Can we crowdsource some 13" performance/throttling testing?

[mikethebigo]

Hi guys!

I snagged a 13" i7 off the shelf at the Apple Store, not too fussed about the i7 itself but wanted the extra RAM/SSD. I'm seeing a lot of talk about the 15", but not too many comparisons about the i5 vs i7 on the 13" MBP.

If someone could give me instructions, I'd be happy to test performance and thermals and give you results. I'd love if someone here with a 13" i5 could also do the same testing and we could do a direct comparison.

Thanks!

 

[matt_k]

Hi guys!

I snagged a 13" i7 off the shelf at the Apple Store, not too fussed about the i7 itself but wanted the extra RAM/SSD. I'm seeing a lot of talk about the 15", but not too many comparisons about the i5 vs i7 on the 13" MBP.

If someone could give me instructions, I'd be happy to test performance and thermals and give you results. I'd love if someone here with a 13" i5 could also do the same testing and we could do a direct comparison.

Thanks!

1. Get Intel Power Gadget for Mac

2. Get Prime95 for Mac

Run Prime 95, Go to Option -> Torture Test, chose "Inplace Large FFT" with 8 threads to run and post screenshot of the Intel Power Gadget immediately after start and after 10 - 20 minutes of running

 

[CodeJoy]

Hi guys!

I snagged a 13" i7 off the shelf at the Apple Store, not too fussed about the i7 itself but wanted the extra RAM/SSD. I'm seeing a lot of talk about the 15", but not too many comparisons about the i5 vs i7 on the 13" MBP.

If someone could give me instructions, I'd be happy to test performance and thermals and give you results. I'd love if someone here with a 13" i5 could also do the same testing and we could do a direct comparison.

Thanks!

Great idea! Just send some MBP 2018 models over to me and I'll do any testing anyone could ever ask for. ;-)

 

[lankox]

Here is some information concerning the i5 in the 2018 13". If you could run this same test and compare with the i7 that would be great.

https://forums.macrumors.com/thread...o-13”-throttling.2128110/page-2#post-26260505

 

[mikethebigo]

They stock, basically secretly, the i7/16/1TB option. If you go to the store website, and pick the highest end ($1,999) MacBook Pro 13, and then change the options to that configuration, you'll suddenly see the pick up in store option change and you can see what local stores have it in the back.

Some Specialists at the store are literally unaware they stock it. I had to show her this info to convince her to look in the system, and then she was like, wow, there it is!

Regarding the rest of this thread, I'll try to get you guys some of that benchmark data when I have time!

 

[DeanL]

Here are my tests
https://imgur.com/a/aXmPJYi
[doublepost=1531958976][/doublepost]

They stock, basically secretly, the i7/16/1TB option. If you go to the store website, and pick the highest end ($1,999) MacBook Pro 13, and then change the options to that configuration, you'll suddenly see the pick up in store option change and you can see what local stores have it in the back.

Some Specialists at the store are literally unaware they stock it. I had to show her this info to convince her to look in the system, and then she was like, wow, there it is!

Regarding the rest of this thread, I'll try to get you guys some of that benchmark data when I have time!

Yup that's how I got mine

 

[Drmash]

They stock, basically secretly, the i7/16/1TB option. If you go to the store website, and pick the highest end ($1,999) MacBook Pro 13, and then change the options to that configuration, you'll suddenly see the pick up in store option change and you can see what local stores have it in the back.

Some Specialists at the store are literally unaware they stock it. I had to show her this info to convince her to look in the system, and then she was like, wow, there it is!

Regarding the rest of this thread, I'll try to get you guys some of that benchmark data when I have time!

Thank you! I will add my test once I have the machine in hand. Though I will admit I'm a little nervous about spending the extra money on the i7 due to thermalgate

 

[Elwe]

Thank you! I will add my test once I have the machine in hand. Though I will admit I'm a little nervous about spending the extra money on the i7 due to thermalgate

While the 13-incher may indeed throttle at some point for some perioed of time, I think you can relax a bit as you remember 1) this is (only) a quad core cpu, and the outrage has mostly been about the six core i9 model; and 2) you can use a third-party fan utility to adjust the fan speeds/curve.

 

[Picapau21]

They stock, basically secretly, the i7/16/1TB option. If you go to the store website, and pick the highest end ($1,999) MacBook Pro 13, and then change the options to that configuration, you'll suddenly see the pick up in store option change and you can see what local stores have it in the back.

Some Specialists at the store are literally unaware they stock it. I had to show her this info to convince her to look in the system, and then she was like, wow, there it is!

Regarding the rest of this thread, I'll try to get you guys some of that benchmark data when I have time!

You’re so right!! Changed to store pick up. Would normally buy but really don’t think I will ever need 1 TB in storage

 

[joelhinch]

Also very interested in a comparison here.. I have orderd an i7/16/512 but considering how well the i5 seems to perform and the concern for throttling on the i7.. it may turn out the i5/16/512 is a better option.

 

[lankox]

Notebookcheck.net posted the following scores in Cinebench for the i5: OpenGL = 33.91fps, cpu = 608 and single core = 154. Now we need someone to do the same with the 2108 i7 13".

 

[mikethebigo]

Notebookcheck.net post the following scores in Cinebench for the i5: OpenGL = 33.91fps, cpu = 608 and single core = 154. Now we need someone to do the same with the 2108 i7 13".

I’ll get Cinebench and run it on my i7 13” in about 2 hours.
[doublepost=1532029948][/doublepost]

Notebookcheck.net post the following scores in Cinebench for the i5: OpenGL = 33.91fps, cpu = 608 and single core = 154. Now we need someone to do the same with the 2108 i7 13".

OK! Here's my results for the i7 13" MBP.

OpenGL score of 37.71 FPS
CPU score of 619
I'm not sure how to run single core? Don't see an option for it.

Enclosed is my Intel Power Gadget info during those tests as well. Briefly turbo clocked to max speed, then it looks like during the CPU test I still averaged a little over 3GHz across the entire test, so still above base clock speed. Pretty encouraging results honestly. The system did get to about 100 C internally, but never felt too hot on the outside and cooled very rapidly once the fans kicked in.

EDIT: figured out how to set to single thread, scored 168. Interestingly, for that test my CPU frequency sat at a consistent 4 ghz, with power draw between 20-25. The fans did kick in, but it looks like for single core tasks the CPU can hold its turbo speed, so it's the multicore stuff that heats it up too much.

 

[lankox]

Thanks mikethebigo for doing theses tests. Though it is somewhat encouraging that the i7 beat the i5 in each test it is still disappointing overall. Comparing your IPG i7 graph with the i5, it is clear the i7 is throttling more. The standard cpu score increase was less than 2% over the i5. Not only that, but the Dell XPS 13" which is using a slower 15W processor still beat the i7 (639). I Would love to know how much increasing the fans would help the i7. If you're willing, you can download the fan control application, and set the fans to max and redo the tests. I have a feeling the cpu score would improve quite nicely. Thanks again.

https://www.crystalidea.com/macs-fan-control

 

[mikethebigo]

Thanks mikethebigo for doing theses tests. Though it is somewhat encouraging that the i7 beat the i5 in each test it is still disappointing overall. Comparing your IPG i7 graph with the i5, it is clear the i7 is throttling more. The standard cpu score increase was less than 2% over the i5. Not only that, but the Dell XPS 13" which is using a slower 15W processor still beat the i7 (639). I Would love to know how much increasing the fans would help the i7. If you're willing, you can download the fan control application, and set the fans to max and redo the tests. I have a feeling the cpu score would improve quite nicely. Thanks again.

https://www.crystalidea.com/macs-fan-control

Honestly under load I don't know why the i7 would throttle any more than the i5. They are identical processors other than the GPU being slightly clocked higher and it having a tiny bit more L3 cache. Neither processor is hitting sustained maximum turbo speeds, so they should sit right at the same average together. I did look at another post that showed their i5 performance and it did seem to fluctuate right around the same average (around 3GHz). Could you maybe link to what you're looking at that looks better?

I will try to do it again with the fans a bit later.

 

[lankox]

When I say throttle, I'm talking in reference to the Base clock speed. The i5 is 2.3 GHz and the i7 is 2.7GHz. The i5 can maintain it's base clock speed under load (actually even higher) while the i7 struggles to maintain it's base clock speed. Essentially, both the i7 and i5 are running at the same speed under load, even though the i7 should be 0.4 GHz faster. Here is the graph from notebookcheck. Notice how the frequency of the i5 stays above the base clock speed of 2.3GHz. Your i7 graph shows the frequency constantly dipping below the base clock speed of 2.7GHz. As it stands right now, spending the extra money for the i7 is a waste when it comes to workloads that max out the processors; but there is value in terms of the iGPU, single core tasks, and burst speeds. With the fans on max, I'm hoping the i7 graph will show speeds above the base clock.

 

[mikethebigo]

When I say throttle, I'm talking in reference to the Base clock speed. The i5 is 2.3 GHz and the i7 is 2.7GHz. The i5 can maintain it's base clock speed under load (actually even higher) while the i7 struggles to maintain it's base clock speed. Essentially, both the i7 and i5 are running at the same speed under load, even though the i7 should be 0.4 GHz faster. Here is the graph from notebookcheck. Notice how the frequency of the i5 stays above the base clock speed of 2.3GHz. Your i7 graph shows the frequency constantly dipping below the base clock speed of 2.7GHz. As it stands right now, spending the extra money for the i7 is a waste when it comes to workloads that max out the processors; but there is value in terms of the iGPU, single core tasks, and burst speeds. With the fans on max, I'm hoping the i7 graph will show speeds above the base clock.

Yeah so basically they're showing same performance under load, instead of the i7 being a little faster under load which you would expect without thermal issues. But, like you said, the i7 will burst faster than the i5, and will show a few more FPS in synthetic graphic benchmarks lol.

It's all rather academic, when I'm doing normal tasks and check my CPU frequency the system has idled it at like 1.2 GHz lol.

 

[Vazza]

So go i5 then?

 

[lankox]

Yeah so basically they're showing same performance under load, instead of the i7 being a little faster under load which you would expect without thermal issues. But, like you said, the i7 will burst faster than the i5, and will show a few more FPS in synthetic graphic benchmarks lol.

It's all rather academic, when I'm doing normal tasks and check my CPU frequency the system has idled it at like 1.2 GHz lol.

You summarized it better than I did!
[doublepost=1532045196][/doublepost]

So go i5 then?

Depends on your work load and the value of your money. Right now, the i7 is approximately 10% faster in single core performance, burst speeds, and the iGPU; while being no faster under heavy loads. I'm hoping that with the fans on max, we will see the i7 about 10% faster than the i5 under heavy load. You will have to decide if that is worth an extra $300 ($270 academic pricing).

 

[yoheywut]

When I say throttle, I'm talking in reference to the Base clock speed. The i5 is 2.3 GHz and the i7 is 2.7GHz. The i5 can maintain it's base clock speed under load (actually even higher) while the i7 struggles to maintain it's base clock speed. Essentially, both the i7 and i5 are running at the same speed under load, even though the i7 should be 0.4 GHz faster. Here is the graph from notebookcheck. Notice how the frequency of the i5 stays above the base clock speed of 2.3GHz. Your i7 graph shows the frequency constantly dipping below the base clock speed of 2.7GHz. As it stands right now, spending the extra money for the i7 is a waste when it comes to workloads that max out the processors; but there is value in terms of the iGPU, single core tasks, and burst speeds. With the fans on max, I'm hoping the i7 graph will show speeds above the base clock.

I can basically throw another data point in agreement with yours. 2018 13" i5, I don't know ways to test the load by rendering, so I just went with trying to convert a 720p MKV to MP4 using handbrake. Using the same Intel power gadget, I looks to me like the i5 did indeed maintain and was actually slightly exceeding it's base clock speed, bouncing around 2.5 GHz and hanging around 95ºC. Only when I played a 4K video in the photos app did I see a drop to around 2.3ish. So I'm happy I guess?

 

[anshuvorty]

I can basically throw another data point in agreement with yours. 2018 13" i5, I don't know ways to test the load by rendering, so I just went with trying to convert a 720p MKV to MP4 using handbrake. Using the same Intel power gadget, I looks to me like the i5 did indeed maintain and was actually slightly exceeding it's base clock speed, bouncing around 2.5 GHz and hanging around 95ºC.

Interesting...I am getting my 13" i5 next week. I will try running these tests and get back with my results then.

 

[lankox]

I can basically throw another data point in agreement with yours. 2018 13" i5, I don't know ways to test the load by rendering, so I just went with trying to convert a 720p MKV to MP4 using handbrake. Using the same Intel power gadget, I looks to me like the i5 did indeed maintain and was actually slightly exceeding it's base clock speed, bouncing around 2.5 GHz and hanging around 95ºC. Only when I played a 4K video in the photos app did I see a drop to around 2.3ish. So I'm happy I guess?

Thanks yoheywut for the information. The more data points the better.

 

[Vazza]

You summarized it better than I did!
[doublepost=1532045196][/doublepost]

Depends on your work load and the value of your money. Right now, the i7 is approximately 10% faster in single core performance, burst speeds, and the iGPU; while being no faster under heavy loads. I'm hoping that with the fans on max, we will see the i7 about 10% faster than the i5 under heavy load. You will have to decide if that is worth an extra $300 ($270 academic pricing).

For my purposes probably not

 

[SBruv]

Got my 13" i7 2.7/16/1TB today, which I'm mostly using for music production and related stuff, and I'm very happy with it indeed.

I'm running some pretty plugin-packed projects, and the 2018 MBP is handling them all so much better than my previous 2016 equivalent. In fact, it gets close enough to my 27" 4GHz 2015 5K in iMac that I'm more than happy to sell that and go fully portable (with a TB3 dock and LG Ultrafine in the studio).

So, on-the-fence musicians, don't stress about all these video rendering benchmarks and whatnot – it's a superb machine for production work.

 

[lankox]

Any way you 13" guys can update your scores with the new fix from Apple? Thanks?

 

[ofarlig]

Any way you 13" guys can update your scores with the new fix from Apple? Thanks?

My i5 gets around 700-720 Cinebench multi and like 155 single.