEN Desktop in some places of the world is sold preconfigured. But here is a link,
where you can build your own configuration from $2900, including a NVIDIA® GeForce RTX™ 3090 24 GB:

https://www.hp.com/us-en/shop/Configure ... quantity=1

Regards Carsten.