SEARCH
diglloydTools™
$220 SAVE $130 = 37.0% Western Digital 16.0TB Western Digital Ultrastar DC HC550 3.5-in… in Storage: Hard Drives
|
|
|
|
|
|
|
|
|
|
2023 MacBook Pro M3 Max: Adobe Camera Raw AI Denoise + Enhance Details
Related: 2019 iMac 5K, 4K and 5K display, Adobe Camera Raw, Adobe Camera Raw AI Denoise, Adobe Camera Raw Enhance Details, Apple 2023 Mac Pro M2 Ultra, Apple MacBook Pro M3, Apple Silicon, Artificial Intelligence, GPU, iMac, iMac 5K, laptop, Mac Pro, MacBook, MacBook Pro, Macs, noise, Other World Computing, Photoshop, thermal throttling, video tech
Please buy your gear at B&H Photo and OWC/MacSales.com using any link from this site.
Consult with Lloyd, to design a complete system for photography or similar including backup and data safety
MPG tested both the 14-inch and 16-inch models Apple 2023 MacBook Pro M3 Max 128GB /4TB, model Z1AU002AK and model Z1AF001AN. See all Apple MacBook Pro M3 Max.
Consult with Lloyd to design a complete system for photography or similar including backup and data safety.
About
Adobe Camera Raw AI Denoise is a GPU-intensive task, the most intensive camera RAW file processing use case out there, taking 10X to 20X as long as normal RAW-file conversion.
The AI Denoise “test” is not a synthetic test. Rather, it’s what I do as a photographer for every RAW file I process. It doesn’t get any more “real world”. Luckily, I typically run it on fewer than 10 files at a time. However, some focus stacks involve up to 40 images.
If below are out of date: view current Mac wishlist and all current OWC wishlists.
Results
Fan noise of M3 Max machines in high power mode is obnoxious to the point of unacceptable.
M3 Max has 40 GPU cores. M2 Ultra has 60 GPU cores. M3 Max should therefore be expected to take 50% longer.
The 2023 Mac Pro M2 Ultra with its 60-core GPU easily outperforms the M3 Max, and without its fan noise changing noticeably.
Discussion continues below...
Thermal throttling?
Both 14-inch and 16-inch M3 Max have a 40-core GPU. The Mac Pro M2 Ultra has a 60-core GPU.
16-inch 2023 MacBook Pro M3 Max takes 78% longer than the 60-core Mac Pro M2 Ultra, rather than the expected 50% longer.
14-inch 2023 MacBook Pro M3 Max takes 91% longer than the Mac Pro M2 Ultra and 7% longer than its 14-inch sibling.
The outrageous 58 db fan noise along with the unexpectedly slow performance of the M3 Max strongly suggest thermal throttling.
The 16-inch 2021 MacBook Pro M1 Max with its 32 GPU cores takes 31% longer than the 16-inch M3 Max , but runs with very low fan noise.
Fake GPU performance?
On short-running benchmarks of this review, the M3 Max generally outperforms the M2 Ultra, at least on a per-core basis and when not all GPU cores can be used.
Claims of GPU performance for M3 Max appear to be marketing bullshit when it comes to GPU-intensive tasks. Because once the M3 Max heats up after 30 seconds or so of 90+% GPU utilization, a big chunk of performance goes missing in high power mode. And the fan noise goes batshit-crazy because the 47-watt GPU is dumping enormous heat into a tiny space.
Is this “gaming the system”? The test footnotes from Apple make no mention of ambient temperature, test duration, etc. You get impressive benchmark numbers, but real-world performance when most needed drops considerably, as proven here with not just one, but two M3 Max laptops.
What happens to M3 Max GPU performance at 80°F or 90°F instead of the ~65°F test environment? I frequently work in such warm environments in my Sprinter van when in the field.
Premature failure is likely, because dust steadily coats the already hot-running chipset. A reasonable viewpoint here is that Apple favors form over function. Proper design would have meant a larger chassis (thicker or larger somehow, for more airflow). Presumably that was unacceptable for a laptop, and hence we have designs that struggle to stay cool with what is essentially a high-powered desktop chipset. That it can work at all is impressive!
High power vs low power
The 2021 MacBook Pro M1 Max outperforms both M3 Max laptops in their low power modes.
Forcing the M3 Max to low power mode results in a 45% and 68% longer running time for the 16/14 M3 Max models. That’s interesting; it strongly supports the idea that the 14-inch model suffers from thermal throttling since even in low power mode it is slower than its 16-inch sibling.
Comparables for Intel machines not available—the 2019 iMac 5K crashed instantly every time the test was tried, and the 28-core 2019 Mac Pro was unavailable. Probably the 28-core 2019 Mac Pro would take 3x to 5X as long in this case.
Seagate 22TB IronWolf Pro 7200 rpm SATA III 3.5" Internal NAS HDD (CMR)
SAVE $100
