Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MBP 15" 2.9ghz gets beat in benchmarks by 2015 2.8ghz?
- Thread starter RumorConsumer
- Start date
- Sort by reaction score
Interesting also that the Mid 2014 machine keeps finding itself in the best bang for the buck category. I think pound for pound thats the best value of all.
id be interested to see how a Maxon Cinema 4D's rendering compares, Physical rendered use CPU and viewport uses GPU. The tests are also easy to do side by side.
I'm considering upgrading if its an improvement.
I'm considering upgrading if its an improvement.
It's the edram cache that the cheaper skylakes lack. I even put some benchmarks up in an earlier thread where broadwell running over 500 MHz slower than haswell was keeping up haswell without edram with as, you guessed it, has an edram cache.
Every single rmbp 15 since 2012 had the highest end igpu. The 2016 dongle edition is the first without top of the line igpu
Every single rmbp 15 since 2012 had the highest end igpu. The 2016 dongle edition is the first without top of the line igpu
I would use R, Matlab, check compilation of large LaTeX documents as well as compiler performance for software builds (the later two will obviously slightly benefit from the SSD as well). These are things that I use daily in my work.
I think its fairly disappointing that the common reviews focus on hard t put into perspective synthetic benchmarks as well as media content creation, but completely ignore another very large and load group of Apple users: the scientific community. A good review should look at the computer from different aspects and evaluate its merits in regards to various possible workflows.
I'm a music producer and just ordered the 2,9 MBP for live and movie studio use, so CPU power is the main concern for me. I did some homework on the 6920HQ and found out that :
-haswell cpu's had higher turboboost, which is shown in synthetic benchmarks as Geekbench, but under sustained load on multiple CPU cores (which is needed for audio workstations) performed at actual 2,8-3,0 Ghz frequency, ref. read this really good review about the 2015 macbook pro in performance and processor sections:
http://www.notebookcheck.net/Apple-MacBook-Pro-Retina-15-Mid-2015-Review.144402.0.html
-confirmed in the same article, because of this, 4980hq where roughly on a par with other 2014 or 2015 options under sustained load (out of experience, in concrete DAW use there is not much difference between different MBP)
-skylake cpu's have the important advantage for notebooks of the lower tdp, resulting in less (no?) throttling when under sustained load. On top of that skylake is presumed to deliver 5 to 10% speed bump per (actual) Ghz vs Haswell perf./Ghz ratio.
--> You can actually see this in the Cinebench R15 CPU scores which should measure real (as opposed to theoretical), sustained performance. 2014 and 2015 MBP's scored around 600 on this multicore cpu test (see article above). While the 2016 mid-tier MBP (6820HQ) as reviewed by Ars Technica scored a promising 700 on the same test (this is the only cinebench CPU test I could find for the new 15 inch MBP). See http://arstechnica.com/video/2016/11/the-2016-13-and-15-inch-touch-bar-macbook-pros-reviewed/
--> I didn't find 6920HQ macbook pro cinebench CPU scores yet (anyone please? or Prime 95 etc.?), but if there are no throttling issues we could be expecting around 750 on the same multicore test (see other 6920HQ laptops). If this is true we could thus be expecting 10 up to 25% faster raw multicore cpu power when on high load when compared to the previous MBP, which would be the first significant speed bump in years for the 15-inches... We payed enough for it, so we may always hope so
-Perhaps a detail, but i read somewhere that speed stepping is handled differently in skylake cpu (not by OS but within system), resulting in lower latencies when sudden bursts of cpu power are needed. This wouldn't be accurately measured yet in geekbench. (sorry but I lost the source, it was on macrumors however)
-haswell cpu's had higher turboboost, which is shown in synthetic benchmarks as Geekbench, but under sustained load on multiple CPU cores (which is needed for audio workstations) performed at actual 2,8-3,0 Ghz frequency, ref. read this really good review about the 2015 macbook pro in performance and processor sections:
http://www.notebookcheck.net/Apple-MacBook-Pro-Retina-15-Mid-2015-Review.144402.0.html
-confirmed in the same article, because of this, 4980hq where roughly on a par with other 2014 or 2015 options under sustained load (out of experience, in concrete DAW use there is not much difference between different MBP)
-skylake cpu's have the important advantage for notebooks of the lower tdp, resulting in less (no?) throttling when under sustained load. On top of that skylake is presumed to deliver 5 to 10% speed bump per (actual) Ghz vs Haswell perf./Ghz ratio.
--> You can actually see this in the Cinebench R15 CPU scores which should measure real (as opposed to theoretical), sustained performance. 2014 and 2015 MBP's scored around 600 on this multicore cpu test (see article above). While the 2016 mid-tier MBP (6820HQ) as reviewed by Ars Technica scored a promising 700 on the same test (this is the only cinebench CPU test I could find for the new 15 inch MBP). See http://arstechnica.com/video/2016/11/the-2016-13-and-15-inch-touch-bar-macbook-pros-reviewed/
--> I didn't find 6920HQ macbook pro cinebench CPU scores yet (anyone please? or Prime 95 etc.?), but if there are no throttling issues we could be expecting around 750 on the same multicore test (see other 6920HQ laptops). If this is true we could thus be expecting 10 up to 25% faster raw multicore cpu power when on high load when compared to the previous MBP, which would be the first significant speed bump in years for the 15-inches... We payed enough for it, so we may always hope so
-Perhaps a detail, but i read somewhere that speed stepping is handled differently in skylake cpu (not by OS but within system), resulting in lower latencies when sudden bursts of cpu power are needed. This wouldn't be accurately measured yet in geekbench. (sorry but I lost the source, it was on macrumors however)
Yes, but I won't receive it before January, unfortunately.
[doublepost=1479471118][/doublepost]
Yep, that was a big feature in Broadwell and further improved in Skylake. Thats also the reason why the 12" MacBook is so fast for browsing and office tasks, it can give you this burst of performance exactly when needed and then clock down just as quickly before the heat becomes an issue.
I am no expert on any of this, but the lack of progress is why I am still sat here working on my cMBP (mid-2012).
I work docked 95% of the time, so retina is not relevant for me, nor would the Touch bar be.
I have a processor with reasonable specs (2.6 GHz Ivy Bridge), with GeekBench of 3610 / 11473, so we are talking about a 25% increase for the new machines. I have added a Samsung 1 Tb SSD, nowhere near the R/W speeds of the recent MBPs but at over 500 Mb/sec no slouch either.
I have a second HDD in there (no optical drive anymore), so could if I want put a second SSD and a RAID array (internal!). For me it is a useful internal backup. Loads of ports and 16 Gb RAM (Apple says 8 Gb max, but it takes 16Gb).
No reason to spend £3000 for me...
I work docked 95% of the time, so retina is not relevant for me, nor would the Touch bar be.
I have a processor with reasonable specs (2.6 GHz Ivy Bridge), with GeekBench of 3610 / 11473, so we are talking about a 25% increase for the new machines. I have added a Samsung 1 Tb SSD, nowhere near the R/W speeds of the recent MBPs but at over 500 Mb/sec no slouch either.
I have a second HDD in there (no optical drive anymore), so could if I want put a second SSD and a RAID array (internal!). For me it is a useful internal backup. Loads of ports and 16 Gb RAM (Apple says 8 Gb max, but it takes 16Gb).
No reason to spend £3000 for me...
You are absolutely right. Computers are tools and not toys. People should upgrade to a new laptop only when performance becomes unacceptable or is hurting their workflow.
I have waited 4 long years before taking the plunge on a retina screen. Most times waiting (and extending the use of current setup) is the right choice.
Agreed, and I was almost surprised about his positive spin. He addressed all the concerns but with an eye to the not-so-distant future instead of being tied to the past.
It's mostly true but there are certain multimedia improvements in each generation, but normally Windows is better optimised for those than macOS.
This is indeed mostly true, especially for desktops. For notebooks however the cooler, more efficient operation contributes to better actual performance.