Mac mini i5 Tests

[F-Train]

I purchased an i5 today, with 8GB of RAM, and have two weeks to decide whether to keep it or exchange it for an i7.

In that time, I need to determine whether the i5 processor will work for me, and I'd like to get a sense of what the limits are these days of 8GB of RAM. I may install 32GB of RAM before the two weeks are out in order to test that as well, although the possible need to exchange the machine has me thinking twice about doing so.*

I have an iMac 27" 5K Retina, i7 4GHz, 32GB of RAM (late 2014), and I'll probably do some comparisons with that as well.

If anyone has particular tests that they would like me to run, let me know. In addition to the standard suite of Mac applications, I run Final Cut Pro X, Compressor, Motion, Lightroom/Photoshop, Logic Pro X and iZotope's RX Advanced and Ozone.

I sometimes record audio via an audio interface (Sound Devices USBPre 2) to Logic, so I also want to find out whether the Mac mini is audible when I do so. I'm not anticipating a problem, but I'll be checking this.

* The Apple store employee told me, without prompting, that installing one's own RAM does not void the warranty. Doesn't mean that I will necessarily do it before I decide whether I'm keeping this machine.

 

[Mol1n]

Can you run Cinebench and let us know of the temps?

 

[tpivette89]

Can you run Cinebench and let us know of the temps?

I too would like to know how it performs during a Cinebench test. I’m on the fence about the i5 vs i7 as well. We’ve all seen the geekbench scores between the two, but running other benchmarks would help in the decision making process also.

 

[jrlcopy]

CPU Cinebench on an i7 & GPU Cinebench on i7 + eGPU

 

[F-Train]

CPU Cinebench on an i7 & GPU Cinebench on i7 + eGPU

Great. If there's interest, I could expand the thread title to include the i7. Which eGPU?

 

[Jorbanead]

I purchased an i5 today, with 8GB of RAM, and have two weeks to decide whether to keep it or exchange it for an i7.

In that time, I need to determine whether the i5 processor will work for me, and I'd like to get a sense of what the limits are these days of 8GB of RAM. I may install 32GB of RAM before the two weeks are out in order to test that as well, although the possible need to exchange the machine has me thinking twice about doing so.*

I have an iMac 27" 5K Retina, i7 4GHz, 32GB of RAM (late 2014), and I'll probably do some comparisons with that as well.

If anyone has particular tests that they would like me to run, let me know. In addition to the standard suite of Mac applications, I run Final Cut Pro X, Compressor, Motion, Lightroom/Photoshop, Logic Pro X and iZotope's RX Advanced and Ozone.

I sometimes record audio via an audio interface (Sound Devices USBPre 2) to Logic, so I also want to find out whether the Mac mini is audible when I do so. I'm not anticipating a problem, but I'll be checking this.

* The Apple store employee told me, without prompting, that installing one's own RAM does not void the warranty. Doesn't mean that I will necessarily do it before I decide whether I'm keeping this machine.

Interested in Logic and FCP performance! As well as noise while recording.

 

[tpivette89]

Great. If there's interest, I could expand the thread title to include the i7. Which eGPU?

Looks like he had a Vega 56 based on the cinebench results. Dont know what box it was installed in, though
[doublepost=1541822832][/doublepost]

CPU Cinebench on an i7 & GPU Cinebench on i7 + eGPU

There was some speculation on whether or not the i7 would perform similarly to the dual X5690 Xeon setup in a 4,1/5,1 Mac Pro. Even though the two setups performed similarly in Geekbench multi-core scores, the cinebench CPU tests clearly favors the dual X5690s over the i7 in the new Mini.

Not that there is anything wrong with that... it is 12 cores vs 6. That being said, it seems as though the cMP still has alot going for it even beyond the expandability aspect.

 

[jrlcopy]

Great. If there's interest, I could expand the thread title to include the i7. Which eGPU?

Looks like he had a Vega 56 based on the cinebench results. Dont know what box it was installed in, though

I'm rocking the Razer Core X + Sapphire Vega RX 56.

If I was OP, I'd 100% exchange for the i7, just having double the threads alone on the processor would go towards longevity.

 

[marc_b]

Do your i7 fans spin up a lot in regular use?

And am I seeing that right that adding an eGPU reduces CPU temps a lot?

 

[jrlcopy]

Do your i7 fans spin up a lot in regular use?

And am I seeing that right that adding an eGPU reduces CPU temps a lot?

Barely spin up, this thing is pretty silent with my 50+ tabs, spotify, typing going on.
And yeah, the first night I had it set up, this thing was VERY hot to the touch, but ever since I had the eGPU hooked up, no real thermal issues other than that 100F throttling, but honestly it's pretty much a non issue.

 

[Mol1n]

but ever since I had the eGPU hooked up, no real thermal issues other than that 100F throttling

How did the eGPU help the core temps of the processor under load?

 

[Mol1n]

Here is my Cinebench run on my i5. Definitely hit the 100C mark quick and the clock speed took a dive.

Will be going for the i3.

 

[F-Train]

Here is my Cinebench run on my i5. Definitely hit the 100C mark quick and the clock speed took a dive.

Will be going for the i3.

Thanks for that.

After an amusing but instructive experience getting my i5 running (see https://forums.macrumors.com/threads/three-pieces-of-advice-when-setting-up-a-mac-mini.2153979/), I'll start posting my own test results tomorrow.

 

[tpivette89]

Here is my Cinebench run on my i5. Definitely hit the 100C mark quick and the clock speed took a dive.

Will be going for the i3.

Do you have any fan control software to maybe kick the fan on sooner so the throttling isn’t as bad, or possibly avoided alltogether?

I’m guessing the cinebench test was pretty much a wash after the clock speed plummeted.

 

[Mol1n]

Do you have any fan control software to maybe kick the fan on sooner so the throttling isn’t as bad, or possibly avoided alltogether?

I’m guessing the cinebench test was pretty much a wash after the clock speed plummeted.

The temp vs clocks comparison shows that despite the fan ramping up, the temps were never brought down even at the lowered clock speeds. That's why turboboost didn't ever boost back up to 3.9 again. All the fan did was to keep the cpu from down clocking even more. I think 3.7ghz@99C is all this computer can do under high loads.

 

[Spectrum]

The temp vs clocks comparison shows that despite the fan ramping up, the temps were never brought down even at the lowered clock speeds. That's why turboboost didn't ever boost back up to 3.9 again. All the fan did was to keep the cpu from down clocking even more. I think 3.7ghz@99C is all this computer can do under high loads.

Which is fine isn't it? That is still above its base clock. And you had 6 cores active whereas the i3 will only have 4. So even if the i3 manages to stay at 3.6 ghz, the i5 will still have completed the task in 2/3 the time.

Or am I misunderstanding the graph? I assume the render task was only active during the bock of high wattage activity, and then it finished. At which point the temp, wattage and speed all dropped off to idle.

 

[F-Train]

I'm keen to see how people interpret this data, given that I know nothing about Cinebench. I exited all other applications before running this.

i5 Cinebench OpenGL:

[doublepost=1541947363][/doublepost]i5 Cinebench OpenGL Reference Match & CPU:

[doublepost=1541947421][/doublepost]i5 Cinebench Single Core:

 

[archer75]

Do your i7 fans spin up a lot in regular use?

No. Mines at 85c right now and the fan is just under 1700rpms. I can't hear it at all.

 

[tpivette89]

I'm keen to see how people interpret this data, given that I know nothing about Cinebench. I exited all other applications before running this.

The higher number the better with Cinebench.

The scores aren't bad for what it is... 6-core CPU with no hyper threading and an integrated GPU.

For comparison, here is my early 2009 Mac Pro with the best CPUs that can be installed (6-core 3.46ghz x 2), and a Radeon HD 7970 graphics card

 

[gusping]

No. Mines at 85c right now and the fan is just under 1700rpms. I can't hear it at all.

What is the temp like under light use? E.g. spotify, 10 tabs open in safari including videos, outlook and 3 word docs. My 2014 MBP can go pretty crazy in terms of fans and temps just from that.

 

[pl1984]

I'm keen to see how people interpret this data, given that I know nothing about Cinebench. I exited all other applications before running this.

i5 Cinebench OpenGL:

Here is another Mac Pro point of reference (stock 2010 model except memory was upgraded to 32GB):

Edit: From my 2013 Mac Pro:

 

[archer75]

What is the temp like under light use? E.g. spotify, 10 tabs open in safari including videos, outlook and 3 word docs. My 2014 MBP can go pretty crazy in terms of fans and temps just from that.

I can give you a better idea when it's setup. Right now I'm transferring a ton of data over from 3 different sources. 9 tabs open in safari, Plex serving a 4k movie to ATV4k, transmission open, Sonarr open, activity monitor, intel power gadget, mail, textedit all open. Temp is bouncing all around but I would say 82c average and the core is cranked to the turbo boost frequency of 4.3ghz.
CPU proximity at 60c
memory proximity at 57c
airport card 51c
gpu 75c

Fan is at 2100rpm and I don't hear it.

 

[Jorbanead]

I can give you a better idea when it's setup. Right now I'm transferring a ton of data over from 3 different sources. 9 tabs open in safari, Plex serving a 4k movie to ATV4k, transmission open, Sonarr open, activity monitor, intel power gadget, mail, textedit all open. Temp is bouncing all around but I would say 82c average and the core is cranked to the turbo boost frequency of 4.3ghz.
CPU proximity at 60c
memory proximity at 57c
airport card 51c
gpu 75c

Fan is at 2100rpm and I don't hear it.

This sounds promising! I've been seeing a lot of people complaining about the thermals, but half of them are stress-testing the CPU until it hits 100C which is only pertinent to a small number of workflows. From your early findings, it's looking more like it can handle boost clocks for much longer. 4.3ghz is good!

 

[pl1984]

This sounds promising! I've been seeing a lot of people complaining about the thermals, but half of them are stress-testing the CPU until it hits 100C which is only pertinent to a small number of workflows. From your early findings, it's looking more like it can handle boost clocks for much longer. 4.3ghz is good!

I would recommend benchmarking with something other than Geekbench and Cinebench. Both tests do not tax the system for very long and thus might hide thermal throttling (or other) issues.

Perhaps something such as Handbrake as it's known to utilize a lot of cores for long periods of time. It's free software anyone can easily download and run. We can find a sample video to download and transcode.

 

[F-Train]

I would recommend benchmarking with something other than Geekbench and Cinebench. Both tests do not tax the system for very long and thus might hide thermal throttling (or other) issues.

Perhaps something such as Handbrake as it's known to utilize a lot of cores for long periods of time. It's free software anyone can easily download and run. We can find a sample video to download and transcode.

I have both Handbrake and Compressor. Which do you want to use, what video and what transcoding?

I could also use a video of my own that's on YouTube.