This was by far the most clean and concise comparison video I’ve ever seen and I’ve been consuming them all so great work! Also your explanations about the tech were super clear. Is there anyway you can do a similar video for After Effects (m1 native beta)? AE and C4D are the two programs I’m upgrading for and it’s great to see these 4 computers go at it to help make an informed decision. Thanks!
Great comparison video. 👏 just wondering if you could do some after effects playback, performance and export especially with adding some heavy effects. Would really appreciate it man 👏
Hey @jJoseph from LearnSketch. Great video. according to your chart+ higher block size 2x3090 '1:49" and m1 max with higher block size "6:00". that means 1x3090 is just 1.75x faster then m1 max?
Is this comparable to Blender cycles GPU compute renders? As Blender announced that we'll have to wait the 3.1 to get GPU compute render possibilities I'm looking for comparable renders to get an idea on how good the new macbook pros could possibly be
Can you make a video with the m1 max 32 core MacBook running cinebench and 3d mark wildlife test at the same time and see if the cpu or gpu throttle in terms of clock speed. Really debating on getting the m1 max 32 core but want to see how it performs under full cpu and gpu load at the same time.
Is Cinema4d and Redshift optimized for Apple Silicon or are they running thru the Rosetta 2 emulator? Would be curious to see blender tests as well. And how this stacks up compared to your Desktop PC with only one 3090 in it
Still trying to wrap my head around the Mac M1s. I am contemplating purchasing a Mac Studio Mini for C4D and Redshift. But I can't quite get a straight answer as to what GPU would be the best... and with r26, Redshift is now CPU capable... so is the GPU even an issue? I am working on a much older Mac 12 core with Radeon RX 580 (8 GB). Red Shift is not an option and client demands for more realistic renders and frankly the enormous render times I am facing make an upgrade inevitable... mandatory. I am a Mac guy through and through, so PC is not really an option for me. Not a lot of options in terms of GPU in these Mini Studios. No real upgrade path. Any thoughts or suggestions would be greatly appreciated.
Wow this is actually a big surprise for me. It's way faster than I though in redshift. That changing the bucket size would make that big of a difference is very surprising for me. I wonder if the new Mac studio ultra with 128gb can go even up to 1024 buckets but I guess that's a multiple of 4 from the 512 so maybe that is stretching it a bit. But if we are gussing that the ultra will be close to twice as fast as the max, then that also means that a ultra comes close to at least 1 x 3090 which is still very nice and definitely not what I expected it to in something like Redshift. That has to be tested out, that makes me consider getting a fully spaced ultra.
Thanks homie, this is a huge help - I currently have that exact Intel macbook pro. I work in After effects mainly but have been needing to render more C4D stuff recently. It help's knowing I can still work decently without having to buy a whole other desktop.
Well, a monster desktop rig vs a laptop... it was never going to be a fair comparison. But it was fun, and it does bring some of the hype down to earth. A fairer comparison would be the monster rig vs an M1-Max or M2 Mac Pro, whenever that gets released.
But he said this exact test uses almost no system ram, and that it's also not pushing the vram. So probably another test would make more sense for such a comparison?
Would have been great to see your render specs as that makes a huge difference. Are you doing GPU rendering with all of them or are you comparing GPU rendering on the PC with CPU rendering on the Macs?
The larger bucket will help even in the dual GPU. the performance loss from only being able to use 1 GPU at the end is far and away outstripped but the performance advantage from using the larger bucket.
When the GPU market cools off, hopefully a 3090 can be found for $1500 for 3.5x the rendering performance of the M1 Max. Right now the prices of these cards is just outrageous.
You’ve got to lock in a good price for these GPUs when they first come out. I bought 3 for the price of a loaded Apple laptop. So no, it’s no comparison…
I do not agree - if you have like 1000 frames to render then 3090 is the way to go when it comes to timings. Also even a cheaper 3080 wil be faster than M1 Max chips.
@@nyambe I don't use Vray anymore, but it should work without issues. Apparently an update for Vray is coming soon with native Apple Silicon support, so aside from a lack of GPU support, at least the CPU will be fully used.
Is redshift utilizing hardware raytracing? I’m personally betting that Apple will put some of that capability in top end M2/M3, since video and AI acceleration is such a success
No, but Metal already supports raytracing. You don’t need dedicated RT cores to do RT, so Apple could technically work with render engines to implement this somehow. As of right now, RT won’t work until that’s figured out.
@@Notacet Of course, and we only really have the performance of dedicated RT cores to go off on at the moment. However, it is technically possible for ray tracing without dedicated RT cores, and it seems like Apple is looking into this.
If we assume that with single GPU, it will take double the time, 3:30 minutes then M1 MAX is like 60% of RTX 3090. Coming iMac Pro and Mac Pro will have 2X and 4x cores. So the 4x Max version will be faster than this dual 3090 workstation.
You are comparing 2 gpu’s vs 1 with shared memory. Deactivate 1 gpu on your pc in RS and rerun the test to make it apples to apples. Don’t think would be faster than the 3090 but would be a fair comparison.
The reason is this is what you can get on each computer maxed out. This is a better real-world scenario than the one you're proposing. With a PC you CAN double the video cards, with eGPUs you can do even more. With a Mac, it's all integrated technology, so you get what you get. But it's really helpful to show, "with this computer I can get this, and this computer I can get this".
That's pretty absurd performance given the power consumption difference. 2x3090 would be about 700w or so. The M1 Max peaks at around 80w I believe(saw somewhere that the M1 Pro is rated for 60w).
So, to recap: a MacBook Pro isn't EVEN in the same league as a Ryzen Threadripper with dual RTX3090. This surely must have been one of the biggest letdowns of 2021, and I thank RUclips for providing this valuable insight. Bullet dodged, money saved.
None of these Apple computers will work for me in a production environment for clients. I need render times below a minute a frame. Multi-GPU PCs are still the way to go. As far as bucket size goes the rule of thumb is bucket size set to small for CPUs and larger for GPU rendering.
Of course a full desktop PC would be faster than a laptop but my reason for getting one is that I need a laptop and I wouldn't dare go close to a windows laptop-those things are nasty
Is it just me, or is comparing a laptop with a single GPU to a desktop with two gfx cards, a bit of a silly comparison? Really, for this to be an equivalent, you should have used two laptops with team render utilising both chips, or just compared a single gpu. Of course a computer with two gpus will be faster than a computer with one gpu, that's kinda obvious. Otherwise, very grateful for the video. I think this is what most redshift users have been after. I would prefer to know single high-end gpu vs the pro and max scores, but otherwise very helpful. Also yes, costs of the comparison system would be helpful too. Probably worth doing this again in a few months when redshift have refined how it works on m1s. Right now we are just seeing out the box stats I think. As opposed to the Nvidias etc which have had years of tweaking at this stage. Just my opinion tho. Thanks again for video!
I don't see a point of this comparison - Redshift is written mainly under CUDA technology and it is obvious that it will work much faster on Nvidia's cards (especially on such gpu like 3090) cause of the best optimization for these devices. Or if you wanted just to show how SLOW Apple's tech is then ok :)
Why compare a MacBook against a dual dedicated graphics desktop PC. At least try comparing against a 3080 laptop which are over £2000 and more relatable. Pretty sure redshift isn’t designed for apple silicon - try blender.
This was by far the most clean and concise comparison video I’ve ever seen and I’ve been consuming them all so great work! Also your explanations about the tech were super clear. Is there anyway you can do a similar video for After Effects (m1 native beta)? AE and C4D are the two programs I’m upgrading for and it’s great to see these 4 computers go at it to help make an informed decision. Thanks!
This was a great comparison video! Really helped a lot. Thanks!
Great informative video! Thanks, where can I find the video about the PC with the dual 3090s?
Great comparison video. 👏 just wondering if you could do some after effects playback, performance and export especially with adding some heavy effects. Would really appreciate it man 👏
It would be nice to put final graphs again with latest render times . Also, is redshift fully optimized for the new MacBook Pro max ?
Great video, exactly what I was looking for! Thanks!
Hey @jJoseph from LearnSketch. Great video. according to your chart+ higher block size 2x3090 '1:49" and m1 max with higher block size "6:00". that means 1x3090 is just 1.75x faster then m1 max?
Would i notice any BIG difference between the Studio Ultra and Studio Max when rendering out of After effects or C4D?
Thx for sharing! Did you change the bucket size to 256 or 512?
Is this comparable to Blender cycles GPU compute renders? As Blender announced that we'll have to wait the 3.1 to get GPU compute render possibilities I'm looking for comparable renders to get an idea on how good the new macbook pros could possibly be
Can you make a video with the m1 max 32 core MacBook running cinebench and 3d mark wildlife test at the same time and see if the cpu or gpu throttle in terms of clock speed. Really debating on getting the m1 max 32 core but want to see how it performs under full cpu and gpu load at the same time.
What was the base for comparison? the price maybe? all of them cost the same? or what is the base.
Great video, question: are those Macbooks both the 16 inch models? Since the 14 inch has less cooling and throttles faster.
That’s right Rick! All of these are 16” models
Would love to benchmark my Mac Pro with your files and see what happens! Is it downloadable?
Is Cinema4d and Redshift optimized for Apple Silicon or are they running thru the Rosetta 2 emulator?
Would be curious to see blender tests as well. And how this stacks up compared to your Desktop PC with only one 3090 in it
Cinema 4d and redshift are native on the m1
Still trying to wrap my head around the Mac M1s. I am contemplating purchasing a Mac Studio Mini for C4D and Redshift. But I can't quite get a straight answer as to what GPU would be the best... and with r26, Redshift is now CPU capable... so is the GPU even an issue? I am working on a much older Mac 12 core with Radeon RX 580 (8 GB). Red Shift is not an option and client demands for more realistic renders and frankly the enormous render times I am facing make an upgrade inevitable... mandatory.
I am a Mac guy through and through, so PC is not really an option for me. Not a lot of options in terms of GPU in these Mini Studios. No real upgrade path.
Any thoughts or suggestions would be greatly appreciated.
hi, i have m1 max, do you know if this will perform better than a good spec pc with a 1080ti?
Any interest in doing a follow-up with the M1 Ultra?
Can you try C4D ProRender rendering test
Wow this is actually a big surprise for me. It's way faster than I though in redshift. That changing the bucket size would make that big of a difference is very surprising for me. I wonder if the new Mac studio ultra with 128gb can go even up to 1024 buckets but I guess that's a multiple of 4 from the 512 so maybe that is stretching it a bit. But if we are gussing that the ultra will be close to twice as fast as the max, then that also means that a ultra comes close to at least 1 x 3090 which is still very nice and definitely not what I expected it to in something like Redshift. That has to be tested out, that makes me consider getting a fully spaced ultra.
Thanks homie, this is a huge help - I currently have that exact Intel macbook pro. I work in After effects mainly but have been needing to render more C4D stuff recently. It help's knowing I can still work decently without having to buy a whole other desktop.
Thanks for this. Really useful to see where it sits compared to your dual card PC setup. It's a high bar and useful to see how far off the Max is.
Cool video, but to be fair... A single 3080 or 3090 would have better a better comparison.
Agreed! Just working with what I have in my little room here 🙂
@@josephatodaro For sure! Great video. I currently run a 3080, but have been considering the M1 MAX for a mobile setup.
Well, a monster desktop rig vs a laptop... it was never going to be a fair comparison.
But it was fun, and it does bring some of the hype down to earth.
A fairer comparison would be the monster rig vs an M1-Max or M2 Mac Pro, whenever that gets released.
is redshift optimized for m1 chips?
Great video! Would love to see a this exact test but with m1 max 32gb vs 64gb ram specs head to head
But he said this exact test uses almost no system ram, and that it's also not pushing the vram. So probably another test would make more sense for such a comparison?
The beefy maxed out PC didn't really have any business in this laptop test, kinda disappointed.
Would have been great to see your render specs as that makes a huge difference. Are you doing GPU rendering with all of them or are you comparing GPU rendering on the PC with CPU rendering on the Macs?
Redshift is GPU only... don't think it taps the CPU at all for computation.
@@studiomiguel that's actually not true. It's ridiculously slower than GPU, but the CPU *can* be used to render in Redshift.
Bro can you please do a M1 Ultra Comparison?
thnx mate
The larger bucket will help even in the dual GPU. the performance loss from only being able to use 1 GPU at the end is far and away outstripped but the performance advantage from using the larger bucket.
Those Pc specs are crazy!
It would be great to share this scene, so we can compare on our machines.
single 3090 costs the same as M1 Max. So when it comes to Performance/Price ratio M1 Max beats it absolutely.
When the GPU market cools off, hopefully a 3090 can be found for $1500 for 3.5x the rendering performance of the M1 Max. Right now the prices of these cards is just outrageous.
laptop 3080 16G im 3 time faster them M1 Max and in 2 time cost less
You’ve got to lock in a good price for these GPUs when they first come out. I bought 3 for the price of a loaded Apple laptop. So no, it’s no comparison…
I do not agree - if you have like 1000 frames to render then 3090 is the way to go when it comes to timings. Also even a cheaper 3080 wil be faster than M1 Max chips.
@@baldoski But we are comparing the GPU card here with the entire laptop. So I still think the Performance/Price ratio is better.
Indeed, increasing the bucket size can really make a huge difference on M1!
Great tests, I try to tell motion designers just go for the MAX over the PRO for $400 it's a no brainer.
How about sketch with vray? have you test test it?
Vray on Mac only supports CPU rendering at the moment
@@RubmaLione Does it render ok with the new pro/max m1?
@@nyambe I don't use Vray anymore, but it should work without issues. Apparently an update for Vray is coming soon with native Apple Silicon support, so aside from a lack of GPU support, at least the CPU will be fully used.
Holy balls, batman! An integrated laptop GPU is half the speed of a 3090?! At _waaaaay_ less power?! No wonder Intel is sh!tting themselves...
I would love to see the 16 inch intel compared with the m1 ultra 16 inch
Is redshift utilizing hardware raytracing? I’m personally betting that Apple will put some of that capability in top end M2/M3, since video and AI acceleration is such a success
No, but Metal already supports raytracing. You don’t need dedicated RT cores to do RT, so Apple could technically work with render engines to implement this somehow. As of right now, RT won’t work until that’s figured out.
@@RubmaLione well, support and dedicated hardware support are 2 different things. There’s almost a order of magnitude difference in performance.
@@Notacet Of course, and we only really have the performance of dedicated RT cores to go off on at the moment. However, it is technically possible for ray tracing without dedicated RT cores, and it seems like Apple is looking into this.
Cool comparison. It would be nice to know the price of each of the computers.
The subtle downplaying of the M1 Max against a desktop computer with twice the RAM and dual 3090s. Try and work past you biases at least.
If we assume that with single GPU, it will take double the time, 3:30 minutes then M1 MAX is like 60% of RTX 3090. Coming iMac Pro and Mac Pro will have 2X and 4x cores. So the 4x Max version will be faster than this dual 3090 workstation.
The thing about that assumption is that GPU performance does not tend to scale linearly, so single GPU performance isn't quite double the time.
You are comparing 2 gpu’s vs 1 with shared memory. Deactivate 1 gpu on your pc in RS and rerun the test to make it apples to apples. Don’t think would be faster than the 3090 but would be a fair comparison.
3090's don't have pooled memory I have a 4x3090 setup and I really wish it did.
The reason is this is what you can get on each computer maxed out. This is a better real-world scenario than the one you're proposing. With a PC you CAN double the video cards, with eGPUs you can do even more. With a Mac, it's all integrated technology, so you get what you get. But it's really helpful to show, "with this computer I can get this, and this computer I can get this".
That's pretty absurd performance given the power consumption difference. 2x3090 would be about 700w or so. The M1 Max peaks at around 80w I believe(saw somewhere that the M1 Pro is rated for 60w).
So, to recap: a MacBook Pro isn't EVEN in the same league as a Ryzen Threadripper with dual RTX3090.
This surely must have been one of the biggest letdowns of 2021, and I thank RUclips for providing this valuable insight. Bullet dodged, money saved.
i have NV of your 3090s
None of these Apple computers will work for me in a production environment for clients. I need render times below a minute a frame. Multi-GPU PCs are still the way to go.
As far as bucket size goes the rule of thumb is bucket size set to small for CPUs and larger for GPU rendering.
Of course a full desktop PC would be faster than a laptop but my reason for getting one is that I need a laptop and I wouldn't dare go close to a windows laptop-those things are nasty
Is it just me, or is comparing a laptop with a single GPU to a desktop with two gfx cards, a bit of a silly comparison? Really, for this to be an equivalent, you should have used two laptops with team render utilising both chips, or just compared a single gpu. Of course a computer with two gpus will be faster than a computer with one gpu, that's kinda obvious.
Otherwise, very grateful for the video. I think this is what most redshift users have been after. I would prefer to know single high-end gpu vs the pro and max scores, but otherwise very helpful. Also yes, costs of the comparison system would be helpful too.
Probably worth doing this again in a few months when redshift have refined how it works on m1s. Right now we are just seeing out the box stats I think. As opposed to the Nvidias etc which have had years of tweaking at this stage.
Just my opinion tho. Thanks again for video!
The M1 Max has been great for the last two years. But I still LOOOONG for those Nvidia numbers. How does apple get there? Do they even want too?
I don't see a point of this comparison - Redshift is written mainly under CUDA technology and it is obvious that it will work much faster on Nvidia's cards (especially on such gpu like 3090) cause of the best optimization for these devices. Or if you wanted just to show how SLOW Apple's tech is then ok :)
redshift is based on GPU so It nothing to do with intel, it should be RX5500M VS M1MAX/PRO.
I'm so sick of PC, can't wait to go back to Mac
Why compare a MacBook against a dual dedicated graphics desktop PC. At least try comparing against a 3080 laptop which are over £2000 and more relatable. Pretty sure redshift isn’t designed for apple silicon - try blender.
suh dude...
What was the price of the pc? You get thumbs down for even comparing the pc to laptops. Too bad I can’t give you two thumbs down
It’s for reference. More than appropriate for a rendering test as Nvidia based PCs remain the gold standard for Redshift rendering.