Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1110
Article Thumbnail

Lightroom Classic CC Version 7.2 Performance

Written on February 8, 2018 by Matt Bach
Share:

Introduction

Software developers are often very closed-lipped and secretive with patches and updates before they go live, but the Lightroom team at Adobe has given us the go-ahead to publish some benchmark results for the upcoming 7.2 update to Lightroom Classic CC. To be clear, at the time this article was published this build is not available for public download (so don't go hitting refresh on the Creative Cloud app) but it should be available soon.

Since we are testing with a non-public update, there are unfortunately no published update notes quite yet. There should be at launch, but one thing we can tell you is that based on our testing this new version improves performance dramatically - especially with higher core count CPUs. Like always, the tasks we will be testing in this article are:

  1. Importing images
  2. Exporting images
  3. Convert RAW to DNG
  4. Generate 100 Smart & 1:1 Previews
  5. Create HDR image
  6. Create Panorama image

If you would like to skip over our test setup and individual benchmarks, feel free to jump right to the conclusion section.

Test Setup

Listed below are the three test platforms we will be using in our testing. We opted to test a number of modern platforms along with a wide range of CPU architectures and core counts.

The images and settings we used in our testing were:

Test Images

18MP (5184x3456)
Taken on a Canon EOS REBEL T3i

HDR settings

5x 18MP RAW (5184x3456)

Auto Align, Auto Tone, No Deghost

Panorama settings

11x 18MP RAW (5184x3456)
Merged to 187MP (48085x3898)

Spherical Projection, No Crop

Import 100 Images

Lightroom Classic CC 7.2 vs 7.1 Benchmark Importing Images
Since importing images is typically the first thing you do, we thought that this would be a good place to start our testing. Overall, this test is not very interesting with minimal difference in performance between the two versions.

One thing we do want to make sure is clear is that this test is measuring the time it takes to both copy images off a USB device and add to the Lightroom catalog. So technically this is not a straight import test. In future tests, we are planning on solely testing the time it takes to import images that are already on the local drive since using a USB device introduces a number of variables that can affect performance.

Export 100 Images

Lightroom Classic CC 7.2 Benchmark Exporting Images
Exporting images is an area where we have expected Adobe to make improvements for a while now but the initial release of Lightroom Classic CC didn't give us quite what we were hoping for. The 7.2 update, however, delivered in spades. The lower core count CPUs like the Core i7 8700K and Core i7 7820X only saw a marginal 6% and 22% increase in performance respectively, but the higher core count CPUs (especially the Intel CPUs) saw huge performance gains.

For the higher core count CPUs, the Intel Core i9 7900X about a 32% increase in performance while the Intel Core i9 7940X saw a terrific 55% increase. This is a pretty absurd performance improvement and for users that export a huge amount of images may finally make higher core count CPUs (especially those from Intel) worth the investment. Unfortunately, the AMD Threadripper 1950X didn't get as much of a boost as we expected, coming in at just 19% faster in the new version.

Convert RAW to DNG

Lightroom Classic CC 7.2 Benchmark Convert to DNG
The results for converting RAW images to DNG is not nearly as exciting as the exporting results, but there was still a small performance gain across the board. This time, however, the lower core count CPUs were the ones that saw the biggest performance gain.

Generate Smart Previews

Lightroom Classic CC 7.2 Benchmark Generate Smart Previews
When Lightroom Classic CC was first launched, Smart Preview generation was one area that saw a large performance increase. Apparently, there was still plenty of room for improvement as we saw even better performance in the new 7.2 update. Interestingly, the performance gains appears to be focused around multi-core performance as the higher core count CPUs are the ones that saw the largest benefit. Where the Core i7 8700K 6 Core only saw a 6% increase, the Core i7 7820X 8 Core was 18% faster in the new update. Even better, the Core i9 7900X 10 Core and Core i9 7940X 14 Core saw a 50% and 67% performance improvement respectively.

One trend that we are starting to see is that while the AMD Threadripper 1950X 16 core is definitely faster with the new version of Lightroom Classic CC, the performance gain is not as large as what we are seeing with the Intel CPUs as it is only 44% faster.

Generate 1:1 Previews

Lightroom Classic CC 7.2 Benchmark Generate 1:1 Previews
While generating 1:1 Previews is faster in the new update, unlike exporting and generating Smart Previews the performance gain was pretty consistent across all the CPUs we tested. There was of course some variance, but overall each CPU saw between a 20% and 25% performance improvement.

Generate HDR Image

Lightroom Classic CC 7.2 Benchmark HDR Image
Creating HDR images is also faster in the the new 7.2 version, although the performance gain is a bit inconsistent. Interestingly, since the initial release of Lightroom Classic CC saw a small performance drop when creating HDR images, this puts the performance roughly back to where it was in the last version of the old Lightroom CC.

Generate Panorama Image

Lightroom Classic CC 7.2 Benchmark Panorama Image
Wrapping up our testing is creating Panorama images, which is pretty significantly faster in the new 7.2 update. There is still no additional benefit to having a high core count CPU for generating Panoramas, but we did see a great 30-40% performance improvement across all the CPUs we tested.

Conclusion

Starting with Lightroom Classic CC and continuing in the new 7.2 update, one thing is very clear: the Lightroom developers are focusing heavily on improving performance. To get a broad overview, lets take a look at each task we tested to see how the new 7.2 update compares to the previous version:

Lightroom Classic CC 7.2 vs 7.1 Benchmark

Overall, the performance gains in the new 7.2 version of Lightroom Classic CC can be split into three categories:

  1. Minimal performance improvement - Importing and Converting RAW to DNG fall firmly into this category with either no or minimal performance gains. However, our import benchmark is likely being impacted by the fact that we are doing a copy and import from a USB device which we will be addressing in future testing.
  2. Substantial gains across all CPUs - Generating 1:1 Previews, creating HDR images, and especially creating Panorama images fall into this category with all the CPUs we tested seeing a significant performance improvement.
  3. Substantial gains focused on higher core count CPUs - Exporting and Generating Smart Previews are the two tasks where improving multi-core performance was clearly a priority. These tasks are where we saw the largest performance gains in our testing, although they were focused on the higher 10+ core CPUs.

This new update is substantial enough that it will likely influence what CPU you would use in a new workstation for Lightroom Classic CC. If your workflow involves heavily modifying a small number of photographs then the Core i7 8700K is likely still the best CPU choice, but if you export a large number of images then investing in a higher core count CPU may be worth it.

Lightroom Classic CC 7.2 CPU Performance Benchmark
As we stated earlier, for most Lightroom users the Core i7 8700K is still the best overall CPU choice as it gives great performance in a wide variety of tasks. However, this new 7.2 update is a huge boon for users that work with large numbers of photographs. While an Intel Core i9 7900X or 7940X is certainly an investment, being able to export hundreds or thousands of images twice as fast as you can with a Core i7 8700K could be a massive time savings for some users.

We have to hand it to the Lightroom Classic CC developer team - since Lightroom Classic CC initially launched just 4 months ago, it has been clear that improving performance is a major focus. Just for fun, we went back and compared the results from this new 7.2 version of Lightroom Classic CC to the old Lightroom CC 2015.12 using the Core i9 7940X. Comparing the two versions, exporting is now almost twice as fast, generating 1:1 Previews is about 5x faster, and generating Smart Previews is an insane 7x faster! This is an incredible improvement and a trend that every Lightroom user should be very excited about.

Tags: Lightroom CLassic, Coffee Lake, Skylake-X, Threadripper
Jakub Badełek

Matt are you planning to throw in very popular Ryzen 7 processors? I wonder if they gained any advantage against Intel with this new release. Also, dPreview reported supposedly huge boost in Importing images - do you have any thoughts on why you didn't see any improvement here? the difference seems to be gigantic... quoting dPreview: "When importing 130 Raw files from the Fujifilm X-T2 (7.6GB in total) and building "Standard" previews, we saw a major performance boost in LR Classic CC 7.2 on our quad-core 2015 MacBook Pro. Roughly 80%, in fact." - I know, they threw in Building Previews... but they did that on a "quad-core Mac-book"

Posted on 2018-02-09 14:43:14

That's actually my mistake - our testing right now is importing from a USB device, so it is actually a Copy+Import test. I forgot to make note of that, but I added text just now to explain that. We are actually in the middle of completely revamping our Lightroom testing, and in future tests we are planning on solely testing the time it takes to import images that are already on the local drive since using a USB device introduces a number of variables that can affect performance.

Edit: Sorry, forgot about your Ryzen question! At the moment, we don't have plans to test Ryzen in Lightroom. Our previous testing showed that Ryzen is not that great for Lightroom ( https://www.pugetsystems.co... ) and given that the Threadripper CPU didn't see anywhere near the performance gains of the Intel CPUs, it is pretty safe to assume that Ryzen still isn't a great choice. If we had unlimited time I would love to include Ryzen, but our time is a very limited resource so we have to prioritize testing platforms that are the most likely to be good performers.

Posted on 2018-02-09 17:09:36
Jakub Badełek

Many thanks for your honest reply! Very good point on Ryzens, i was just wondering how are they catching up, but you're not a testing and reviewing site ;-) (a pity though...).

Looking forward to the revamped tests next time!

Posted on 2018-02-11 10:35:21
cnccnc

Matt - as a fairly heavy Lightroom user, one of my biggest complaints is perceived speed in the Develop module. Lightroom tends to be slow making changes to images that already have a lot of modifications, including lots of use of the adjustment brush and spot cleaning. Speed may also vary based on monitor resolution. Secondarily, scrolling to the next image in Develop and switching from Library to Develop to edit an image can be slow.

Importing and Exporting images I can start and walk away. I have to sit and watch Develop while at the computer, so I would contend that this is the most important factor in evaluating Lightroom performance. Develop module performance may be harder to quantify than the others, but you're basically the leaders in testing Lightroom performance, so as you modify your testing, please take Develop into consideration.

Thanks!

Posted on 2018-02-13 18:31:32

The develop module is definitely really high on our to-do list. The problem is that almost everything in that module lacks a clear "I'm done" kind of notification that we can use to benchmark automatically. The good news is that in a few weeks I will be talking a bit with the Lightroom development team to get their assistance with figuring out how to benchmark things in the develop module. In fact, my current plan is to move to a largely plug-in based benchmark that we can put up for public download. We even just the other day got a new batch of images to use for our testing that we have full permission to distribute (Including some insanely high-res stuff from a Nikon D850) that I think will really improve our benchmark.

Posted on 2018-02-13 19:04:08
Byron Du Bois

Ye, I'm using the Nikon D850, and the previous version of LR 7.1 absolutely cried help. Exporting was basically a bottleneck. Machine was so so sluggish. Develop tab took forever to render back to what you were exporting. And I'm using an 8700k 16gigs of memory 960 pro Nvme itx MSI gaming pro board.

I was really happy to see this update. Because to be quite honest. All the other copy cats just can't compare to what LR is as a complete great workflow that I'm use to using.

LR must just have more features with regards to performance and let the user decide how much via the use of a slider, Like they have in Photoshop with regards to how much memory you want to use.
Like 50% 75% or 100% CPU core/thread usage.

I would really also like to see Adobe fix the stupid issues of when pushing the alt key...and then your tabs at the top get selected. I have to use a third party app to suppress this, (APassistant)

Posted on 2018-02-13 22:50:35

Thanks for your great review. I'm a super-heavy Lightroom user, importing 5000-10 000 photos per shoot and exporting 200-300 per edit. My greatest concern with Lightroom is actually inconsistent and slow culling behavior. Sometimes its OK, but some photos are slowing time (loading...). It would be nice if you are able to quantify culling speed, going through 1000 images in Library module or smth. I have to use PhotoMechanic to make my first selection and then I import to LR.

Secondly, I really believe that core-heavy tasks have also quite good improvements on Ryzen 7. Just thinking, seems logical.

Posted on 2018-02-14 22:26:11
JohnN

Hi Jack. Have you tried importing using “Embedded & Sidecar” You select it at the top of the right-hand panel in the import dialog where you normally choose 1:1, Standard or Minimal previews. This relatively new choice is said to be faster than PM although I haven’t timed it personally. It uses the full size preview created by the camera within raw files or the full sized sidecar when shooting raw+jpeg. I found it a great timesaver, with culling becoming available almost instantly.

Posted on 2018-02-16 10:28:07

Yes indeed, this option came w 7.1 if I'm correct. At first I was super happy that I can leave PhotoMechanic aside and use LR Classic CC only, but it still was not consistently fast. In PM you literally can go forward and back in milliseconds and it is essential in order to make painless first culling based on your first instinct. In LRC it still had some unexpected delays, even using embedded preview.

Posted on 2018-02-16 13:45:17
JohnN

OK cool, as long as you know about it. It’s really fast with my Canon 5Dmk4 files as long as I stay in the Library Loupe. Trashing the rejects and building previews for the picks to be edited seemed a breeze to me.

Posted on 2018-02-16 15:53:16
Dave Stromberger

Were these tests done on systems with the Spectre/Meltdown OS and BIOS updates installed?

Posted on 2018-02-10 00:18:36

Meltdown/spectre is a pretty big mess at the moment. Most motherboard manufacturers launched BIOS and/or Firmware updates to fix the issues, but they have caused a ton of problems. Linus Torvald (creator of Linux) used some... strong language... describing what he thought of the fixes ( https://techcrunch.com/2018... ). Microsoft even launched a Windows update blocking the firmware fixes since it was causing reboot and stability issues. At this point, a lot of motherboard manufactures have pulled down the meltdown/spectre fixes due to all the problems surrounding them.

So at the moment we are continuing testing without the "fixes". In our internal testing the performance difference was minimal with the fix so I don't expect them to impact Lightroom users very much in most cases.

Posted on 2018-02-10 00:31:50
pillybilly

Patch and fix "mitigate" the problem don't resolve the problem at all. On linux the fix will cause to lost a general 16% of performance on Sky-lake cpus, on coffelake less.
The truth is until anyone will not harm thousand of computers there will be no patch, then probably you will loose till 50% because they will have to deactivate branch prediction totally...

Posted on 2018-02-11 13:56:57
Brian J.

As I read in another post Adobe worked with Intel to improve performance, but not AMD. Seems to me that the author and Adobe have lost objectivity. Not sure whats happens behind the scenes with either, but I wish people in our industry would hold Adobe accountable for not making sure their code was optimize.

Posted on 2018-02-11 22:19:07

Without knowing everything - which only the developers at Adobe know - it is really hard to know the full story. It sure looks like Adobe spent more time optimizing for Intel products, but it is entirely possible they tried to work with AMD and didn't get the kind of collaboration they were looking for. Or it could be that the type of things they are doing is simply better suited to the architecture of Intel CPUs. Again, no way for anyone outside of Adobe to know for sure.

All we can do in our testing is to report the results as found in our testing. Maybe AMD could perform better if the software was optimized for those CPUs, but the fact is that with the current version of Lightroom as it is today, Intel CPUs simply offer better performance. If that ever changes, we will be very happy to report it since competition between Intel and AMD is always a great thing for consumers.

Posted on 2018-02-12 01:05:21
Laca Port

Thanks for that!
In the conclusions, you specify that the first bar graph is showing "% faster than 7.1" but the following text implies that those numbers might (sadly) relate to something else. If you really meant to say "% faster than 7.1" this would mean that a value of 100% stands for a x2 improvement, i.e. the job takes 1/2 of the time on 7.2 than on 7.1. However, the following text implies that the value 100% might, in fact, stand for a case in which there was no difference between the performance of the two Lr versions. Which of the above interpretations is correct? what did you mean? If the second interpretation is the correct one, the legend isn't correct.

Posted on 2018-02-12 09:56:31

Opps, yea, that should be "Performance relative to version 7.1" similar to the second chart comparing CPU performance. So 100% is the same, 200% would be twice as fast. I am at a conference for the next few days but I will get that corrected as soon as I get back. Thanks for pointing that out!

Posted on 2018-02-12 15:54:05
Jozsef Weigert

I think, the biggest problem with Lightroom is that moving in the Develop module can be utterly slow. We spend most of our time in that module. In your earlier test you also included a Develop module scrolling speed test, it would be great to have that again. Displaying clock speeds would be also useful on the charts as we can see how core speed and core count influences the results.
You mentioned that Lightroom is not optimized for Ryzen so well, however many of us would be curious, how the upcoming Ryzen+ will perform in Lightroom. While Intel might be quicker with LR, it would be great to know, whether Ryzen can offer us an acceptable speed for our purposes or not? Usually we don't buy a CPU only for Lightroom, that's why a Ryzen test might be handy, too.
The other thing is that I noticed that on my new 4K display scrolling and working in the Develop module is definitely slower, than in 1080p. It would be interesting to test, how 1080p, 2560p and 2160p affects speed in certain scenarios and on different CPUs.
You have the greatest test available, anyway, really useful for choosing the right hardware!

Posted on 2018-02-13 06:33:53
Roberto Berdini

I agree 100% specially in Develop module Test and resolution Tests. It would be very very interesting

Posted on 2018-02-13 16:07:13

Develop module testing is definitely on the to-do list. It is just really hard to accurately measure most things since there isn't a clear "task done" like there is with things like exporting. The image scrolling we meant to have in this test, but it was acting really inconsistent so we decided to remove it so that we could at least get the rest of the results up.

The good news is that in a few weeks I will be talking a bit with the Lightroom development team to get their assistance with figuring out how to benchmark things in the develop module. In fact, my current plan is to move to a largely plug-in based benchmark that we can put up for public download. We even just the other day got a new batch of images to use for our testing that we have full permission to distribute (Including some insanely high-res stuff from a Nikon D850) that I think will really improve our benchmark.

Posted on 2018-02-13 16:41:37
Dragon

Matt, in future LR testing, it would be helpful to include runs with larger RAW files. In a quick check, the new version seems somewhat faster, but LR has long been much slower handling big files (by more than the difference in pixel count). It has been a dog handling Canon 5DSR files, and while those are some of the largest around, most recent cameras sport more than 18 MP, so either run down the middle with 30 MP or best of all, do a run with 16 or 18 MP files and another one with 50 MP files. With the big files, your plan to test import from the HD rather than from the flash card is probably a necessity for a fair evaluation since differences in cards and card readers can have a big impact on 75 MB files. Also, keeping a quad core in the mix would be helpful to those of us contemplating an upgrade. When you get into testing the Develop module, I think GPU performance will come into the equation as well and possibly dominate. One last question. Since certain operations are noticeably faster on the 8700k than the difference in boost clock speed would suggest, a test run on the 8700k with the onboard GPU disabled would be instructive to see if Adobe is quietly using the onboard GPU even though the user setting is pointing to the 1080ti (that would clarify the "working with Intel to optimize performance" line). Thanks for the ongoing updates on LR performance. These articles are most helpful.

Posted on 2018-02-13 20:23:31
Jim Kiefer

Great article! I saw somewhere that all things being equal (e.g., megapixels), the different raw formats result in different LR performance. Like Nikon/NEF are slower to process than Canon Raw files and Sony Raw files, etc. If you can show results for the different formats in future testing, that would make the results more relevant. I do like the public plug-in that you mentioned.

Posted on 2018-02-14 06:01:19

I feel like god has existed in some form already though. I use it a lot. I can go back look at anything I’ve processed in any given time range.

Posted on 2018-02-17 21:44:12
pillybilly

I have seen every your publications but now i have a big doubt. Why do you choose only the 8700k in your configurations?
When it has a barely 6% more power than a 8700 but with a +30W tdp, it costs more and considering workstations are not machine for overclocking?

Posted on 2018-02-18 23:49:27

If you are buying a workstation that is likely in the $1500-3000 range (for a serious Z370 / Coffee Lake build) why would you not pay ~$50 more for a ~5% increase in performance? That is a 5% increase in performance (roughly) for a 1.6-3.3% increase in price (depending on the cost of the whole system)... which seems like a good deal. Yes, the 8700K will run a little hotter and use a little more power... but it would take a long time for the added cost of electricity to really add up to anything substantial. And just because it can overclock, doesn't mean you have to :)

I wrote a blog post about this which you might find interesting: https://www.pugetsystems.co...

Posted on 2018-02-19 16:48:59
pillybilly

Yes maybe you are right, the difference in price was not important but in some constipated micro or mini atx configurations it was more useful to have 30 watt less, that's all.

Posted on 2018-02-19 17:12:35

That is a good point, actually! In systems where the power consumption or heat output are limiting factors, then yes - by all means, the 8700 (non-K) is a fantastic choice. That is the reason our smallest system, the Echo, doesn't list the 8700K as an option :)

Posted on 2018-02-19 17:18:22

Remember that the TDP is based on the base clock, not any of the Turbo speeds. So in reality the power and heat difference between two CPUs is probably closer to 5 watts or so. That really shouldn't make much of an impact on cooling in most cases.

Posted on 2018-02-19 17:19:34
Stidi

Could you keep test these Lr tests also with previous gen 4 core processor? I think that many peoples using still 4 core processor (example i7-6700 and i7-7700) and considering how much power full processor is latest 8700K processor with 6 core. This 4/6 compare measure on the basic level how much Adobe has improvement multi threading.

And I still suggest you to test newest high speed (+4000GHz) memory modules because I have got over 30% Lr speed up with +4000GHz memories.

Posted on 2018-02-21 12:04:34
Mathias Appel

Hi! Very nice article, thanks for your effort :)
I have one question though, what settings are you using for the 1:1 previews and JPEG exports?
For example, with pretty similar hardware the export of 100 JPEG images takes about 3 and a half minutes. Sure, my D7100 has a higher resolution and the files are about 90% larger but even then that is relatively slow compared to your result.

Posted on 2018-03-05 16:58:08

We keep everything at default. So for exporting that is same resolution, JPEG with a quality of 80. No sharpen or anything like that. 1:1 Previews are also default (I'm not even sure if there are any settings you can change for those)

Posted on 2018-03-05 18:43:52
Mathias Appel

I see, thanks for the reply!
Setting the JPEGs to a quality of 80 makes quite the difference, the exported files are about 10mb smaller on my end.

Posted on 2018-03-05 19:20:12
photodude

Very nice review, gives me a lot to think about. Very interesting that the i7-8700K performs so well in the image import and conversion to DNG but is much lower in performance in most of the other general use cases (export, preview, etc)

My current rig
i7-4770 CPU @ 3.40GHz, 32GB ram
256gb OS SSD, 512gb LR catalog SSD, a 3TB and 6TB internal 7200RPM image storage drives
GTX 1070 graphics

working with
24.6MP raw files

I definitely see slowdowns in some areas of LR

Posted on 2018-03-18 23:44:13
MastaUvWaffles

We have the same system! I started to notice some slowdowns as well. Primarily while working in photoshop, which, I am fairly certrain does utilize multicore performance better. My workflow typical goes above the 4gb limit these days, so I'm looking for something more recent.

Posted on 2018-04-10 20:15:23
pauldeleon.com

Thanks for the awesome article.
Some questions for the crew...

Will there be a follow up with larger file sizes (my D850 is on the way!) and where the files start on the same drive as LR?

Thanks again for the great article.

Posted on 2018-03-22 17:38:11

Yes, we will be doing more testing in the future that will include larger files. I actually already have the files ready to go including a set from a Canon 5D MKII (22MP) to cover a bit older cameras and a set from a Nikon D850 (45MP). Right now we are spending a lot of our time getting ready for NAB (video/broadcast tradeshow in April) but after that I'm hoping to get back to some Lightroom testing. I want to revamp a lot of our testing process in the hopes of making it stable and straight-forward enough to put up for public download, however, so it may take a bit of time to get that completed.

Posted on 2018-03-22 17:49:02
photodude

Matt, Will the new tests cover disk performance issues?
I'm mostly working with
24.6MP raw files (converted to DNG) and a few small video files

From what I can tell LR performance is highly affected by Disk operations even with automatic XMP writing turned off
I tend to see very high disk uses across all my image drives.
256gb OS SSD,
512gb LR catalog SSD,
a 3TB and 6TB internal 7200RPM image storage drives

Posted on 2018-03-24 16:43:05
Borconi Szedressy Emil

Hi Matt.

Thank your for all your nice tests you are doing. I know it's slightly off topic, but I want to ask you (and all other members) which of the following CPU's you recommend for a Lightroom laptop, 7700HQ or 8550U I'm just about to replace my current laptop which has a 4712MQ in it and I'm not sure which is a better choice from Lightroom point of view. I'm not working with RAW files, I only import JPEG's apply some presets, add a watermark and export them, but I do this in bulk in some occasions with 60k+ images, so for me JPEG handling and import/export speeds are the crucial bit. Thanks for the advise.

After reading more about it I decided to go with the 7700HQ one.

Posted on 2018-05-03 11:27:29
DmitM

The major drag in LR for me personally is going from photo to photo, zooming in-out and working with adjustment brushes, culling 1000-2000 images. I can wait for the previews and export,but working in the development mod is always very slow on my 3770K. I am hoping you can include tests that somehow measure all of these in Lightroom. I have been waiting to upgrade for about 3 years as I haven't seen any tests indicating better performance in those cases.
Most just talk about import/export and while speeding up those is a good thing, 90% of all people who work in Lightroom would appreciate faster develop module speeds instead; and that's where LR lacked all these years. It's a real pain to work on 24-40mb raw files! It would be amazing if you can come up with tests that can measure CPUs somehow in the develop mode.

Posted on 2018-06-26 01:37:08
Mark

DmitM.... your work flow is an issue along with needing to work with raw files. There is a reason for smart previews.
My Lr cruises along with an 7820X processor and 32gig of ram all on a SSD.
Plus I have a properly sized graphic card for my monitor. so you can have a fast system but if the GPU can not push 4K/5K monitor... it will seem slow.

a nice feature about the skylake X CPU is the turbo 3 that allows 1-2 cores jump up to 4.5Ghz. this mimics running with a 4 core in an over-clocked computer.

Posted on 2018-07-26 04:49:31
Mark

Late to the discussion. Thanks Puget System for the test results.
I find my 7820X moves along very well.
I edit 2000-4500 images for several clients. 800-1500 for myself.
I do cull in PM. just better than Lr. Plus I create separate catalogs for each wedding.
I only work on raws for 100 images or less. Everything else is smart reviews.

I have found Lr 7.2 has pretty darn good performance. Still an occasional issue on heavy edited images.
Exports have basically doubled if not a little faster from LR 6 and 7700K cpu + 16 gig ram to current platform + Lr7.2

Really appreciate the testing you do.

Mark
wedding photographer
viaweddingphoto(dot)com

Posted on 2018-07-26 04:57:06
Mac McGommery

Hi Matt,
are there big differencies in the import/generating benchmarks when using higer MP count raw's ? Like ones from a Sony A7RIII ?

Posted on 2018-08-28 12:39:17

Hey Mac,
Higher MP images will definitely take longer to import and do things like generate previews, but in general there isn't too much difference in what hardware works best. For example, from what I've seen, the best CPU for working with lower MP RAW files is also the best CPU for working with higher MP RAW files. The only possible exception to this I've seen is that higher MP files might benefit from having a faster storage drive. If you have an SSD, however, you should be perfectly fine with pretty much any kind of image file.

Posted on 2018-08-29 22:14:31
Mac McGommery

Thank you for the reply Matt :)

Posted on 2018-08-30 09:26:40