rMB 2016 benchmarks question

[pmouritz]

So I finally did what I've been wanting to do for the past year.. I sold my 4 years old but very loved 15" rMBP (mid 2012) and bought the new 12" rMB (early 2016) with M5 processor and 512 GB RAM. And it has been great so far! I really love the new MacBook in every aspect and does not feel that it is significantly slower for the work I use it for.
However, I've run several Geekbench 4 (64-bit) tests and come of with various results. I'm aware of the fact that all processors varies individually but I'm wondering how my rMB can score ≈ 5400 in Geekbench 4 (64-bit) when many others have reported scores in the high 5000's and some even as high as 6200? Is this something that I should be worried about? I mean, the m3 has been reported in the 4900's.

I still have the 14-day return policy, meaning that I have the chance to return my currently new rMB and try to buy another, in the hope of getting a 'better' chip in it.

It is not because I think that my rMB is too slow.. But I guess I just want the most out of my money

EDITED: So this is my results, https://browser.geekbench.com/v4/cpu/432409, seems weird that all the tasks under 'Multi-core performance' are high except Memory Copy, Latency and Bandwidth.

 

[mrex]

i quess it depends how cool the cpu was in the beginning and how cool it stays: cooler environment, the longer it can push the limit, the higher room temperature, the lower clock speed you will see...

if the ambient temperature is "high" already, it doesnt cool the cpu enough... so i think the variation is due to the initial temperature and an ambient temperature.

 

[mrex]

is the geekbench score for m5 generally or just for the mbr m5?

 

[pmouritz]

Yes, I figured that the ambient temperature and the current temperature when the test was initiated have an influence. I've done the Geekbench 4 at various times and many just after I booted the computer to ensure that temperatures had not been raised significantly to reduce the performance.
I see a lot of different scores on Geekbench's database ranging from 2000s to 6000s. I guess, I just expected that the score would be around 5900 as that was what had been reported on various sites reviewing and testing the rMB.

 

[fisherking]

a lot too depends on what apps are open, what processes are running... when you do the GB test. i just got a score of 5204 (this is the entry-level MB), with all apps closed (and crashplan closed). am sure it would be different later...

 

[pmouritz]

i just got a score of 5204 (this is the entry-level MB).

Well that is not too shappy! That is a great score for the entry level imo. Maybe I've should have saved the extra bucks and got the entry level..
I also run Geekbench 4 with no other programs running than the test and get 5400 consistently which frankly is a bit disappointing considering the extra cost for the m5. Especially when you can report that your m3 delivers 5204

 

[KPOM]

Well that is not too shappy! That is a great score for the entry level imo. Maybe I've should have saved the extra bucks and got the entry level..
I also run Geekbench 4 with no other programs running than the test and get 5400 consistently which frankly is a bit disappointing considering the extra cost for the m5. Especially when you can report that your m3 delivers 5204

I'd run it first thing when you open it up "cold" and see what happens. It should be scoring a bit higher than that. In real-world tests usually the difference between the m3 and m5 is more pronounced than the difference between the m5 and m7.

 

[pmouritz]

I'd run it first thing when you open it up "cold" and see what happens. It should be scoring a bit higher than that. In real-world tests usually the difference between the m3 and m5 is more pronounced than the difference between the m5 and m7.

Yes, that is exactly what I've been doing. Opening it up and starting the Geekbench 4 test without having any other programs open and when it is 'cold'. It gets 5400 consistently which what I'm kind of worried about. However, in my daily usage (web browsing with many tabs, Maple computations, Office suite, Netflix, etc.) I don't see that it being slow at all. But of course I want the most out of my money, ie. my new rMB to be as future proof as possible.

 

[fisherking]

my score just now, on battery & crashplan NOT disabled, was 5195. so, varies.

really, though, it's the experience that matters most. i use my macbook pro for heavy lifting (logic X, finalcut pro), and the MB for everything else (including affinity photo). it never feels sluggish, and i can do what i need to do.

just saying; don't get caught up in geekbench scores, focus on how it feels to use the macbook. if it feels good, it is good.

 

[pmouritz]

So I just noticed that the Memory Copy, Latency and Bandwidth are the ones that pulls the Multi-core performance score down the most. As seen, https://browser.geekbench.com/v4/cpu/432409, they are very low compared to the other tasks. HTML DOM is also as low as the Memory Copy, Latency and Bandwidth.

 

[KPOM]

Yes, that is exactly what I've been doing. Opening it up and starting the Geekbench 4 test without having any other programs open and when it is 'cold'. It gets 5400 consistently which what I'm kind of worried about. However, in my daily usage (web browsing with many tabs, Maple computations, Office suite, Netflix, etc.) I don't see that it being slow at all. But of course I want the most out of my money, ie. my new rMB to be as future proof as possible.

I got 3242/6018 on Geekbench 4 64-bit. Check Activity Monitor to see if you have something running in the background that might be using up the CPU.

 

[fisherking]

I got 3242/6018 on Geekbench 4 64-bit. Check Activity Monitor to see if you have something running in the background that might be using up the CPU.

nice score; which MB are you running (what processor speed?).

 

[KPOM]

nice score; which MB are you running (what processor speed?).

2016 m5.

 

[pmouritz]

I got 3242/6018 on Geekbench 4 64-bit. Check Activity Monitor to see if you have something running in the background that might be using up the CPU.

Damn that's the kind of score I expected! Great score you have there.
Upon boot, the only one pulling CPU power is Kernel Tasks with around 8-18%. However, I don't think that is uncommon?
I really don't understand my scores as they should be significantly higher.

 

[KPOM]

Damn that's the kind of score I expected! Great score you have there.
Upon boot, the only one pulling CPU power is Kernel Tasks with around 8-18%. However, I don't think that is uncommon?
I really don't understand my scores as they should be significantly higher.

Are you within the return period? It's possible that the thermal paste is defective and causing the processor to throttle. If I were to run the benchmark over and over again the scores would drop to about what you are seeing.

 

[pmouritz]

Are you within the return period? It's possible that the thermal paste is defective and causing the processor to throttle. If I were to run the benchmark over and over again the scores would drop to about what you are seeing.

Yes I'm still within the return period, I bought it last wednesday. This is also the reason for this thread as I would like to know whether or not it is a reason to return it. I figured that there might be varians but 5400 benchmark score seemed too low to me
So the benchmark lays the ground for a return? I'm so happy for the computer in every other aspect, it has great battery life, great keyboard, doesn't feel slow etc. But it would not be acceptable if anything is wrong with either the thermal paste or other things, limiting the performance of the processor.

 

[KPOM]

Yes I'm still within the return period, I bought it last wednesday. This is also the reason for this thread as I would like to know whether or not it is a reason to return it. I figured that there might be varians but 5400 benchmark score seemed too low to me
So the benchmark lays the ground for a return? I'm so happy for the computer in every other aspect, it has great battery life, great keyboard, doesn't feel slow etc. But it would not be acceptable if anything is wrong with either the thermal paste or other things, limiting the performance of the processor.

I'd say go for it. If they ask, be upfront, but in any case Apple will refurbish your return for sale later in its online store. If the replacement m5 exhibits the same I'd suspect software.

Do you have cloning software and an external drive that you can back up your drive with? That would make swapping it out a breeze. If so, before you return it, you might try wiping the internal drive and doing a clean install to see what the benchmark is like without anything else installed, just to rule out software.

 

[KPOM]

Yes I'm still within the return period, I bought it last wednesday. This is also the reason for this thread as I would like to know whether or not it is a reason to return it. I figured that there might be varians but 5400 benchmark score seemed too low to me
So the benchmark lays the ground for a return? I'm so happy for the computer i

I should clarify I am running the golden master build of macOS Sierra. It's possible that is affecting the scores somewhat (updated drivers, etc.). If you are updating to Sierra tomorrow, try re-running it and see if you get better results.

 

[Mike Boreham]

Try installing Intel Power Gadget to show instantaneous CPU temp and processor speed.

https://software.intel.com/en-us/articles/intel-power-gadget-20

The tops of the CPU spikes in IPG during a GB4 run should be at the spec turboboost speed of 2.4 GHz for a multicore test of an M5.

My best Geekbench 4 score is 5992. 2016 MacBook with M5/512

In my experience it does not throttle as much as people think, and first start etc are not that important. I have run HIARCS chess in analysis mode for 15 mins. This uses 350% CPU. Intel Power Gadget shows it running at the expected turbo boost of 2.4Ghz initially and for several minutes. After about 5 mins it is temperature throttled back to 2.2Ghz.

EDIT I just did a Geekbench 4 run with IPG open. GB4 is not particularly demanding because it runs lots of tests of short duration which appear as spikes. Because they spikes the CPU does not get hot enough to throttle. The highest temp I have seen is 70C, whereas the temperature throttling does not happen until about 90C. On my 15 min run of HIARCS at 350% CPU the temp was 88C while throttled back to 2.2.
[doublepost=1474324458][/doublepost]Here is an IPG screenshot during my Geekbench 4 run, showing the Turboboost peaks at 2.4 GHz.

 

[pmouritz]

I'd say go for it. If they ask, be upfront, but in any case Apple will refurbish your return for sale later in its online store. If the replacement m5 exhibits the same I'd suspect software..

Yes, I think that I will just wipe the harddrive and go return it tomorrow and buy a new one from another store. There are no Apple Stores in Denmark. However, the 14-day return policy still stands as long as the computer is clean and does not show any signs of being used.
I don't think that it is worth all the trouble to try to update, reinstall etc. when I can just go get a new one. Then I will also know that if the problem remains it has something to do with the software and not an actual problem with the computers hardware.
[doublepost=1474371473][/doublepost]

My best Geekbench 4 score is 5992. 2016 MacBook with M5/512

Also a great score! That is exactly what should be expected from the M5/512 GB 2016 rMB.. I reckon that Geekbench 4 does not require that much CPU computing power for a longer period of time as my Macbook does not even get hot when I run the benchmark. This is also something that made me wonder since it should not be throttling when doing the benchmark as it does not get hot..

What does it say under processor in your Geekbench 4? For some odd reason mine says 'Intel(R) Core(TM) m5-6y54 @ 1.10 GHz @ 1.2 GHz'. Is it supposed to say this? I've linked a screenshot of my Geekbench information upon running the benchmark.

 

[Mike Boreham]

Try installing Intel Power Gadget to show instantaneous CPU temp and processor speed.

https://software.intel.com/en-us/articles/intel-power-gadget-20

The tops of the CPU spikes in IPG during a GB4 run should be at the spec turboboost speed of 2.4 GHz for a multicore test of an M5.

My best Geekbench 4 score is 5992. 2016 MacBook with M5/512

In my experience it does not throttle as much as people think, and first start etc are not that important. I have run HIARCS chess in analysis mode for 15 mins. This uses 350% CPU. Intel Power Gadget shows it running at the expected turbo boost of 2.4Ghz initially and for several minutes. After 15 mins it is temperature throttled back to 2.2Ghz.

EDIT I just did a Geekbench 4 run with IPG open. GB4 is not particularly demanding because it runs lots of tests of short duration which appear as spikes. Because they spikes the CPU does not get hot enough to throttle. The highest temp I have seen is 70C, whereas the temperature throttling does not happen until about 90C. On my 15 min run of HIARCS at 350% CPU the temp was 88C while throttled back to 2.2.
[doublepost=1474324458][/doublepost]Here is an IPG screenshot during my Geekbench 4 run, showing the Turboboost peaks at 2.4 GHz.

If your IPG spikes are at 2.4Ghz I would suspect other processes are dragging your score down. Have you had a good look at Activity Monitor? as Spotlight finished indexing?

Yes, I think that I will just wipe the harddrive and go return it tomorrow and buy a new one from another store. There are no Apple Stores in Denmark. However, the 14-day return policy still stands as long as the computer is clean and does not show any signs of being used.
I don't think that it is worth all the trouble to try to update, reinstall etc. when I can just go get a new one. Then I will also know that if the problem remains it has something to do with the software and not an actual problem with the computers hardware.
[doublepost=1474371473][/doublepost]
Also a great score! That is exactly what should be expected from the M5/512 GB 2016 rMB.. I reckon that Geekbench 4 does not require that much CPU computing power for a longer period of time as my Macbook does not even get hot when I run the benchmark. This is also something that made me wonder since it should not be throttling when doing the benchmark as it does not get hot..

What does it say under processor in your Geekbench 4? For some odd reason mine says 'Intel(R) Core(TM) m5-6y54 @ 1.10 GHz @ 1.2 GHz'. Is it supposed to say this? I've linked a screenshot of my Geekbench information upon running the benchmark.

That is odd, but I have seen something about this elsewhere. Here is my screen shot of the Geekbench CPU info window. I just got 6043 by turning off some third party background processes I had installed.
[doublepost=1474372424][/doublepost]

If your IPG spikes are at 2.4Ghz I would suspect other processes are dragging your score down. Have you had a good look at Activity Monitor? as Spotlight finished indexing?


That is odd, but I have seen something about this elsewhere. Here is my screen shot of the Geekbench CPU info window. I just got 6043 by turning off some third party background processes I had installed.

EDIT: This Intel chart about the three processors in the MacBooks, shows the M5 with a base CPU frequency of 1.1Ghz, implying Apple overclocks it to 1.2 Ghz. Doesn't explain why Geekbench shows yours different from mine.

 

[pmouritz]

If your IPG spikes are at 2.4Ghz I would suspect other processes are dragging your score down. Have you had a good look at Activity Monitor? as Spotlight finished indexing?


That is odd, but I have seen something about this elsewhere. Here is my screen shot of the Geekbench CPU info window. I just got 6043 by turning off some third party background processes I had installed.
[doublepost=1474372424][/doublepost]

EDIT: This Intel chart about the three processors in the MacBooks, shows the M5 with a base CPU frequency of 1.1Ghz, implying Apple overclocks it to 1.2 Ghz. Doesn't explain why Geekbench shows yours different from mine.

That is so weird that my Geekbench says something different from yours.. I can't seem to understand it.
I tried to install and run IPG and I just ran Geekbench 4 with it open. The highest frequency I obtained was 2.4 GHz for 3 peaks, otherwise they were way lower except for 2.2 GHz a couple of times. I think that your IPG and mine shows pretty much the same.
Once again my Geekbench 4 score was around 5400, this time at 5417 with no other applications running other than IPG and Geekbench 4.

My spotlight have been indexed and I've tried to take a good hard look at the Activity Monitor but can't seem to identify anything using too much CPU.

Ah okay, so Apple overclocks the m5-6y54 processor Maybe it has to do with this overclocking? I don't know.. I don't what to do, either I should spend time wiping down the harddrive and computer and go return it and buy a new one. Then I have to reinstall and download all of my stuff once again since I have no adapter to USB hehe. Or just live with the fact that my Geekbench 4 score is lower than should be expected. There doesn't seem to be any slow downs in my daily usage - I guess, I would just feel silly to not have gotten the entry-level computer if my Geekbench score isn't that much better hehe.

 

[Mike Boreham]

Personally I would not be happy if my new computer was doing what your is. I would want to either fix it or return it.

How did you set it up? did you migrate from an old computer, or install everything fresh?

If you migrated then I would try erasing (booted from the Recovery Partition) and reinstalling OSX and then just installing Geekbench. If that gives a good result then you know it something you have installed.

If you did a fresh install when you got it last week then there is less likelihood of it being something migrated, but could be something else installed.

Do you an external you can clone to so that you can go back to your full install after erasing and reinstalling?

Note my results are with Sierra but I don't think that is a big factor. Sierra is due to be released the evening.

You may just prefer to return it and get another one to remove all doubt.

 

[KPOM]

I think what is happening is that your CPU is not overclocking properly. GeekBench 4 reports 1.2GHz for me, not "1.1 GHz @ 1.2 GHz." Also Intel Power Gadget shows that GeekBench 4 maxes out at 2.4GHz (the maximum dual-core Turbo Boost for the m5) and bounces between 2.2GHz-2.4GHz most of the time. As I recall, GeekBench 3 sustained the CPU a bit more, so perhaps GeekBench 4 is designed to reduce throttling.

 

[pmouritz]

Personally I would not be happy if my new computer was doing what your is. I would want to either fix it or return it.
.

No exactly! I'm not satisfied so I think I'm going to allocate some time tomorrow for the whole return process.. Just very annoyed that it is not working out but I guess I'm glad that I have the opportunity to get it returned and buy a new one. Hopefully the new one will live up to its expectations!

I think what is happening is that your CPU is not overclocking properly. GeekBench 4 reports 1.2GHz for me, not "1.1 GHz @ 1.2 GHz." Also Intel Power Gadget shows that GeekBench 4 maxes out at 2.4GHz (the maximum dual-core Turbo Boost for the m5) and bounces between 2.2GHz-2.4GHz most of the time. As I recall, GeekBench 3 sustained the CPU a bit more, so perhaps GeekBench 4 is designed to reduce throttling.

I think you are totally right. It seems very weird that mine says "@ 1.1 GHz @ 1.2 GHz". I've checked with Geekbench 3, it says the same and the score is even worse. It would make sense that if the CPU did not clock at 1.2 GHz, the performance would be lower. And it is about 10% lower of what it should be which I think is reasonable if the CPU did not clock at the right frequency. The CPU stops at 2.4 at its max, al though it should be able to go as high as 2.7 GHz which I find rather odd.