C20 not switching to discrete GPU on 16" MBP OSX Catalina
I've been monitoring C20 (20.0.1) with the Activity Finder and noticed that the GPU does not automatically switch to or engage with the Radeon 5500M but instead remains only with the Intel UHD. Performance is sluggish on the Intel UHD as the effects of the sliders respond with noticeable lag. The lag is improved when I turn Automatic Graphics Switching off in the System Preferences and force the system to use the Radeon, however I've also noticed that both the Intel UHD and Radeon 5500M are being utilized in the Activity Monitor - with the 5500M utilized to max and the Intel UHD at around 25%.
Despite the improved lag with the GPU switching off - C20 seems sluggish overall on this system for it's specs (i9, 32GB RAM, 5500M), however I don't have a reliable base-point to compare it with as my main work station is a Windows 10 based 9900K and I realize this is OSX on a laptop.
Despite the improved lag with the GPU switching off - C20 seems sluggish overall on this system for it's specs (i9, 32GB RAM, 5500M), however I don't have a reliable base-point to compare it with as my main work station is a Windows 10 based 9900K and I realize this is OSX on a laptop.
0
-
That is interesting. There are several threads here that report sluggish performance of Capture One 20 on the new 16" MacBook Pro.
The fact that the AMD Radeon GPU is not automatically engaged is something other users might suffer from without knowing.
Still strange why the overall performance is still well below expectations.0 -
I can confirm the above observation. I'm also on a specced out 16 inch Macbook Pro and Capture One feels slow and I cannot see it is ever using the GPU in the activity monitor. Processing Images is generally quite slow. Too bad C1 is not making proper use of graphic cards. In comparison Davinci Resolve seems so be so well optimized to take advantage of modern GPUs. Would love Phaseone to work only on performance optimization under the hood for the next version. I dont think we need new tools with every release. 0 -
Because, they insist not to update their code around OpenCL and Metal. They insist on sticking old version of OpenCL and they have a culture of blaming users for having under specced GPUs and high resolution screen i.e. 4K/5K. Below is from their website;
If working with high-resolution images and doing calculation-heavy adjustments we recommend the following:
Intel CPU with 4+ cores
16+ GB of RAM
Solid State Disk (SSD)
Dual (matching) Graphics card from NVIDIA or AMD with 4GB+ RAM per card
Calibrated monitor with at least 1920x1200, 24-bit resolution at 96dpi
NOTE: Due to the significant additional calculation overhead, systems using 4k/5k monitors will require additional resources in addition to the above-recommended spec (both in GPU and CPU power). For professional experience, we recommend at least doubling the specification above for these configurations.
They ask us to have 4 GPUs, I mean what ....... 😊
Also below what OpenCL1.2 supports;
OpenCL 1.2 support[edit]
(2011+) for some AMD GCN 1st Gen some OpenCL 2.0 Features not possible today, but many more Extensions than Terascale
(2009+) AMD TeraScale 2 & 3 GPU's (RV8xx, RV9xx in HD 5000, 6000 & 7000 Series)
(2011+) AMD TeraScale APU's (K10, Bobcat & Piledriver-based)
(2012+) Nvidia Kepler, Maxwell, Pascal, Volta and Turing GPU's (GeForce 600, 700, 800, 900 & 10-series, Quadro K-, M- & P-series, Tesla K-, M- & P-series)
(2012+) Intel 3rd & 4th gen processors (Ivy Bridge, Haswell)
(2013+) Qualcomm Adreno 4xx series
(2013+) ARM Mali Midgard 3rd gen (T760)
(2015+) ARM Mali Midgard 4th gen (T8xx)
As you see the code does not support new AMD processors. The processor might support the old version but how optimized the old code for the new architecture is a million dollar question 😊0 -
d42b wrote:
I can confirm the above observation. I'm also on a specced out 16 inch Macbook Pro and Capture One feels slow and I cannot see it is ever using the GPU in the activity monitor. Processing Images is generally quite slow. Too bad C1 is not making proper use of graphic cards. In comparison Davinci Resolve seems so be so well optimized to take advantage of modern GPUs. Would love Phaseone to work only on performance optimization under the hood for the next version. I dont think we need new tools with every release.
In the preferences under the options for Hardware Acceleration do you see a message that it is active?
I know nothing about video editing but just reading up a little I sort of suspect that in the modern era the processing overhead has been greatly reduced whereas the rather different needs of photo editing with larger files and a frequent requirement to view them at full size or produce high definition output files is generally increasing processing overhead on a "frame by frame" basis. So I'm not sure that the methods required (and therefore the potential performance available) are comparable. However I freely admit that I need to understand the differences better than I do currently.
As for under-the-hood changes - I tend to agree but the masses and the marketers probably would not.
That said the under-the-hood changes that might appeal to one person may mean nothing to the next person.
I certainly don't run on state of the art hardware but I don't personally perceive a problem with the performance of C1.
So improved performance might be nice but the only way any performance improvement would really give pay pack for me would be if my own mental and optical faculties and dexterity could be re-engineered. Even then I'm not sure it would really make much difference.
I think the majority of users, if polled, would say they were seeking changes in areas that were not deeply buried in the esoteric areas related to the the architecture deep in the body of process management and OS integration.
I also note, along with a few others, that this question about performance seems to be a more often raised subject for Mac users than it is for Windows users. Whether that is in fact true I can't say - it may be just a perception.
But if it is the case one wonders why. There could, no doubt, be several reasons.
For an "under-the-hood" performance enhancement to be a good marketing pitch would require a universal benefit to be on offer and I simply don't see that as being likely to apply. Not all users would care. Not all hardware would be able to show the benefits. That makes for a difficult sales pitch.
Show a fairly wide set of new function and features where 2 or 3 of them (at least) will likely be attractive across the broad user base and one gives the product an greater chance of appealing to a broad range of users and prospective users.
That's just the way humans are.
Grant0 -
Grant,
Your response is going against human's desires since the dawn of time. If we all enjoyed a leisural pace through life we'd be riding bicycles in Denmark.
Others are trying to reach Mars. We fly instead of take a boat. Some of us work against a clock. Some of us are using the hardware to its extremes. Some of us just need to get to work by 8 and taking a walk means a 4 hour journey -- so we take a bus.
If you're basically trying to say that performance improvements are a low priority and doesn't affect everyone and therefore shouldn't be prioritised (?) then I'd love to know.0 -
Well given the fact that most of the users on Mac forum are complaining about Catalo speed, GPU utilization, etc. that mean we DO care about under the hood changes not some UI changes. Also, if you don't use Mac or a new GPU machine, I don't think you can understand us, and all your posts won't make any sense. As I have said earlier, some users are defending PO pointlessly for unbeknown reasons to me. 0 -
Just providing a different perspective chaps.
One way to be fairly sure of a significant potential increase in performance would be to switch the application onto a server based database that was designed to cater for large number of "records" and deliver a speedy response.
Presumably one can still get Macs that offer such opportunities.
However, it's all a bit expensive. Still, if it saves some time it may pay dividends.
Perhaps this is a route that the Capture One team may take for the Enterprise version?
As for human desires - very true though I think a one way ticket to Mars is probably the most ridiculous objective that Humanity could aspire to.
Sadly the current dominant trend in social mobility and philosophy seems to be at odds with the concept of ever increasing reach and scope of human endeavour as it has been practised so far.
Humans are fickle creatures.
We say we want certain things and will 'buy' if they are provided. But the majority don't really mean it when they say that.
We want the speed, we want the new tools we want the quality - but many don't want to pay for it when the time comes.
It's all about what has visually changed and how little does it cost. Or at least it is in the mass market.
I'm sure there are élite users who could probably justify a business case for a higher spend for the faster, better, more comprehensive solutions they seek. Whether there are enough of them with enough of a need to make that combination work I don't now. There might be. But I bet they don't spend their time in forums discussing the whys and wherefores of Catalogues, OpenCL and the cost of annual upgrades.
FWIW the last time I found myself using software that was being totally re-written in order to add new performance and functional improvements (but also forced on them by MS and Apple technical requirements) the product failed to arrive before they ran out of money.
Many more people, notable Apple users of Aperture, have missed out on continuity of product for other reasons. Apparently. One assumes that lack of return on investment was still part of the problem.
These are all things to consider in my experiences. Your experiences may be different - which was the point I was making.
Grant0 -
Crikey,
And here I thought if you're the kind of company that tries to push the envelope of technology with regards to sensors & image quality, you'd be interested in having that technology supplemented well with the software you make for it.
But then again, here I stare at an XF150 that can't even perform live view correctly.
It's like being an F1 team and excusing improving the mapping software because it already goes 200mph. So we'll concentrate on the helmet design instead. What kind of sense does that make?0 -
photoGrant wrote:
Crikey,
And here I thought if you're the kind of company that tries to push the envelope of technology with regards to sensors & image quality, you'd be interested in having that technology supplemented well with the software you make for it.
But then again, here I stare at an XF150 that can't even perform live view correctly.
It's like being an F1 team and excusing improving the mapping software because it already goes 200mph. So we'll concentrate on the helmet design instead. What kind of sense does that make?
It depends where the envelope is being pushed to.
One wants the mass market, or what is going to be left of it by the time any really serious re-write could be delivered, one would presumably do what Apple did with Aperture and push out Photos instead.
If you want to specialize - as a company with significant hardware content well might - you could perhaps look at specialized markets that suit the hardware concept well.
For example Industrial and Cultural Heritage.
In the modern era the mass market seems less clear - especially for the camera manufacturers.
Even the big budget end of the Fashion industry seems to be showing signs of re-thinking it purpose and methods.
https://time.com/5760459/vouge-italia-illustrations-photoshoots-sustainability/
Your observation about F1 is interesting.
As one of the top car designers said a few years ago;
"The unfortunate thing about the changes coming in next year is that they are all restrictions" : Adrian Newey.
There have been recent comments that the organisers might be thinking about changing to 2-Stroke engines.
The massive vehicle manufacturer support it currently lives from is being forced to move to electric power for its day to day market and the business units are back to consolidating to survive, hardly a great time to find it necessary to pump a lot of resource into a short lived engine project with no obvious route to commercialisation.
Marketing via Helmet colour tweaks race by race may generate more useful brand awareness for manufacturers and sponsors alike.
Plus they will need to deal with the "Carbon" thing.
It will be interesting to see how it changes in the next decade. In fact whether it survives in anything like its current form, especially in its traditional geographic market.
Whether, today it is still a sport, a successfully extended sport compared to its past development or mainly a marketing exercise for the sponsors and some Governments, we will probably discover as it is goes through its next phases of change.
The same as any other business.
Grant0 -
My analogy is about winning the race and ignoring the go faster stripes, not about the politics and restrictions put in place for competitions to be fun and fair to watch. 0 -
photoGrant wrote:
My analogy is about winning the race and ignoring the go faster stripes, not about the politics and restrictions put in place for competitions to be fun and fair to watch.
Well ok, but that's very 60s and 70s.
Since somewhere around the early 80s it's been more about the money and eventually TV coverage for the teams and sponsors.
For the drivers the helmet is the only things the cameras really see when they are racing.
Cutting edge is fine so long as it does not compromise reliability too readily.
Some of the modern controls - F1's digital age rules - allow little latitude until the sport's owners and managers agree to changing core technology. The equivalent of Sony releasing new Sensor tech. Back in the day one could change film stock. Today to change sensors one has to change the camera body.
The modern era makes on much more difficult just to change stuff - in many different fields. Not ujust F1 and photography.
Grant0 -
SFA wrote:
photoGrant wrote:
My analogy is about winning the race and ignoring the go faster stripes, not about the politics and restrictions put in place for competitions to be fun and fair to watch.
Cutting edge is fine so long as it does not compromise reliability too readily.
Exactly my point.
Supplement your hardware to perform at its best and ignore the new features. That's what we want.0 -
I am wondering what the current situation is in regard to graphics support and Capture One 20. Have there been any improvements over the last year? I am also wondering if choosing the 5600M option would result in better performance on 5K screens.
0
Post is closed for comments.
Comments
13 comments