It’s an interesting approach, but ultimately flawed. We know that the nVidia cards have dedicated RT cores which are known to be superior at rendering type tasks.
This is similar to the hardware accelerators for video encoding/decoding on the M1 generation chips. You could measure the time to render video and the power used and conclude that the M1 chips have amazing efficiency. The 16" Macbook Pro has 21 hours of video streaming battery life. Amazing!
So let’s compare video encoding tasks of M1 with 3090 and we get a different story. But honestly, Apple could have made a chip with no GPU cores and only media engines, and still achieved this result.
Back to the point, nVidia and AMD have been chasing the highest benchmark scores for years, and their approach has mostly been: throw more power at it. It’s refreshing to see Apple to come at the problem from the opposite side of the story.
Well, I got a 6900XT coming in tomorrow and we’ll finally find out if a Hackintosh with that GPU is the absolute fastest (and possibly most affordable) way to render with Cycles in MacOS.
I fully agree - amd and Nvidia only get better with more power but the technological limits remain.
Apples tech also is interesting because it can be a basis for lower power computation too
I agree that for high end gpu render Apple is far off but when it comes to lower power computation apples has a tech now that offers them to explore form factors Intel and amd can forget about
Remember those useless atom netbooks as an example ?
I nevertheless would not say that Nvidia or AMD products are “inefficient”.
Nvidia do this power increase (and Intel always did for CPUs !)
just because being a low hanging fruit to stay ahead of AMD.
Efficiency was never on top of their list in the past of course.
But after Apple’s ARM release, Intel is in deed going the efficiency track
too.
For how it is now, I think Bullit isn’t that wrong overall. Maybe not in numbers
but AMD and later Nvidia did a more effective hardware Raytracing.
The overall efficency to Render a whole image or movie may be still a bit better
for Apple - but a very large render time penalty. Where the one or other may
finally choose to accept a bit more energy cost to keep timelines.
Looks like Apple has lots of room to gain speed by Software Optimization.
But if that always means that they force developers to their own standards
and island solutions like forcing to Metal, which may have a 5% advantage
over Vulkan, insisting on making use of incompatible Tiled Memory and such
(in my eyes just) marketing and anti competition attempts, it might not really
happen in the end.
(Like with my double D700s ?)
And on the other hand,
Apple with ARM went the efficiency trip for mobile and handheld devices first,
with is now also getting more important for desktops for ecological facts.
And Apple really worked around many other common x86 bottlenecks,
which is great !
But now as Intel does with CPUs, if Nvidia will get interest in marketing Efficiency,
for any reasons, I would assume they would be pretty capable to nevertheless
bring GPU performance that is ahead of Apples.
And I am sure also AMD has ARM experience and it will not be easy for Apple to
follow their current development speed.
That may also be a reason why Apple went that anti-compatibility way after 2012.
To avoid fair comparisons ?
Before 2013 I was pretty agreeing with all Apple Intel road maps AND Marketing !
Apple was performant, compatible and offered things you would not find elsewhere.
You just contradicted yourself. They became inefficient to get the best benchmark.
Barely. They just released a chip with efficiency cores, many years after Apple, and it still sucks power like nothing else in the high end configurations.
How is this any different from Microsoft and DirectX?
On what basis do you think that nVidia’s chips will become more efficient than Apple’s. Where do you get that reasoning from? What evidence do you see?
Are you talking about Metal again? Why not discuss this in light of DirectX that Microsoft have been pursuing for 26 years. Do you think they developed that technology also to avoid fair comparisons. Or maybe they developed it to gain a performance edge.
It’s not a contradiction, though it could’ve been worded a little better. In this case, Apple has gone for a more elegant, holistic approach to their hardware, which ended up punching well above it’s power envelope relative to their competitors. Their efficient design provides more power per watt. Intel and Nvidia, on the other hand, are taking a brute force approach to gaining performance by throwing MOAR at it. They ended up taking the performance crown, but at the cost of power hungry, hot running pieces of hardware that require a number of noisy fans to keep cool. It’s powerful, but inefficient.
AMD at least is combining the high power-consumption with major design changes from generation to generation. RDNA3 for instance is going to a full MCM design, so AMD is not using brute force so much as pushing the wattage purely to maximize the performance uplift over the previous generation.
The company also builds power saving plans into their hardware as well, so those who express concern and worry for Mother Earth whenever the state says so can run their machines in eco-mode for the purpose of feeling better.
It’s not merely about the hippie dippie happy good fun time earth love. There are other costs to be considered. Like the wear and tear a hotter machine will accrue relative to a cooler running one, or how your machine will heat up your study, which really sucks when you live in an old house without central air, so you have to crank a window unit AC and a tornado fan to keep from dying of heat stroke, which in turn leads to savings on your power bill for not having to run that AC 6-12 hours a day alongside your 1000w PC.
To simplify, a computer that can do just as much with less is an overall better machine for a variety of reasons.
Comparatively is the key word here. No one’s saying that Nvidia produces badly engineered GPUs. Anyone who claims such is probably starved for attention, and wants to rile up people on a messageboard.
But compared to what Apple’s produced, and their response to such being to take their already good cards, and up the wattage requirements, it’s just an inelegant solution.
Nvidia has a totally different philosophy (dominate the GPU market).
But Nvidia and AMD build high end GPUs (or CPUs) for decades,
they know what they do.
They just have, for now, a different target group.
So as an abstruse example, if the EU says one day that you must not sell
anything GPU like that needs more that 100 Watt.
Do you really think after 1.5 years of adaption, Nvidia’s solution would be
less FPS than Apple’s GPU cores ?
I think they are just in a different market environment and Apple is still
a niche market currently, not important to really compete with.
I wouldn’t say Apple is a niche market exactly. Rather, it’s niche for our particular niche.
If you’re into video compositing, photography, or graphic design, you’re more than likely using a Mac. Apple is practically the standard there, and it’s where they tend to focus most of their attention. But for us, Macs are merely so-so.
What I am currently annoyed with is Apples compatibility with older hardware.
I meanwhile have 3 still working multi purpose printer+scanner units here
that may print easily but will not ever scan again with my M1 Mini.
They won’t also with my trash can as I run it on 12.4 too.
They won’t even run on my father’s 20" iMac 2012 - as unfortunately it updated
even to Catalina.
WTH, it only runs still on my mother’s white plastic iMac from 2006 (OSX 10.7.5 yeah !)
Of course all these devices run flawlessly on Windows 11.
“But I do not want to start Windows - to be able to scan my office and tax stuff
once in a while …
Do I really need to replace all existing devices or reactivate my 2009 Mac Mini Server,
2006 plastic iMac 20” or even the cheese grater Mac Pro 2.1 ?
Cannot agree more. This is the main reason I bought a mac studio.
Even on idle my PC uses vastly more power and generates more heat than my mac studio. I don’t have AC at home (not possible to install in my apartment due to strata regulations) and last summer I literally couldn’t work on my PC because of the heat it generates.
I am not sure.
Wasn’t the majority leaving with the FCX debacle ? Did they all come back ?
And aren’t the majority of YT content creators on Da Vinci anyway ?
Compositing, really? (I have no clue)
Photography yes, I think a good amount will be Mac affin. Although they will
just work in Lightroom.
(Man, am I annoyed that they deprecated Aperture …)
For serious 3DCC I assume the smart majority left the boat at least starting 2014.
2D Graphic Design, I am pretty sure (from my own experience with the graphics
people - yes
