Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I'm more concerned about the 6200 supposedly being faster than the 390 o_O

I understand the whole i7 hyperthreading in benchmarks vs an i5 but this gpu thing is worrying...

Also, how the hell is the 3.1 i5 broadwell in the 4K beating the 3.2 i5 Skylake in the 5K?
What I meant to say was that I don't understand what exactly they are saying but the fact that the M290X is so much faster than any other GPU shows that their results aren't overly meaningful. Maybe it was actually accurate for what they have done, but even then it seems like some freak occurrence rather than a meaningful comparison. This goes for the M290X as well as the 6200. I haven't seen anything remotely similar in any other published benchmark.
 
Last edited:
Does OS X make use of hyperthreading in any way, such as multitasking, or is it just apps?

In a quad-core hyperthreaded CPU, the four physical cores are presented as eight logical cores. They all appear identical to the app and OS from a thread scheduling standpoint. Any multithreaded app or OS component can potentially use all logical cores. There are no hyperthreading APIs to take advantage of the extra logical cores -- it happens automatically.

The OS X thread scheduler is hyperthread-aware. It knows if a hyperthreaded CPU is present and it will sometimes elect to only schedule every other logical core. This is because in some cases two threads on one physical core can cause "cache thrashing".

Even though an average Mac may have thousands of threads, at any instant only a few of those are typically in a "runnable" state. The others are blocked waiting for I/O, system call, user input or a synchronization signal from another thread, etc.

So the simple answer is all logical cores of a hyperthreaded CPU can be used by both OS and app threads. However this does not necessarily translate to dramatically higher performance.

Aevan gave a good summary of the i5 vs i7. The only additional complication is the i7 has a faster clock speed. So disregarding hyperthreading, the i7 base clock speed is 21% faster.

The turbo boost clock speeds are closer between i5 and i7 but this doesn't mean much without knowing what the turbo boost behavior is for your workload. The only way to determine this is running it on an i5 and i7 while monitoring the turbo boost behavior with a tool like Intel Power Gadget: https://software.intel.com/en-us/articles/intel-power-gadget-20
 
After much deliberation I feel this years iMac is not a milestone release and therefore not a wise purchase. The Skylake chips don't offer much of a jump from last years i7s and the GPUs are just a rebrand with very minimal gains, with the exception of better thermal properties across the board, but then I wouldn't call Apple fixing a design flaw an upgrade...

I have lived without an iMac to this point so I'm sure I'll manage another year. From an investment standpoint I would rather wait for a milestone refresh with half decent gains in the CPU/GPU department, preferably involving a switch to nvidia, and then chuck everything I've got at it, maxing it out!

I had considered getting a lesser model this year and then selling it next year but to be honest I don't want to lose money on it and frankly I just can't be bothered. If the iMac was user upgradable in more areas and/or the overall price of the upgrades from Apple weren't so ludicrous it wouldn't really matter so much but as we all know, Apple is very overpriced and as such a purchase ends up being more of a strategic move than a fun experience.

So until next year. Peace :p
 
Seriously I wish you would participate in more discussions............you put a real dose of reality into the discussions and describe it in a way that the average user can understand.

Kudos to you and your posts....

Hahah! Thanks a lot, friend. It's a hobby of mine and I do it for fun, but I do like to help out if I can. So glad if someone finds it useful. :)
 
After much deliberation I feel this years iMac is not a milestone release and therefore not a wise purchase. The Skylake chips don't offer much of a jump from last years i7s and the GPUs are just a rebrand with very minimal gains...
Despite having to return it, at least you tried one and had it in your hands. I think the lag problem was an anomaly -- I've never see that. But you experienced the retina display, etc. You are making an informed decision based on your own experience and research.

That said, you must expect incremental improvements from this point on -- at least on the CPU side. CPU architects have used nearly every available trick to wring out more IPC and there is very little left. Each new process node is increasingly expensive and difficult -- for the first time in many years Intel is breaking their "tick tock" development cadence because the 10nm process is so late. The Skylake successor will not be a die shrink but a new microarchitecture on the same 14nm process. It will probably be a few % faster but no huge gains.


On the GPU side there are more possibilities but for many workloads, the GPU only helps so much.

My top-spec 2015 iMac 27 is not hugely faster than my top-spec 2013 model, but it is *usefully* faster and the retina display is a huge improvement. On the LuxMark GPU tests the M395X is much faster than the nVidia 780m in my 2013 iMac (see attached).
 

Attachments

  • iMac27Benchmarks.jpg
    iMac27Benchmarks.jpg
    741.1 KB · Views: 148
<snip>
Aevan gave a good summary of the i5 vs i7. The only additional complication is the i7 has a faster clock speed. So disregarding hyperthreading, the i7 base clock speed is 21% faster.
<snip>
EXACTLY. You will certainly find those 21% higher clockspeed handy, IF you have cpu-bound tasks in your portfolio. In general, there are more important upgrades (such as the SSD) but upgrading to the i7 will probably be a good investment in future proofing as well, besides the higher core frequency.
 
On the GPU side there are more possibilities but for many workloads, the GPU only helps so much.

My top-spec 2015 iMac 27 is not hugely faster than my top-spec 2013 model, but it is *usefully* faster and the retina display is a huge improvement. On the LuxMark GPU tests the M395X is much faster than the nVidia 780m in my 2013 iMac (see attached).

It's the 990M I keep hearing so much about that I am intrigued by...
 
EXACTLY. You will certainly find those 21% higher clockspeed handy, IF you have cpu-bound tasks in your portfolio. In general, there are more important upgrades (such as the SSD) but upgrading to the i7 will probably be a good investment in future proofing as well, besides the higher core frequency.

21% under non turbo boost conditions. joema2 is right to point out that we can't deduct too much from the stated turbo boosted frequencies, but it's save to assume that they are more applicable the less cores are used. I know from other computers with intel CPUs that run almost exclusively at turbo boosted speed as long as only one core is stressed, but this doesn't necessarily carry over to the iMacs since it largely depends on the involved cooling solutions.

But all these numbers are just indications of what someone might be able to expect, yet benchmarks demonstrate clearer what these CPUs are actually able to deliver. And from what I have seen so far, if you expect more than a 10% advantage for single core tasks, you'll likely going to be disappointed.
 
...as long as only one core is stressed...from what I have seen so far, if you expect more than a 10% advantage for single core tasks, you'll likely going to be disappointed.

The only caveat is many of the time-consuming operations discussed are heavily multicore, e.g, Lightroom import and preview generation, export, video transcoding, etc. In those cases where all cores are heavily used, wouldn't the CPU base clock freq. dominate over the turbo frequency?
 
The only caveat is many of the time-consuming operations discussed are heavily multicore, e.g, Lightroom import and preview generation, export, video transcoding, etc. In those cases where all cores are heavily used, wouldn't the CPU base clock freq. dominate over the turbo frequency?

Absolutely, and I'm quite sure that this, combined with HT is the reason why a 50% advantage isn't uncommon in multi threaded benchmarks. I'm not even going to argue that the i7 would be beneficial if this is a big part of what you are doing. I just like to encourage people to think about whether this is actually the case.
For me, while I do some photo work, I do it on a rather casual level. The time I would have saved by using an i7 would be too negligible to justify the cost, importing and exporting is still quite fast. My main bottleneck is to actually dig through the photos I made and edit the ones I like, a task where I am 99.9% of the bottle neck. But it would be a completely different story if I would have to import hundreds of new photos on a daily basis.
 
After much deliberation I feel this years iMac is not a milestone release and therefore not a wise purchase. The Skylake chips don't offer much of a jump from last years i7s and the GPUs are just a rebrand with very minimal gains, with the exception of better thermal properties across the board, but then I wouldn't call Apple fixing a design flaw an upgrade...

I have lived without an iMac to this point so I'm sure I'll manage another year. From an investment standpoint I would rather wait for a milestone refresh with half decent gains in the CPU/GPU department, preferably involving a switch to nvidia, and then chuck everything I've got at it, maxing it out!

I had considered getting a lesser model this year and then selling it next year but to be honest I don't want to lose money on it and frankly I just can't be bothered. If the iMac was user upgradable in more areas and/or the overall price of the upgrades from Apple weren't so ludicrous it wouldn't really matter so much but as we all know, Apple is very overpriced and as such a purchase ends up being more of a strategic move than a fun experience.

So until next year. Peace :p


Last year's iMac was a milestone. This year's was just a minor improvements in multiple areas(which most were expecting)
 
Last year's iMac was a milestone. This year's was just a minor improvements in multiple areas(which most were expecting)

I agree with last year's iMac being an absolute revelation but I feel like a lot of the improvements were needed to make the iMac a much better machine such as the skylake processor, better heating and better SSD speeds. Furthermore I feel like the pricing this year is better (I think last year's model started at 2499).

Well, I still went for it this year as I simply can't imagine waiting another year for a refresh for my dream machine.
 
  • Like
Reactions: OddyOh
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.