My new base 2.2 blows away my i9!

[drvelocity]

This might just be the silicon lottery in action, but I just got my base 2.2 Ghz today and it runs significantly faster, with higher sustained clock speeds, and at lower temps than my i9 - even after I repasted the i9 (CPU and GPU).

My average Cinebench score is 1050 on the 2.2 where it was only around 1010 on the i9. The i9 could sustain 2.9Ghz at 100% load pretty much maxing out temps. The 2.2 is sustaining 3.3Ghz and ramps up much slower to 100C although it gets there eventually (as should be).

Turbo boost on the 2.2 only hits 3.9Ghz or so so you lose a split second of .2Ghz from the i9.

Both systems are updated to the newest system update Apple just pushed out.

As far as I'm concerned the 2.2 is absolutely the CPU to get.

Edit: Added screenshots of Cinebench/Power Gadget:

The 2.2 Ghz base:

The 2.9 Ghz i9:

I'm in the process of installing UE4 for a longer benchmark.

 

[mr.anthonyramos]

This might just be the silicon lottery in action, but I just got my base 2.2 today and it runs significantly faster, with higher sustained clock speeds, and at lower temps than my i9 - even after I repasted the i9 (CPU and GPU).

My average Cinebench score is 1050 on the 2.2 where it was only around 1010 on the i9. The i9 could sustain 2.9Ghz at 100% load pretty much maxing out temps. The 2.2 is sustaining 3.3Ghz and ramps up much slower to 100C although it gets there eventually (as should be).

Turbo boost on the 2.2 only hits 3.9Ghz or so so you lose a split second of .2Ghz from the i9.

Both systems are updated to the newest system update Apple just pushed out.

As far as I'm concerned the 2.2 is absolutely the CPU to get.

This is exactly why I honestly cannot trust Cinebench solely. Do a real life comparison by rendering something or applying a work task that you normally do that loads up the CPU.

I have both systems on hand and I’m getting better processing times with my i9 compared with my company issued 2.2. Both updated this morning with the patch. 200 (80mb) RAW files processed in 3 minutes 22 seconds on my i9 and my base i7 work computer does it in 4 minutes 17 seconds.

 

[Standard]

This might just be the silicon lottery in action, but I just got my base 2.2 today and it runs significantly faster, with higher sustained clock speeds, and at lower temps than my i9 - even after I repasted the i9 (CPU and GPU).

My average Cinebench score is 1050 on the 2.2 where it was only around 1010 on the i9. The i9 could sustain 2.9Ghz at 100% load pretty much maxing out temps. The 2.2 is sustaining 3.3Ghz and ramps up much slower to 100C although it gets there eventually (as should be).

Turbo boost on the 2.2 only hits 3.9Ghz or so so you lose a split second of .2Ghz from the i9.

Both systems are updated to the newest system update Apple just pushed out.

As far as I'm concerned the 2.2 is absolutely the CPU to get.

This is interesting. Would you be able to test light bakes in Unreal Engine with both these machines to see how they perform? I don't know what machine to get!

 

[The Mercurian]

This might just be the silicon lottery in action, but I just got my base 2.2 today and it runs significantly faster, with higher sustained clock speeds, and at lower temps than my i9 - even after I repasted the i9 (CPU and GPU).

My average Cinebench score is 1050 on the 2.2 where it was only around 1010 on the i9. The i9 could sustain 2.9Ghz at 100% load pretty much maxing out temps. The 2.2 is sustaining 3.3Ghz and ramps up much slower to 100C although it gets there eventually (as should be).

Turbo boost on the 2.2 only hits 3.9Ghz or so so you lose a split second of .2Ghz from the i9.

Both systems are updated to the newest system update Apple just pushed out.

As far as I'm concerned the 2.2 is absolutely the CPU to get.

You opened up your brand new i9 and redid the thermal paste ???

 

[patchedreality]

This is interesting. Would you be able to test light bakes in Unreal Engine with both these machines to see how they perform? I don't know what machine to get!

Unity 3D dev here. I would also be super interested in real world tests for game development tasks. I have an i7 2.6ghz on order, but I'd consider exchanging it for the i9 if I can shave a minute or 2 off of tasks like building out a Unity scene to XCode, doing an "Assets->Reimport all", doing a clean build to an Android device, building a Unity Xcode project, etc.

Doing an "Assets->Reimport all" and building out to a Mac app, of something like Unity's "Book of the Dead" project might be a good indicator of relative performance? https://assetstore.unity.com/packag...-projects/book-of-the-dead-environment-121175

 

[Standard]

Unity 3D dev here. I would also be super interested in real world tests for game development tasks. I have an i7 2.6ghz on order, but I'd consider exchanging it for the i9 if I can shave a minute or 2 off of tasks like building out a Unity scene to XCode, doing a "Assets->Reimport all", doing a clean build to an Android device, building a Unity Xcode project, etc.

Awesome! Hopefully we can get some more game related tests done on these machines. I'll be using Unity very soon myself, but for now I created a UE benchmark thread. I feel pretty confident if they are quite capable with UE, Unity should have no issues, then.

Perhaps you can share your testing once your machine arrives?

 

[patchedreality]

Awesome! Hopefully we can get some more game related tests done on these machines. I'll be using Unity very soon myself, but for now I created a UE benchmark thread. I feel pretty confident if they are quite capable with UE, Unity should have no issues, then.

Perhaps you can share your testing once your machine arrives?

Sure thing. However, I'll be comparing it to my current late 2013 maxed out MBP 15", which I assume will be blown away by the 2018 i7 2.6ghz... But maybe someone else will have the i9 to run similar tests on.

 

[macintoshmac]

You opened up your brand new i9 and redid the thermal paste ???

The things we apple fans will do for gains in performance.

 

[IronManFanatic]

Is this after the patch? lol.

 

[maflynn]

Here's my cinebench results and I'm seeing 1014 (post patch)

 

[Elektrofone]

Really need to do a series of tests and average the number. I was getting anywhere from 1010-1050 on my i9. Disappointing that the 2.2 is also getting 1050, but also not a real world test.

 

[Chughes]

This is exactly why I honestly cannot trust Cinebench solely. Do a real life comparison by rendering something or applying a work task that you normally do that loads up the CPU.

I have both systems on hand and I’m getting better processing times with my i9 compared with my company issued 2.2. Both updated this morning with the patch. 200 (80mb) RAW files processed in 3 minutes 22 seconds on my i9 and my base i7 work computer does it in 4 minutes 17 seconds.

Interesting, what program do you use to process files? If you don't use Capture One, it would be really helpful to know the speed difference of the same files processed through C1. You can PM me if you want any instructions, if you're willing to test it out. And your photos will thank you if you learn C1, it's much more powerful than LR.

 

[jerryk]

Here's my cinebench results and I'm seeing 1014 (post patch)

View attachment 772713

Nice. I think I am going to pass on the i9.

Likely do the 2.6 pre-configured with 512 and 16GB. I am not convinced that 32 GB is worth the money for me. Most of my VM type work resides in the cloud since I need all the security stuff plugged in.

 

[zshane1125]

Curious about the 2.6ghz benchmarks post update.

 

[maflynn]

Likely do the 2.6 pre-configured with 512 and 16GB. I am not convinced that 32 GB is worth the money for me. Most of my VM type work resides in the cloud since I need all the security stuff plugged in.

I think as it currently stands the 2.2 is as fast as a 2.6. I'd save your money, unless you're looking for the 512 and 560x GPU

 

[Elektrofone]

I think as it currently stands the 2.2 is as fast as a 2.6. I'd save your money, unless you're looking for the 512 and 560x GPU

Nice. I think I am going to pass on the i9.

Likely do the 2.6 pre-configured with 512 and 16GB. I am not convinced that 32 GB is worth the money for me. Most of my VM type work resides in the cloud since I need all the security stuff plugged in.

Struggling with the 16GB vs 32GB myself. It'd be nice to have the overhead, but my Memory Pressure rarely goes above 50%

 

[jerryk]

I think as it currently stands the 2.2 is as fast as a 2.6. I'd save your money, unless you're looking for the 512 and 560x GPU

Yes, I definitely want the 512, and the 560 might be nice to have. I don't care much about the 2.6, but it looks like the readily available system with 512 has it.
[doublepost=1532539772][/doublepost]

Struggling with the 16GB vs 32GB myself. It'd be nice to have the overhead, but my Memory Pressure rarely goes above 50%

Same here. Of course, like a good OS, MacOS allocates as much memory as possible to the running processes. But, I don't think that my memory pressure was ever not green. So little, if any, rapid page faulting.

 

[Elektrofone]

Yes, I definitely want the 512, and the 560 might be nice to have. I don't care much about the 2.6, but it looks like the readily available system with 512 has it.
[doublepost=1532539772][/doublepost]

Same here. Of course, like a good OS MacOS allocates as much memory as possible to the running processes. But, I don't every think that my memory pressure was not green. So little, if any, rapid page faulting.

I did notice a large amount of Swap Memory when rendering in After Effects which is very confusing to me Considering my pressure is never not green.

 

[jerryk]

I did notice a large amount of Swap Memory when rendering in After Effects which is very confusing to me Considering my pressure is never not green.

I am somewhat familiar with the internals of linux systems (since sources and documentation are available), but not MacOS. Each Linux system handles swap a little differently. But, some allocate swap space proactively just in case it is needed. This makes page swapping, if needed, more efficient since it is already been allocated for the process.

Since how MacOS works is not as well documented I have been following along the guidelines they listed on this page, https://support.apple.com/en-us/HT201464#memory. It says "

• Swap Used: The space used on your startup drive by macOS memory management. It's normal to see some activity here. As long as memory pressure is not in the red state, macOS has memory resources available."

I am not sure if I am being overly optimistic, but so far, so good.

 

[mr.anthonyramos]

This is one of the reasons why I don't trust benchmarking applications and will take them with a grain of salt.

I have attached a screenshot from pcmag.com, link below and another screenshot my Cinebench score. The Alienware running the i9 got 1036 and my ("terribly thermally engineered") i9 got 1037. Top of the line Alienware 17 inch, with the i9, 32GB RAM, a WAY WAY more efficient cooling system built for gaming and an overclock feature of up to 5.0ghz gets such a similar score to my i9 MacBook Pro...which will definitely blow my MacBook Pro in most real world tests. So does your 2.2 i7 blow away your 2.9 i9? It does get pretty close to the i9 Cinebench results, but I can bet my life, results will be very different in the real world testing.

https://www.pcmag.com/news/362510/apple-macbook-pro-benchmarked-fast-but-surprise-youll

 

[rsths12]

This is exactly why I honestly cannot trust Cinebench solely. Do a real life comparison by rendering something or applying a work task that you normally do that loads up the CPU.

I have both systems on hand and I’m getting better processing times with my i9 compared with my company issued 2.2. Both updated this morning with the patch. 200 (80mb) RAW files processed in 3 minutes 22 seconds on my i9 and my base i7 work computer does it in 4 minutes 17 seconds.

I know you also had access to a 2.6 i7. Have you tried this test on that machine? I am just curious because this is a real world use for me as I export a lot fo photos from Lightroom for timelapses.

 

[Kpjoslee]

This is one of the reasons why I don't trust benchmarking applications and will take them with a grain of salt.

I have attached a screenshot from pcmag.com, link below and another screenshot my Cinebench score. The Alienware running the i9 got 1036 and my ("terribly thermally engineered") i9 got 1037. Top of the line Alienware 17 inch, with the i9, 32GB RAM, a WAY WAY more efficient cooling system built for gaming and an overclock feature of up to 5.0ghz gets such a similar score to my i9 MacBook Pro...which will definitely blow my MacBook Pro in most real world tests. So does your 2.2 i7 blow away your 2.9 i9? It does get pretty close to the i9 Cinebench results, but I can bet my life, results will be very different in the real world testing.

https://www.pcmag.com/news/362510/apple-macbook-pro-benchmarked-fast-but-surprise-youll



View attachment 772723View attachment 772726

That is like lowest score I have seen with Alienware 17R5. Most of other reviews I have seen scores from 1150 to 1300.

Only advantage i9 has over i7 2.2 and 2.6 is its ability to boost clocks higher, but in thermally constrained chasis like Macbook Pro, its advantage simply disappears. I expect any workload that stresses 6 cores long enough, there wouldn't be much difference between 2.2, 2.6, and 2.9.

 

[mr.anthonyramos]

I know you also had access to a 2.6 i7. Have you tried this test on that machine? I am just curious because this is a real world use for me as I export a lot fo photos from Lightroom for timelapses.

Yup my colleague has the 2018 2.6 i7. Have not tried the i7 on Cinebench yet but I probably won't bother.

And yes! The i7 does the same task with 200 RAW files an extra 32 seconds finishing at 3 minutes and 54 seconds. And consecutive tests show it stays within the 3 minutes and 50 seconds or longer area.
[doublepost=1532545133][/doublepost]

That is like lowest score I have seen with Alienware 17R5. Most of other reviews I have seen scores from 1150 to 1300.

Only advantage i9 has over i7 2.2 and 2.6 is its ability to boost clocks higher, but in thermally constrained chasis like Macbook Pro, its advantage simply disappears. I expect any workload that stresses 6 cores long enough, there wouldn't be much difference between 2.2, 2.6, and 2.9.

Which of course strengthens my case, one time it does 1030 then next time it does 1150 then sometimes it goes to 1300 HUGE differences but the same machine. Currently my highest Cinebench Score is at 1046 and I can't get higher than that, and averages around 1020-1030. Lowest I think I got was around 980 all of which are post patch.

So with that said, I really wouldn't put so much weight on these Cinebench scores and make a claim that this blows that away because in reality, it really doesn't. As I posted above, 32 seconds difference in LR RAW Processing. This was enough for me to say, my 300 bucks was worth it.

Other applications might have less of a difference, while others might have more.

 

[drvelocity]

i just ran the Unreal Engine production lighting test and the 2.2 Ghz is finishing in one minute and 50 seconds while the i9 is taking 2:10.

The i9 for some reason started clocking down halfway through. Could be that I just got a real dud of a processor there.

 

[The Mercurian]

i just ran the Unreal Engine production lighting test and the 2.2 Ghz is finishing in one minute and 50 seconds while the i9 is taking 2:10.

The i9 for some reason started clocking down halfway through. Could be that I just got a real dud of a processor there.

Maybe repasting it made it worse?