iMac Pro 18-core Follow Up Review
hrtapps.com
hrtapps.com
That's nice but it's not Apple's market. Apple's market is not other PCs, and it's not build your own.
This criticism of Apple has been around ever since Apple has been around (1970s). While the criticism has basically stayed the same since then, Apple has not stayed the same, they have grown. Their products have improved. Their market share has grown and it continues to grow.
Some people pay more money for a Benz, and other people say Hey that's a ripoff, I can buy a Toyota that gets me to work and back, just like your expensive Benz, for half the price! What a ripoff!
Apple is not in the business of making inexpensive PCs (and accepting all of the tradeoffs therein). The trends show that more and more people are willing to spend their money on Apple products.
They are clearly doing something right.
If you are enraged by their success, perhaps you are not their target market. And that is ok.
> They are clearly doing something right.
Mac sales are down 5%.
I'm still using my 2010 15" MPB (i5 duo core). When I was browsing for a computer at the time it was a product I could see myself using for a decade. I hope that turns out to be true--it works great for what I need to do. Their recent updates have made me shrug and think, "meh, I can wait." If I'm going to shell out $2k+ for a new computer it better be damn near perfect.
Which isn't to say that the Mac business is as good as it could be—surely it isn't, and they have several models they need to address with updates. I don't really like the Touch Bar. But the Mac is doing fine.
Mac sales and revenue were both down 5% from last year.
Mac sales down ~200,000 units.
Q1-2018 numbers, reported 02 Feb 18: https://www.macworld.co.uk/news/apple/apple-financial-result...
And that 3-8% market is an 70-90% of the high end expensive market, and commands 90% of the profits -- more profits than the next 3-4 players combined.
Let alone the OS feeling less polished, the removal of time remaining on battery and other user hostile changes, the touch bar being poorly thought out net loss (e.g. the fact that you can’t adjust the brightness so in low light typing, it messes with your eyes), I’ve had to warranty 3 MacBook pros in the past year. I’d never had to warranty an apple product before.
I don’t want another iPad or iPhone. I want a laptop that just works and I’m willing to pay for it. Please somebody deliver that, since apple isn’t chasing it anymore.
I do, but I’ve always preferred less travel.
Outside of the moaning online bubble (like HN) the MBP is a raging success.
I have no feeling with it.
So yeah, for the first time in years, I'm looking around at other brands.
I'm just locked in with iPhoto.
Large key travel is only good for those hitting they keys way too hard imo
(Laptop keyboards also don’t have the space to properly raise the keycap tops on keys in further rows, because they don’t have space for it; this makes reaching for the top 2 rows of the keyboard noticeably slower and less comfortable)
The best keyswitches actuate about halfway through the stroke, have a distinct drop in force at the actuation point, and have a significant mechanical bounce to them on the upstroke, launching your fingers up toward the next keypress. My all-time favorites are Alps “plate spring” switches used on certain Japanese IBM computers and some IBM luggables in the late 1980s (I can type about 20–30 wpm faster on these than on any laptop keyboard). You can also get this effect (with a very different overall feel) from Topre switches (still produced), or from IBM beam spring keyboards (from the 1970s). Some other keyboard enthusiasts have different preferences, e.g. for simple springs with no additional tactile mechanism (“linear switches”) optionally with a clicker or buzzer to give audio feedback of key actuation. Such switches are slower to type on than a snappier alternative, but some people prefer the feeling.
Nobody should be hitting any type of keys “way too hard”. You want to use just enough force to actuate the key and not push the key hard all the way through the bottom of the stroke. This is much easier to judge when the switch has relatively dramatic tactile feedback and a lot of post-actuation travel which ramps up in resistance toward the bottom.
For example: the IBM Model M is one of the single worst typing experiences imo. The only way to type on that thing is use about 5x the force I normally do.
No wonder our industry is plagued by RSI - it lionizes horrible things like this due to misguided nostalgia.
In any event, actuation force is an orthogonal concern to travel distance, tactile feedback, post-actuation force curve, upstroke bounce, keycap shape, etc.
Relatively fewer people typing on a such IBM keyboards got RSI compared to other keyboards, because the extreme tactile snap makes it very obviously unnecessary to type with much force more than barely required to actuate them.
But the cause of RSI is largely down to posture and typing style more than keyswitch design, though the latter can contribute (the worst are cheap rubber dome boards from the 90s-2000s, which often force a hard smashing typing style). We would have much less RSI if computer users kept their wrists straighter, pulled the keyboard closer, tilted it parallel to their forearms, and stopped slouching as much.
“Nerve issues”, hah, what a trolly answer.
Could it have a slightly better power/price ratio? Yes. But it’s so much better than my pervious model MBP it’s rediculous (previous was two releases behind but one physical revision).
Neither has any developer I know at my work. Are you somewhere with a lot of dust or something?
I’ve seen issues like that with literally every laptop or regular pc keyboard since the 80s. Regular maintenance of a computer (any one) includes gentle cleaning of debris like that.
The problem is not whether your environment is pristine. The mechanism is just unreliable. Some people end up with keys not working after a few weeks of limited and careful use. My guess is that the tiny parts relatively easily get into some kind of subtle misalignment.
The basic design is sound enough. I think with another few generations of engineering tweaks they can get these keyboards working as expected. In the mean time it’s pretty frustrating though.
I don't wanna play no more. Because of the atrocious nature of things vis a vis MBP2017, and onwards, I'm leaving for something else.
I say this, typing uncomfortably on a MBP2017 with a touchbar I never use, making constant typos, and feeling just plain ripped off.
So I'm curious how you find it so useful?
Most of these MBP users are what I call new users, they are new to the Apple ecosystem. Mainly pulled in by the iPhone hollow effect. That is why Mac as a segment is still doing so well, there are lot of new users coming in, especially from China.
The sales data doesn't tell how angry some of these old Apple users are. If it wasn't a friend of someone within the Apple executive circle told them how they were planning to leave the Mac ecosystem because the current Mac Pro didn't work, they will never have known they had a problem.
It took 2 years for Apple to release a new product after they knew its has problem. Once Steve Jobs knew how many people hated the Apple Mouse, they had news and Interview saying they knew people didn't like it, and they are working on it.
The new Apple now, it is coming up to 5 years and Mac Pro hasn't seen a update.
My first Mac was an LC475 upgraded with a real '040 to run MATLAB, back in oh, '94 I think. Also had a second-hand SE30 to run Word 5.1 (still the best version). Next up was a Power Mac 7600 when I needed more grunt to run FORTRAN code. Amazing machine, the price/performance was incredible. It was faster than a DEC Alpha 3000 at a fraction of the price! A few more both desktop and laptop between then and now. So I think I count as a long-term Mac user and loyal customer.
My current machines are a 2008 MBP and a 2014 Mac Mini. Apple simply haven't given me any compelling reason to upgrade, in fact have pushed me to offload my compute needs to Azure. Now that I have made that leap, I'm actually not particularly interested in a real Mac Pro anymore, I just don't need it. I can't be alone, and this will bite Apple in the ass at some point.
Who's buying the current MBP then? Not professional users, just people looking to flash the brand. Which is fine, but the fashion crowd is fickle. Apple has chosen to court them rather than 20+ year customers. And that's OK too, it's their business to run as they see fit. But a third thing that is fine is for people to post their opinions of it...
I bought an MBP last year, because I needed a new computer and I still think the Macs are the best for my needs, in both software and hardware. And I am most definitely a professional user.
Do I wish it had a few more ports and a quieter keyboard? Yes!
Did that ever seriously threaten my purchase? No!
I don't think anybody who wants to just show off the logo is going to buy an MBP over a straight MB, which is lighter and (IMO) better looking.
But is that even "a thing" with computers in 2018?
I don't know where you are but in the UK there is a TV ad suggesting a MBP is the ideal machine for parent to buy a first-year undergrad.
The prioritization of thinness above everything else - when there is already the Macbook Air for that market segment - is a further example. I'll buy a brand new MBP as soon as there is a compelling reason to replace my 10 year old one. That just hasn't happened yet.
If you're satisfied with the performance of a Core 2 Duo, still able to get fresh batteries, too old for your eyes to benefit from a Retina display, and don't want to shave off 25-40% of the weight, then why do you expect that there ever could be a compelling reason to replace your current MBP? If you don't want anything more from your computer, then how is anyone supposed to entice you with a computer that does more?
My reasons for upgrading to a MacBook Pro were: (i) the better screen (which matters when you're looking at it eight hours a day), (ii) improved performance compared to the Air or the MacBook, (iii) the ability to drive a couple of high-definition external screens, and (iv) all the above on on a portable device that runs MacOS. Number iv is probably the most important reason: I prefer MacOS.
If you construe "professional" as someone with high CPU and GPU processing demands, then perhaps I don't count. But that's a tiny market: most professionals who choose a MacBook choose it for the same reasons I did.
Yes, me too, so I find myself in a weird no-mans-land - I don't want to switch to another platform, and I don't want to buy anything in the recent line-up. I will be forced to when my Mini becomes unsupported for OS updates, then I will spend as little as I can to get back in the game, because they have already pushed me into the mindset and workflow of offloading the heavy lifting. But offer me a machine as radical my old 7600 was relative to what I had previously and relative to the rest of the market, and I'll jump at it.
But for people that want a new laptop, the MBP is the best machine available by many metrics, especially the screen - it's finally bright enough I can work outside on a sunny day.
A manufacturer that in 10 years hadn’t produced a compelling reason to upgrade doesn’t strike you as a little odd?
Nevertheless the point stands; its not “haters” criticising Apple here, but long term fans and customers
You didn't even explain what you want from Apple. Is it about performance or some groundbreaking new feature?
I think most professional MBP users would happily trade 5mm thickness for 32 or even 64G RAM, and a better CPU/GPU.
Edit: Also, I know people love to complain about how thin it’s gotten, but as someone who has to drag it to meetings a lot, I kind of appreciate the thinness (along with the one-port connection when I get it back to the desk).
Not true. Apple sold 5.112M Macs units in 13 weeks recently and 5.374M Macs in 14 weeks a year ago. That is 0.393M/week recently vs. 0.384M/week a year ago.
https://www.apple.com/newsroom/pdfs/Q1_FY18_Data_Summary.pdf
Besides, without checking how did the overall industry go, it can't tell much.
Mac sales are down 5% in Q1 2018, after being up 25% in Q4 2017. Its hardly a convinceable narrative that market share is steadily decreasing, because it isn't (incidentally, Apple moved up the ranks to #4 PC maker because the whole PC market took a dive).
Yet they’ll think they have a better car, because it’s got a Ferrari under the hood.
However I do consider the Mac OS to be very Benz-like, which is why I think it’s terrible.
The Apple Tax used to be pushing 100%. Now it’s somewhere around 15-20%. It’s amazing. It’s the reason I got the iMac Pro, despite having a booting Hackintosh OS on very close to the same hardware. It’s just not that much of a premium anymore.
PS: anyone want to buy a PC? it’s in a really nice case and has lots of ram.
The fact is modularity is the niche, not the other way around.
Most top professionals don't actually upgrade they machines often, because you usually want new everything, and saving a few $1000's every year isn't worth the hassle.
More and more evidence comes out suggesting that it is not as dominated as you think. But if you can substantiate your claim I'm all for it.
In my mind, there are two main advantages - one is that you can save money by buying only the capacity you need with an option to increase the power later if you need it; and the other is that you can extend the lifespan of the system by upgrading it after a few years instead of buying a new one.
Both of these things don't actually get you a better workstation - they get you cost savings. If you max out the system initially, and when you need something more then replace it with a brand new one, then you get better results, though at a higher cost.
I'd say that the market of $13000 workstations is one where you've already stated that you don't care about savings that much, and you're willing to pay a premium for a better workstation.
Enthusiast home users tend to care about expandability, since they want maximum power at a limited price, and would be willing to spend some time and hassle to save, say, $2000. However, they're not the target market of the premium pre-packaged workstations; they're less likely to pay the "Apple tax" and more likely to build something with comparable power cheaper. On the other hand, companies buying $13000 workstations would rather pay $2000 extra to save time and hassle. If they had a decision "upgrade vs replace" they're going to replace anyway, so the ability to upgrade is literally worthless as they're not going to use it.
Still, the iMac Pro can be considered a good buy if you do like many Apple users and sell it after a few years for a hefty percentage of the purchase price, using the funds to offset its replacement. It's how I progressed from a PowerBook Pismo G3 -> eMac G4 -> Mac mini G4 -> two Intel Mac minis (2006 & 2011) -> Mac Pro tower, only keeping the first gen Intel Mac mini when I moved away from the platform for daily use. Because I recovered at least half the purchase price of each successive unit (again minus the first Intel mini) I was able to upgrade for more or less the same cost as a comparable PC workstation.
Conversely, good luck selling a used custom built PC for pennies on the dollar even if you used top of the line components. Even name brand, quality PC workstations don't have a healthy used market; eBay is full of 2016 era corporate machines that were $800 new, now going for $200 or less.
Well lucky for you: you can add four 4K external displays in addition to the built-in 5K monitor. And you have a PCIe bus coming straight out the back which you can use to plug in GPUs, drives etc.
That being said the machine may not be right for you, but a big corporate shop is probably quite happy to have a bunch of identical machines it can swap out instead of fixing them in situ, so getting the user up and running faster.
Supposedly they learned from the "trash can" debacle and have something more cheese-grater-esque in the pipeline for later this year.
Sounds like this iMac Pro is not for either of us. I'm going the other way: I develop all my code in Emacs on a MacBook (not "Pro") and when I want to crunch, do the crunching on a cluster on my network. I don't even need the 10G ethernet the Mac Pro has, and it wouldn't be fast enough to do the crunching I need anyway. Would I benefit from a faster machine? Sure, but I don't want to carry even a macbook pro any more -- this ultraportable is so tiny I can take it anywhere.
I remember this being an issue on the older Mac Pros; it's interesting that this isn't an issue anymore. What's the reason behind it? Better thermals? Turbo Boost?
Yes.
>I imagine most benchmarks do single-core runs by running one process of work, and do multi-core runs by maxing out the cores and then totaling the performance.
Aggregate throughput, yes.
>Do any benchmarks max out the cores, but only report the performance of one process of work?
Sort of. You can get that figure by dividing the aggregate throughput by the # cores. Sometimes it is reported explicitly.
What are the requirements on running, say, TensorFlow? Isn’t it just the GPU there?
https://www.anandtech.com/show/12032/amd-announces-wider-epy...
The main tensorflow branch does not support AMD GPUs at all.
There's a bunch of experimental code by AMD that possibly works in certain conditions, but is neither properly supported nor widely used. For example, there's hiptensorflow which seems to be a TF port stuck on 1.0.1 (TF is now up to 1.5, with many major releases and key changes), not being actively developed (the last commit seems to be at the end of 2017) and core parts of it (ROCm) don't work in OS X (tensorflow as such runs just fine on OSX, just CPU-only, which is a major performance limitation compared to any machine with a good nvidia gpu), and thus are not that relevant for discussing Apple systems. There are also some OpenCL port efforts, none of which are stable, up to date, and actually usable to get the expected performance improvements. All these projects are interesting, promising, difficult, and their developers deserve all the best, but calling them "Tensorflow has been supported" is an extreme exaggeration, PR vaporware.
Proper support is not there yet. I believe that there are likely ways to jury-rig some code together and run it on AMD hardware, but that's not comparable to the maturity level of these tools on nvidia (essentially, run the install, and it works). AMD still has to do their homework and fulfil their promises. They've spent far too long saying that the software support is coming soon, and now their claims have to be taken with a grain of salt. They can hope that the community will take over effective development a year or two after they do that and people in the ML scene start widely using their hardware, but not earlier. Currently, even if you have a nice AMD GPU, it's cheaper to just buy another high-end GPU from nvidia rather than spend a week of engineer-time to possibly maybe get their shoddy software to do what you need.
AMD has a lot more work to do in this area before people will seriously consider them in GPU computing.
We did use a compute cluster, but using it drastically reduced the turnaround time during development in my experience.
That being said I do know one such scientist with the same workflow but who remoted into a Mac Pro, not sure what they are doing now.
It would be convenient if I could run machine learning tasks locally, but that's limited to very small models (since it doesn't have an nvidia GPU and AMD hasn't yet done their homework to ensure that common software packages work), so those tasks need a remote server anyway.
20 years ago in 1998 SMP (before we decided the word "core" sounded nice) was quite common, certainly in machines used by developers and graphics designers.
30 years ago in 1988 SMP in microcomputers was uncommon, but it was very common in minicomputers and mainframes.
Whether software has been written for SMP scaling is quite another thing. Plenty of software written today isn't.
So that's 1999, 19 years ago. You're absolutely right.
Ummm...perhaps the author has a different definition of "real-world" than I do?
Top three consumers of CPU and IOPS on my Mac Pro or MacBook Pro: video editing and transcoding, photo triage and conversion, and Kubernetes clusters (minikube+xhyve) doing whatever.
A really weird machine that seems overly strong in some areas and absolutely weak in fundamentally critical places.
https://www.anandtech.com/show/12032/amd-announces-wider-epy...
All I can say here is "Wow!!!" That's an amazing price for a computer that is completely user hostile to upgrade or repair.
But put in the context of _workstation_ prices -- which I guess is what it's sort of in the category of -- check out what a Sun 2 (68010) workstation cost in 1986:
"A color 2/160 with 8Mb of memory, two 71 Mb SCSI disks and 60 Mb 1/4" SCSI tape cost $48,800 (1986 US price list)."
That's $109,766.16 in today's dollars.
Or more modest and perhaps more comparable, an HP 9000 Model 705 PA-RISC workstation (32mhz, HP-UX) in 1991 was priced around $5000. That's $9000ish in today's dollars.
People bought stuff like that back then. Low volume, prestige item, specialized market, and I guess Apple must figure they will still do that now.
Workstation class machines have always been overpriced.
Besides, that thing is going to get hot, and Apple doesn't exactly have the best track record when it comes to thermal design (see: apple ///)
Think of a photographer: if you charge a fixed fee any boost in productivity allows you to take on more clients. If you’re a freelancer at a public event, getting your final photos out quickly might be what decides whether someone buys from you or a competitor.
For a small business that kind of margin is often the difference between making it or failing.
Even at a low rate of $100 an hour, I could make that on the side with 50 hours of work.
"I could make that on the side with 50 hours of work." if you're already working full time, fitting in another 50 hours of billable work is non-trivial, and may take months to actually do.
But let me give you another real world example of how little $5000 is in the grand scheme of things. Last year, my after tax rate as contract to perm was a little over $50 an hour. I easily put in an extra 100 hours in a month.
I’m not going to spend $5000 on the iMac Pro because I don’t need it. But I probably will spend around $3000 on a 27 inch 5K iMac with 32GB of RAM just because that’s what I want and it should last at least 5 years just as a development machine for hobby projects and learning. What would be an extra $2000 if it were going to make me money?
There are around 700 hours in a month. Most people are already "working" (billable/salary/etc) for ~180 of those. Eat/sleep/travel/family/friends/life eats up a lot more. Fitting in a extra 100 hours of anything in a single calendar month will be extremely difficult for most people. If you have the spare time and money, I'm presuming you may be younger and not have a family that you have responsibilities towards (nothing wrong with that, and just an assumption). The majority of heads of households I know - even professionals earning decently - would indicate they have competing interests for $5-$10k on discretionary purchases.
You went from a 'low rate' of $100/hr to $50/hr in your examples?
I get it - $5k isn't a massive crippling amount of money for most western professionals. Most of those professionals would probably want the $10k pimped out version too, and that expense just isn't trivial for most individuals. Larger companies with equipment budgets and whatnot - less of an issue, for sure.
I went from $100 a month - as an independent consultant where you own your own business and the $5000 would be pre-tax to a real world example for me - Earned $80 hour as a W2 contractor where the consulting company gets a cut and pays the employer's part of the taxes. The $50 is after tax.
I'm not saying the extra 100 hours was easy. But I had to do what I had to do and I was willing to do it for the extra money. It's actually somewhat easier for a married person with a supportive spouse to put in extra hours. My one child who is in the house is in high school.