Q's re performance (MBP faster than a higher spec'd PC)

[longstock]

Hi, I have recently purchased Affinity Photo and have been quite happy with what I can achieve with it. I'm a photographer and I normally go through the workflow of Iridient X-Transformer --> Lightroom --> Affinity Photo on my PC. Affinity Photo is only used for stacking and tone mapping purposes using TIF files generated by Lightroom. More and more I am finding myself going through the same performance pain points where I am basically sitting there for minutes waiting for Affinity Photo to do something (which I will get into below). Last night I downloaded the trial version of Affinity Photo on my Macbook Pro and did a few tests. Turns out Affinity Photo consistently performs up to 3 times faster on the MBP using the same set of files.

 

I am hoping someone with more experience here can give me some advice on how I could improve the performance of Affinity Photo on PC.

 

Below are the specs for my PC and MBP:

PC

Intel i7-4790K OC'd @ 4.40 Ghz (with a large Noctua air cooler)

16GB DDR3 RAM @ 2400 MHz with XMP profile 1 enabled (G.Skill Trident-X 8Gx2 F3-2400C10D-16GTX)

Nvidia 1080 GPU (a Gigabyte OC model)

 

Macbook Pro (Mid 2015, 15'')

Intel i7 @ 2.5 Ghz

16GB DDR3 @ 1600 MHz

AMD Radeon R9 M370X + Intel Iris Pro 1536 MB

 

As you can see the PC's spec is significantly higher than the macbook's. It is winter here at the moment (<0C outside) and the system is well ventilated (has no problem in 144hz gaming) so there is no thermal limitation.

 

The photos that I edit generally sit on a NAS. The PC is connected to the NAS via ethernet and the MBP via wireless AC.

 

I have uploaded a sample of 4 RAF files generated by a Fujifilm X-T2 onto Dropbox. Here's the link.

 

To test the performance. I always repeated the following steps and timed each one on each machine:

1. Create a new stack in Affinity Photo using the 4 RAF files
2. Set Live Stack Group blending to minimum
3. Rasterize the Live Stack Group
4. Tone map the pixel layer

 

Here are some observations I made from the tests:

• Affinity Photo loads up faster on the MBP than the PC.
• Creating a Live Stack Group is consistently quicker on the MBP.
• Live Stack group blend mode live preview is a lot faster on the MBP than the PC.
• Live Stack group rasterizing is significantly faster on the MBP than the PC (10sec vs minutes)
• PC tone mapping consistently takes 00:50 ~ 1:30 to process.
• MBP tone mapping consistently takes 00:20 ~ 00:22 to process.
• CPU heavy tasks will utilise 100% of the CPU on the MBP but only 60~80% on the PC in some cases.
• There is no CPU throttling on the PC at all (winter + large cooler)
• GPU utilisation never goes above 1% on the PC.
• There is very little disk/network bandwidth utilitsation during any of these tasks.

 

All of the above are true regardless of which renderer I use on PC and whether I have hyper-threading enabled or not.

Having the RAF files on SSD's instead of the NAS made no differences either.

I have tried the latest 1.6 beta and it made no differences in performance on PC using the same steps above.

 

I don't understand why a PC with much better hardware is being out performed by a Macbook by 3 times as much and generally have a much smoother experience...

I am getting to the point where I am seriously considering using the MBP as a photo editing machine but that would mean some significant investment into repurchasing software licenses and some new hardware so I could actually dock the MBP to my PC monitors and input peripherals and I would rather not have to do that. Any help is much appreciated!

 

EDIT: more info on the spec's and additional notes on the observations.

[R C-R]

Hi longstock & welcome to the forums.  :)

 

Regarding the "better hardware" aspect, there are some things to be aware of that may (or may not!) account for the difference:

 

Beyond the basic specs you mentioned, several other factors make direct comparisons between Macs & PC's (& between different PC's) more difficult than it might seem. For example, cooling systems must be adequate to prevent CPU throttling, resulting in the CPU running at less than its rated speed for sustained or short term operations. There are two memory specs that affect overall performance, clock speed & CAS latency ("CL"). Speed is usually easy to determine from spec sheets, but latency figures are rarely published for off-the-shelf computers, & not all motherboards are loaded with chipsets that take advantage of the lowest CL memory modules, or must be configured to do so by the user. I/O bus type (PCI, PCI-X, or PCIe), the revision number implemented, available lanes, & other arcane "under the hood" architectural bus features all combine to affect performance in various ways.

 

Because Apple sometimes uses parts with no direct "COTS" (commercial off-the-shelf) equivalents, does not make information about that available publicly, & has a long history of participating in chip design, we can only guess about how that affects performance of its products vs. those of PC vendors or particularly vs. user-built PC's.

 

For this reason, it might be useful for Affinity PC users to include a bit more info about their systems, like brand & model info if they use an off-the-shelf branded computer or motherboard & related info if they built their own system.

[longstock]

Thanks @R C-R. You are right. I have included additional information on the spec and the environment the systems are running in.