Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

SFjohn

macrumors 68020
Original poster
Sep 8, 2016
2,199
4,462
Maxed out 18 core “In Progress”, Credit Card has been Charged. On the other hand my Home Pod is getting ready to ship, my credit card for that was charged 2 days after the iMac Pro, but the iMac Pro is stuck on “In Progress”. Feb 6 for the Pro is impossible I think...
 
Last edited:

yurc

macrumors 6502a
Aug 12, 2016
835
1,014
inside your DSDT
If don't mind, can somebody who have already 10/14/18 core iMac Pro post screenshot iMac Pro's 'About this Mac' ? I'm curious how it's look on iMac Pro on processor section.

Off course block your serial number first. Thank you.
 
  • Like
Reactions: SFjohn

bxs

macrumors 65816
Oct 20, 2007
1,151
529
Seattle, WA
If don't mind, can somebody who have already 10/14/18 core iMac Pro post screenshot iMac Pro's 'About this Mac' ? I'm curious how it's look on iMac Pro on processor section.

Off course block your serial number first. Thank you.

My 10core model...
 

Attachments

  • Screen Shot 2018-02-08 at 5.33.35 AM.png
    Screen Shot 2018-02-08 at 5.33.35 AM.png
    337.4 KB · Views: 293

yurc

macrumors 6502a
Aug 12, 2016
835
1,014
inside your DSDT
My 10core model...

Thank you for answering my curiosity. So it's show with humble Xeon W with processor speed only instead of 10 Core Intel Xeon W. So probably other model would be same (e.g 2.5 Ghz Intel Xeon W on 18 core model)

Well. just cosmetics though.
 
  • Like
Reactions: SFjohn

Beliyaal

macrumors member
Feb 14, 2008
53
19
I have some performance information for the 18 core iMac Pro based on a integer/branchy workload (CPU limited compiling with all cores in Xcode).

I derived this information by building with 36 processes in Xcode, and then disabling cores in Instruments. This probably means that it's not entirely representative of a real x thread workload, because only some parts of the CPU is taking all the thermal load. A real workload should perform slightly better due to more even thermal load.

The max turbo boost for X number of cores enabled:
upload_2018-2-11_10-23-44.png


I started the load when the CPU was at 50 °C or lower. It seems that the maximum sustained power allowed is 133 W, so anything below that will maintain maximum turbo boost indefinitely. For 6 core or less the maximum turbo boost is maintained indefinitely.

For higher core counts the CPU will briefly use more than 135 W, all they way up to 170 W. For my workload this brief period was maintained between 7 and 74 seconds depending on the power usage.

upload_2018-2-11_10-28-41.png


Aggregated frequency, simply multiply frequency by number of cores:
upload_2018-2-11_10-31-23.png


After the brief period of maximum frequency, for higher core counts the CPU will begin throttling to maintain 135 W power usage:
upload_2018-2-11_10-46-22.png


Over time the frequency will will decrease slightly for higher core counts as the temperature increases until eventually the temperature reaches about 88 °C and a steady state emerges:
upload_2018-2-11_10-48-30.png


upload_2018-2-11_10-49-23.png


Aggregated frequency, simply multiply frequency by number of cores:
upload_2018-2-11_10-50-24.png


Some conclusions comparing to https://forums.macrumors.com/threads/imac-pro-18-core-delivery-status.2100673/page-2#post-25715226:

I ran the test again with 10 instances of the Prime Number program running with 10 cores enabled and with Hyper threading enabled.

Frequency was a steady 4.1 GHz

(Assuming max turbo boost for 10 core with all 10 cores enabled is 4.1 GHz)

When you have a highly threaded bursty workload. For example editing filter settings in an image editor, for the best case in favour of the 18 core the 10 core could take about 40% longer time to produce a result (18 has a 29 % shorter response time):
(3.2 * 18) / (4.1 * 10) = 1.40

In reality the difference might be less.

I ran the test again with 20 instances of the Prime Number program running with 20 logical cores enabled and with Hyper threading enabled.

Frequency was varied from 3.49 to 3.64 GHz

For a sustained workload such as a long code compile you could expect the 10 core to take about 30% longer than the 18 core to finish the compile (18 core take 24% less time to finish):
(2.6 * 18) / ((3.49 + 3.64) / 2 * 10) = 1.31

So in conclusion
The 18 core can lower your response times up to 29% vs the 10 core
The 18 core can lower your wait times up to 24% vs the 10 core

Conversely, some single threaded and lightly threaded workloads could take 4% less time to complete on the 10 core, but this is somewhat reduced by the larger L3 cache on the 18 core.
 

SFjohn

macrumors 68020
Original poster
Sep 8, 2016
2,199
4,462
I have some performance information for the 18 core iMac Pro based on a integer/branchy workload (CPU limited compiling with all cores in Xcode).

I derived this information by building with 36 processes in Xcode, and then disabling cores in Instruments. This probably means that it's not entirely representative of a real x thread workload, because only some parts of the CPU is taking all the thermal load. A real workload should perform slightly better due to more even thermal load.

The max turbo boost for X number of cores enabled:
View attachment 750798

I started the load when the CPU was at 50 °C or lower. It seems that the maximum sustained power allowed is 133 W, so anything below that will maintain maximum turbo boost indefinitely. For 6 core or less the maximum turbo boost is maintained indefinitely.

For higher core counts the CPU will briefly use more than 135 W, all they way up to 170 W. For my workload this brief period was maintained between 7 and 74 seconds depending on the power usage.

View attachment 750799

Aggregated frequency, simply multiply frequency by number of cores:
View attachment 750801

After the brief period of maximum frequency, for higher core counts the CPU will begin throttling to maintain 135 W power usage:
View attachment 750803

Over time the frequency will will decrease slightly for higher core counts as the temperature increases until eventually the temperature reaches about 88 °C and a steady state emerges:
View attachment 750804

View attachment 750805

Aggregated frequency, simply multiply frequency by number of cores:
View attachment 750806

Some conclusions comparing to https://forums.macrumors.com/threads/imac-pro-18-core-delivery-status.2100673/page-2#post-25715226:



(Assuming max turbo boost for 10 core with all 10 cores enabled is 4.1 GHz)

When you have a highly threaded bursty workload. For example editing filter settings in an image editor, for the best case in favour of the 18 core the 10 core could take about 40% longer time to produce a result (18 has a 29 % shorter response time):
(3.2 * 18) / (4.1 * 10) = 1.40

In reality the difference might be less.



For a sustained workload such as a long code compile you could expect the 10 core to take about 30% longer than the 18 core to finish the compile (18 core take 24% less time to finish):
(2.6 * 18) / ((3.49 + 3.64) / 2 * 10) = 1.31

So in conclusion
The 18 core can lower your response times up to 29% vs the 10 core
The 18 core can lower your wait times up to 24% vs the 10 core

Conversely, some single threaded and lightly threaded workloads could take 4% less time to complete on the 10 core, but this is somewhat reduced by the larger L3 cache on the 18 core.
Thank you so very much for your amazingly detailed and extremely specific research, and posting all those results. I’d also like to thank everyone who has posted to this thread. The 18 core is faster than I expected, and I’m thrilled about receiving one this coming Monday. I hope this thread will continue with real life comparisons of the two, especially where the 10 core is faster, and where the 18 cores really shine...
 

SFjohn

macrumors 68020
Original poster
Sep 8, 2016
2,199
4,462
Mine arrived today, yea! Beautiful, using migration assistant over Cat 6 Ethernet from my maxed out mid 2010 iMac and it’s slowed to a stop? Called Apple, they advised wait until morning to see if it progresses. For the last 2 hours it’s claimed 1:55 minuets to go at 23 MB/s...
 

Bryan Bowler

macrumors 601
Sep 27, 2008
4,067
4,441
Mine arrived today, yea! Beautiful, using migration assistant over Cat 6 Ethernet from my maxed out mid 2010 iMac and it’s slowed to a stop? Called Apple, they advised wait until morning to see if it progresses. For the last 2 hours it’s claimed 1:55 minuets to go at 23 MB/s...

Congrats on finally receiving your new beast of a machine! I hope the transfer issues sort themselves out overnight.
 

thomas18k

macrumors newbie
Feb 20, 2018
1
1
I got my 18-core/36-threads two weeks ago. Performance might be nice for interactive work with transformations only lasting a few minutes, before the CPU gets time to cool down again. But the performance for complex simulations, e.g. particle dispersion, is quite disappointing. Whenever the task takes more than some 10 minutes real time the large number of physical threads becomes useless. I guess its because Apple prioritize silence over performance.
The optimum between real clock rate and number of threads seems to be at just using roughly 10 of the 36 threads theoretically available for long time calculation. That means more than two thirds of the expensive hardware is to no avail. Has anyone similar observations?
 
  • Like
Reactions: TTOZ

SecuritySteve

macrumors 6502a
Jul 6, 2017
949
1,082
California
Mine arrived today, yea! Beautiful, using migration assistant over Cat 6 Ethernet from my maxed out mid 2010 iMac and it’s slowed to a stop? Called Apple, they advised wait until morning to see if it progresses. For the last 2 hours it’s claimed 1:55 minuets to go at 23 MB/s...
How did it go? I'm really curious - these migration processes always have been finicky for me, and in the end not really worth the time.
 

Bryan Bowler

macrumors 601
Sep 27, 2008
4,067
4,441
How did it go? I'm really curious - these migration processes always have been finicky for me, and in the end not really worth the time.

I agree. I prefer to set-up brand new machines from scratch. It requires far more involvement, but perhaps less chance of bringing over any gremlins to the new system. (Perhaps that is just an urban myth nowadays, perhaps there is some validity to it.)
 
  • Like
Reactions: powerslave65

Macshroomer

macrumors 65816
Dec 6, 2009
1,305
733
I'm interested in this core count cost/benefit analysis as well.

I am a career still shooter at 30 years in, main cameras are 2x 24MP, 46MP and 50MP. I also do large format stitched work that is often 10-15' feet wide @300dpi. Currently I am running a 2017 15" MBP maxed out and a 2011 cMP 5,1 listed below.

I am considering this upgrade to replace the 2011 machine and don't mind paying to max it out if CC LR and PS all run at the top of their game. I'd probably use an OWC EX TB3 1TB SSD for a scratch disk, 4TB of TB3 SSD RAID for workload and put 32-48TB of spinner RAID on as first layer archive.

Keep feeding this beast folks, if not fiscal 2018, I'll upgrade 2019 for sure. I just want to start exploring it now.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.