- Posts: 27
- Thank you received: 0
3 Post Moderation Rule for New Boarders (25 Jan 2021)
New boarders will have their posts moderated - Don't worry if you cannot see your post immediately.
Intel Iris Pro performance (Retina Macbook Mid 2015)
- Montezumeflex
-
Topic Author
- Offline
- Junior Member
-
I'm still working on a regular MacBook Pro so I decided it's time to upgrade to a Retina MacBook Pro. I'm a reporter/director so won't be editing full-time on this Mac. I have always been using macbook pros with a dedicated GPU, but because of the increased prices of the touchbar Macs I'm looking at the mid 2015 MacBook Pros apple still has on sale right now.
I noticed a few comments on Alex4d's post that ran the BruceX benchmark on some Macs in the Apple Store:
MacBook Pro 13 inch, 2.7Ghz i5, 8Gb Ram, Intel Iris 6100 with 1.5Gb = 120s
MacBook Pro 15 inch 2.2Ghz i7, 16Gb Ram, Iris Pro with 1.5Gb = 60s
Imac retina 27 inch, 3,5 GHz i5, 8Gb Ram, AMD Radeon R9 M290X with 2Gb = 43s
These results look really impressive and I feel it won't be worth the 600euro increase to get a new model with dedicated GPU. I don't mind if the render times will be a bit longer, I just need the timeline to be workable while using Color Finale/Grading.
I havent been able to find a lot of users commenting on real life performance. I hope some of you could give me some advice if the Iris Pro is a safe choice.
Kind regards,
Martijn
Please Log in to join the conversation.
Please Log in to join the conversation.
- Montezumeflex
-
Topic Author
- Offline
- Junior Member
-
- Posts: 27
- Thank you received: 0
Please Log in to join the conversation.
But if you need to go mobile, a desktop is out of the question. My rMBP is a nice laptop, I'm very happy with it.
Please Log in to join the conversation.
Montezumeflex wrote: Thanks a lot for taking the time to reply! When you say it's nowhere near the power of a desktop do you mean exporting speed or general speed while working in the timeline?
I have a top-spec 2015 iMac 27 a top-spec 2015 15" rMBP, and a 2012 iMac 27. The GeekBench 4 numbers are:
[tt]
2015 iMac 27:
single-core CPU: 5126, multi-core CPU: 16277, GPU: 85101
2015 MBP:
single-core CPU: 4426, multi-core: 14615, GPU (M370x): 38371 GPU (Iris Pro): 27130
2012 iMac 27 (3.4Ghz i7-3770, 16GB RAM, GTX-675MX)
single-core CPU: 3991, multi-core: 12493, GPU (GTX-675MX): 28376
[/tt]
My BruceX numbers are below, using FCPX 10.3.2 and macOS 10.12.3. Note -- BruceX produces different results on a rendered vs non-rendered timeline. Almost nobody posting BruceX results specifies which they are using, or whether background rendering is on or off. Without this information the results have limited value. There can also be significant run-to-run variation, so ideally taking an average of 3 or 5 runs is best. The results also vary if done without restarting FCPX each time. The BruceX export should be to ProRes 422, not H264. Most people likewise don't specify any of those. BruceX is a mostly GPU-oriented test.
BruceX results, 2015 15" top-spec rMBP, 3-run average, restart FCPX and create new library each time:
Unrendered timeline: 53 sec
Rendered timeline: 23.5 sec
BruceX results, 2015 top-spec iMac 27, 3-run average, restart FCPX and create new library each time:
Unrendered timeline: 31.75 sec
Rendered timeline: 14.45 sec
Comments: Superficially the GeekBench4 numbers might indicate a top-spec 2015 MBP is nearly equal the performance of a top-spec 2015 iMac 27. However in actual usage the iMac feels faster and more responsive. The BruceX numbers are much faster. The iMac is also more pleasant to use because of the larger screen and it's quieter. The MBP fans spin up at the slightest usage. Testing by Max Yuryev found the 2015 MBP is thermally limited in many cases:
That said if you need portability for FCPX editing there's no real choice -- you can't easily lug around an iMac 27. The 2015 MBP is a good machine and it has built-in USB-A and SD card slots.
Re rendering performance vs export performance vs Color Finale performance: technically rendering is resolving the timeline edit and effects directives. Whether this is CPU or GPU-bound will depend on what's in the timeline and the codec used. Exporting (at least to H264) requires encoding, which cannot be GPU accelerated using OpenCL or any other GPU API. It can be accelerated using Quick Sync if the software supports that and fortunately FCPX does.
Color Finale is an effect. It uses the GPU but some effects do not. Some effects like Neat Video are configurable to use either CPU, GPU or a mix. I have not done Color Finale performance testing on my laptop, but it would probably be adequate using either M370X or Iris Pro graphics.
Much of this involves whether you're using 1080p or 4k, and whether you're willing to use proxy. With 4k everything is just much harder for both CPU and GPU. Likewise whether you edit ProRes or H264 has a big impact on CPU usage.
Please Log in to join the conversation.
I couldn't afford the rMBP 2015-16 model with dedicated GPU, so I ended with rMBP 2015 model , 15" 2.2 Ghz, IRIS PRO only.
Its the first time for me too with only one GPU.
IF you are using mostly 1080p footage, you will not have any problems in the timeline (I am speaking always in native footage, not proxy-transcoded).
BUT if you are planning to use it with 4K native footage, then its not ideal.
For only simple editing (cut to cut) and the quality to Better performance is just acceptable.
For anything else 4k... not so!
Please Log in to join the conversation.
- Montezumeflex
-
Topic Author
- Offline
- Junior Member
-
- Posts: 27
- Thank you received: 0
Please Log in to join the conversation.



