Hacker News new | past | comments | ask | show | jobs | submit login
Lightroom $4K iMac VS $4K PC performance test (slrlounge.com)
66 points by rayshan on Apr 25, 2016 | hide | past | favorite | 69 comments

A lot of that iMac's cost is its retina display, and you didn't bother buying a retina display for the PC. Running a speed test is silly, there are already major differences in the build.

To make this remotely more scientific, you'd either need to buy a 4k monitor for the PC, or use a Mac Pro and just take the monitor out of the equation.

That's true. They should have gotten a 5k screen with 5120‑by‑2880 resolution for the PC like the Mac. This Dell costs 2000 which would have reduced the available for the PC components. On the other hand if you don't care for such resolution the PC may be the better choice.

If they wanted a more apples to apples comparison in the hardware department, they probably should have added $600 to the monitor price[1] and paid for it by using the same CPU as the iMac (i.e. a 6700k)

[1] http://www.amazon.com/Dell-Monitor-UP2715K-27-Inch-LED-Lit/d...

Honestly, though, I don't think the point was to do an apples-to-apples hardware comparison. This was on a cost basis only.

The article also leaves out the cost of their IT guy's time to assemble the custom PC components. How much does an hour of Joseph WU's time cost? How many hours did it take for him to order, receive, unpack, and then assemble the components?

My guess would be most of Wu's time was spent researching which components to order, probably several hours. On the other hand, if he's a full time employee, then it does not cost anything extra to have him do his job.

If he is a full time employee then the time he spent researching and assembling this computer was time he didn't spend doing the job he was paid for.

Unless 'researching, building and supporting the IT environment' is his job, as their IT guy

I guess the issue is the driving factor of the iMac cost is the monitor resolution. It would've made more sense (if they didn't need the resolution) to pick a different iMac and spec out a PC at a similar price point.

>On the other hand if you don't care for such resolution the PC may be the better choice.

On the other hand it's a Lightroom test -- the target market for Lightroom is precisely those that do care for such resolution.

Why would speed be silly? That is a business, who saves money with every speed increase.

Also, the end of the article shows the next text is a Mac Pro. I thought commenting without reading was a reddit thing, not hn.

These kinds of posts attract the Reddit-ier crowd of commenters that already exists within HN, if you know what I mean. The title gave me that feeling; the article confirmed it; and the fact that there were enough comments to require scrolling, on an article about “Which computer is faster, Mac or PC???”, hit the nail on the head and sealed the coffin.

Retina only good at pixels but didn't mean its color accuracy better than others, like EIZO.

Its color accuracy is also better than most -- it has 10-bit screen colors for one. EIZO equivalent quality monitors go for $1500 and higher even for crappy resolutions.

But the Mac pro hasn't been updated in years. The gpu is old, and this is mostly a gpu test.

The Mac Pro can dissipate way more heat than an iMac, so between the two GPUs it can have almost twice the FLOPS of the iMac's GPU, even though it's several years older.

The article states repeatedly that it's all about CPU clock speed. If that's inaccurate, could you elaborate on how it's actually a GPU test?

That's a bad argument. They spent $1K on a monitor. That can definitely buy a 4K monitor.

You can definitely buy a 4k IPS monitor for less than $1000, the amount they spent on a monitor. Source:


Maybe people already have good enough monitors they can just plug into a PC without having to buy a new one every time they update the hardware?

I'm not a Apple fan, but I find most of these tests are counting seconds and then extrapolating that to actual work time, which I don't think actually maps as well as the author thinks.

The exception to this is 'smart previews' test which shows a 6.5 minute difference.

I've asked the creator of fileloupe (http://www.fileloupe.com/) if his app would improve the performance of this one test. I know nothing about photos, so I don't even know what this 'smart preview' thing is...

Right. His mapping it 1-to-1 implies that most of the time is bound by computer tasks. So there would be almost no viewing/decision/input task time, just some rapid input, then walking away from the computer.

This is so often the flaw with business upgrades. Its much easier to blame it on old kit than to struggle through improving workflow.

I'm surprised by HN's response here. I always thought it was common knowledge that a custom PC would be faster than a mac. You don't pay for speed with Apple, you pay for user experience.

Not just user experience, lots of things people don't immediately appreciate: construction (e.g. unibody sturdy metal), backlit keyboards, multi-touch trackpads, newest ports, mag safe, better battery life, thinnes, less weight, hi-dpi high quality monitors, seamless "sleep/wake" behavior, etc -- not to mention high resale value.

> you pay for user experience

Which isn't vastly relevant in a single-application-suite workflow. It's end-to-end Adobe so you plug your SD card in, it copies some files out to an Adobe library and you edit them in Lightroom.

I think many of the problems people are describing is more to do with what's being compared in hardware. Different monitor resolution, different CPU, etc.

My own objection is buying top-end rubbish from a photography equipment supplier. You can get similar spec stuff and much cheaper. This seems to be an exercise in trying to spend $4k (which is actually quite hard when you're not buying a Mac) rather than building for value.

This guy managed to edit a 4k video on last years Macbook faster than a much beefier Windows computer:


I think that speaks to the "quality" of the software more than anything. I'm wiling to bet that Adobe Premiere would render that macbook useless. I wonder how much more(I can't imagine how difficult it is to write their software cross-platform) amazing the Adobe suite would be if they only focused on Windows (which allow for these super powered machines).

They could have saved some money on the iMacs by simply buying and installing the RAM themselves. If they opted for the 8GB Apple RAM they'd take $600 off the iMac price.

And the 1TB flash storage option they chose can also be questioned. I mean, 1TB of the iMac's internal flash storage is super expensive, but it's also WAY WAY faster than the Samsung 850 EVO in terms of read/write times. (Samsung 850 EVO has the SATA bottleneck.) My point is that if you know your process is constrained by CPU speed, don't put so much of your budget into fast storage. If they had opted for the 256 GB flash storage option in the iMac they'd take $700 off the price.

OTOH, I don't see a way around the inability to overclock the iMac's cpu. Stuck with 4.0 GHz (or 4.2 in Turbo boost).

> Speaking conservatively, a 25% difference in performance would turn an 8 hour wedding edit into 10 hours.

That's only true if all 8 hours are 100% spent waiting on the machine, which doesn't seem likely. This doesn't seem like a CPU-bound task. Most of the time is probably spent by the human actually looking at the photos and making decisions.

4000USD is a lot of money. That seems like an odd setup.

Not having ever wanted to buy a machine for that amount of cash, I'd expect to do better with multi-socket systems.

Anyone care to chime in? I struggled to spend over 1K GBP with my latest machine. (I think 'Extreme Edition' is basically like buying the S version of a car, mind). i7-5820K, 64GB ram.

One thing that does stick out is an odd monitor comparison. There's an expensive color repro "low" (in comparison) res monitor for the PC, against the standard iMac retina. Aren't there cheaper 4K monitors/TV's about?

Lightroom has a single core CPU bottleneck, multisocket is pointless for the this use case.

The rest is on the GPU

I suspected that, but what's with the monstrous 8 core thing then? Couldn't you equally overclock a 4 core CPU and save hundreds of dollars?

Basically, is the whole article just nonsense? :)

Absolute nonsense.

This stinks of a cross between attempting to do PR and having someone who didn't make a spending decision justify it.

As I said elsewhere, the place to improve is in their human workflows, or there is no improvement to get.

In their situation I think processing time is the bottleneck or at least a major factor. They probably see every penny back in productivity gains.

I'm sure. Are Xeon systems just super expensive, though? Is 8 cores the best you can do for 4K?

The CPU gets progressively expensive; and yes, more than 8 cores will probably blow the $4k budget. However, two things to note:

1) High-end i7 and Xeon with the same core count hover around the same price. However, Xeon are not made for overclocking whereas the i7 can be overclocked (easily).

2) The application they used for benchmarking relies on high CPU speed. Normally what you find is the more cores a CPU has, the lower their clock speed (I assume to keep it within the TDP).

Also, CPU is all but one price component in a Xeon system:

- The base motherboard would cost the same to a high-end consumer motherboard. If you're looking for dual-CPU motherboard it will cost almost twice as much.

- (Most) Xeon systems will require ECC RAM, which is more expensive than the regular RAM.

That $480 gaming motherboard does seem rather excessive too ...

If you want to have a stable system when overclocking 3GHz --> 4.5Ghz, it's a good choice.

Wow, you need to spend that much for decent voltage regulators?

Yeah, the ROG does command a premium. Good hardware quality but I think most of the extra cost is due to all the add-ons (e.g. premium onboard sound chip). It is a gaming motherboard after all.

However, never ever skimp on motherboard if you're planning to overclock :)

These comparisons are silly. Now imagine it's a BMW vs someone's project car both with the same $$$ invested.

Yeah, the project car will probably smoke it. But not everyone wants a project car.

They should be doing this test against a Mac Pro if they want a fair comparison. And yes, the Mac Pro would likely be more expensive, but I don't think it's a surprise to anyone that building your own overclocked PC is cheaper.

Why would it necessarily be cheaper? After all Apple has both economy of scale and the capacity to buy up the entire manufacturing run of many of the components. They should be able to assemble a system as tightly integrated as a Mac much cheaper than somebody could go out and buy all the parts for.

A better example might be the cost delta between buying a car from the manufacturer and buying all the parts to make the same car separately and putting it together. The direct from manufacturer car will be far cheaper.

Apple probably doesn't 'get out of bed' for a profit margin smaller than X%.

That number is probably an order of magnitude higher than the profit margin of component makers.

I agree, perhaps it would be better to compare the iMac's performance against a $4,000 Dell Precision workstation.

I wonder what would be the test results if they used a 4k display in Windows as well. Moving those many pixels has an extra cost.

Other than that, I find the review centered around their use case and well explained. Well done.

The video card in question should handle it without issue...

Final Thoughts > "... Apple is simply adding new products and updating product lines often times without much thought it seems. This can be seen in the horrid pen solution found in the iPad Pro where charging it requires you to have a spear sticking out of your iPad."

The above seem like a non-sequitur, but seems apt. I switched from Windows to OSX in 2007 to work exclusively on iOS. In the past year: I abandoned WatchOS due to artificial constraints, discovered that the Apple Pencil has an annoying skid/squeak, and an Auto-update from Xcode 7.2 to 7.3 broke my C++ toolchain. The last, of which, Apple QA has decided not to fix, even though XC 4 thru 7.2 worked just fine. Meanwhile, can't run Cuda on my year old MBP AMD Chipset. I may be switching platforms. Hence the interest in OP's article, including the tangental Final Thoughts.

I bought my current rMBP in late 2014, and it may well be my last Apple product... I've mainly been a fan because of the ootb unix support, and the touchpad (that's a huge deal for me). The screens are nice too, but others have caught up, and I can't see detail that well anymore.

The SMB/CIFS change from samba is what caught me, that was kind of painful. I finally upgraded to El Capitan, and not sure I like it much better. For the most part, OSX is the one that stands out in terms of how I use my computer... Ubuntu and Windows are closer to each other (with bash) than OSX, and once windows has native Linux subsystem in general availability, I may wind up using it more.

Just switched from the OSX ecosystem to a Dell Latitude in the XPS 15" chassis running Ubuntu. Feels so good to be back on real Linux rather than good-enough-if-you-squint OSX-flavoured BSD. I commend it to you.

This is a qualitative analysis at best. I live in both the PC and Mac world and I'll say this, Some of what you pay apple for is quality. I understand that many cannot fathom this, but their customer service net promoter score is very admirable. You get a quality machine. Style does matter, and you get style with a Mac as well. If style isn't important, explain to me why builders spend hours deciding on cases, lighting, and cable management. At the end of the day, these types of comparisons aren't going to sway anyone.

Of course a purpose-built PC will outperform an iMac, dollar for dollar. That's not exactly news. The iMac also has a severe price-to-performance disadvantage here by including an expensive 5K panel, while the PC is tested with a much cheaper 2560x1440 monitor.

However, the disappointing thing to me here is that you can't reach performance parity with the iMac by throwing more money at it. The iMac tested here is a completely maxed out machine.


The Skylake processor in the iMac is something the end user cannot overclock (without doing something extreme). And none of the Mac Pro cores are as fast. And Apple doesn't water cool their machines. They simply don't offer robust overclocking options to us. So, for a single-core, the 4.0 GHz Skylake in the iMac is the fastest thing Apple offers.

If there is an underlying, valid core criticism in this article, it's the same one everybody already knows about: Apple doesn't offer their end users a truly competitive array of system configurations. Instead, they force their users into carefully defined and limited product tiers, each one costing significantly more than the last.

Apple is delivering a turnkey hardware solution with the 5K iMac. They will also provide an upgrade path with future hardware generations. This article is the same-ole myopic view of DIY specifications and IT enterprise management. Macs are not built for people who can or want to do it themselves. Macs are built for people who ONLY want to focus on the task at hand.

Surprise, surprise. A custom-built, overclocked machine with more of its cost invested in its CPU was faster than one that wasn’t overclocked and is sold for looks and stability over raw performance. It’s not like the choice of almost all the components had absolutely no impact on the final result.

Isn't this more a test of lightroom's performance on a mac vs a PC ? May not really indicate the performance difference between the two.

$630 for the graphics card is a waste of money for Lightroom.

Should have spent it on PCIe SSD or Xeon CPU.

I don't think you'll get higher single core speeds on a Xeon...

Some operations is Light room scale rather well with the number of cores, and some do beyond 8 cores.

I haven't looked at it yet but I'm putting my money on the 4K PC kicking the crap out of the Mac, unless the screen kills the PC budget...

EDIT: Yep, for what it's worth, if you are doing retouch for print, pro studios use EIZO screens and they are usually around 2k-3k+ per unit.

The PC system is overclocked 50%! What a pointless article.

That spec iMac is currently $3879 at MacMall [1], To save money, you could easily spec a 4.0GHz/8GB/1TB FUS/M390 for $2299 from B&H [3], add the same SSD ($300) and ram from OWC ($328 for 32GB / $667 for 64GB). Bringing the total to $3266. We also need to go to a 5K 27 inch monitor on the PC side for a fair price comparison; the cheapest I could find was $1042 HP Z27q [4]. Add $42 to the PC price to make up for this.

Now, the PC price comes out to $4370+42=$4412, so we need to drop $1146 of from the PC to get a realistic price comparison. The RAM, Screen, SSD, and graphics card are anchored due to our Apples to Not Apples comparison. That leaves the motherboard, case, CPU, power supply and water cooler to drop in price.

I don't know water cooling particularly well, but lets say we leave that in to continue overclocking the CPU, but drop to a lesser model, same with the Power Supply and Case. So revised budget for each is:

Case: $125 -> $80 Power Supply: $140 -> $80 Water cooling: $120 -> $80

Now a gaming motherboard was selected because it's stable for overclocking, but let's throw caution to the wind and drop our budget to half that and hope it still overclocks stable. $480 -> $240

That leaves $261 for the CPU, which buys an i5-6600K (3.5GHz), or we drop the water cooling and run everything at stock to get to an i7-6700K (4GHz). In other words, the same processor in the iMac. If this benchmarks any differently, it's the Lightroom developers fault, not the hardware.

So, the upshot is, spending the same amount on a Mac will get you the same specs in a much better looking equivalent device.

[1] http://www.macmall.com/p/product~dpno~13697228~pdp.jhffibb?s...

[2] http://www.bhphotovideo.com/c/product/1190421-REG/apple_z0sd...

[3] http://www.bhphotovideo.com/c/product/1190421-REG/apple_z0sd...

[4] http://www.bhphotovideo.com/bnh/controller/home?O=&sku=11429...

if you buy apple, you pay the extra tax, and the anodized aluminum finish.

And the aluminum body connected to that finish. Why not say the tax is for sturdy, light, consistent (if not attractive) construction?

60 hours is easily lost when some piece on your custom rig dies and you don't have apple care.

Do it like any other responsible place and keep spare machines at hand.

Oh, I forgot, there's nothing as painless as a properly set up Windows Server with auto provisioning...

My primary work machine is a 2011 MBP and it's working like a fucking beast, but still, I fear the moment it will break and require 1-2 days in the IT department for reinstall.

Provisioning a Win7 machine with MSIs is PAINLESS if you know how to deal with the MS toolchain (except Java which is a PITA to get running...)

Ditto for network-backed home directories, as long as you have Cat6 networking and good networking hardware which you should have anyway if you're a serious business, it's a couple of config options set up ONCE for the entire AD and you're done.

Macs, to this day, don't speak PXE but their own Apple protocol, and something even remotely alike to assigning MSIs via Active Directory is not achievable without (paid!) 3rd party tools.

From the article:

"The honest truth is, I have spent just as much time in a Apple store at the Genius Counter getting my Apple machines repaired as I have maintaining my well built PCs."

Apple Care does take sweet time, but you don't need to go to the store. They'll pick your machine up by courier to fix, at least they did for my iMac when the hard drive died.

Hyperbole! What piece could possibly take 60 hours to order, receive and replace?

The same pieces that took four days to arrive via courier in the first place?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact