The standard lowest confit is always for show. Same with Surface Pro. You can buy the 699 version but that’s just a toy. For a real SP, you need to spend 2100 all included. For a real MBP15 you need to spend 4300 minimum all inclusive.
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.
New 8-Core MacBook Pro Offers Solid Performance Improvements According to Benchmark
- Thread starter MacRumors
- Start date
- Sort by reaction score
The standard lowest confit is always for show. Same with Surface Pro. You can buy the 699 version but that’s just a toy. For a real SP, you need to spend 2100 all included. For a real MBP15 you need to spend 4300 minimum all inclusive.
this brings up an important question about Geekbench. It seems that some influencers might be purposely driving down average Geekbench scores, by perhaps running other apps during the test (or otherwise providing ridiculously skewed results). For example, the quoted test results are: "Comparatively, the high-end 2018 MacBook Pro has earned an average single-core score of 5348 and a multi-core score of 22620. Single-core speeds are up almost 10 percent, while multi-core scores are up an impressive 29 percent."
Now look up the CPU 8950HK on the geekbench site http://browser.geekbench.com/v4/cpu/search?dir=desc&q=8950hk+macbook&sort=multicore_score and you get a lot of scores in the >25000 range for the same quoted MacBook. This does not seem like natural variability. In fact if you skip all the high scores (like the first 15 pages or so) and go to the last page (204), you see the results that are ridiculous.
Now having written and performed benchmarks professionally, I will tell you that there are a lot of people who just don't get it when it comes to performing benchmarks. Ideally you test under consistent conditions and eliminate all variables that would skew the results. If there are other factors involved that are relevant, redo the tests measuring those factors and report the results separately. It is downright absurd to report an i9-8950HK CPU with a multicore score of 2316 with no explanation (yes someone did that). Maybe the freezer guy, reran his test with the MacBook in the oven?
Now look up the CPU 8950HK on the geekbench site http://browser.geekbench.com/v4/cpu/search?dir=desc&q=8950hk+macbook&sort=multicore_score and you get a lot of scores in the >25000 range for the same quoted MacBook. This does not seem like natural variability. In fact if you skip all the high scores (like the first 15 pages or so) and go to the last page (204), you see the results that are ridiculous.
Now having written and performed benchmarks professionally, I will tell you that there are a lot of people who just don't get it when it comes to performing benchmarks. Ideally you test under consistent conditions and eliminate all variables that would skew the results. If there are other factors involved that are relevant, redo the tests measuring those factors and report the results separately. It is downright absurd to report an i9-8950HK CPU with a multicore score of 2316 with no explanation (yes someone did that). Maybe the freezer guy, reran his test with the MacBook in the oven?
To be fair, there are a lot of professional workflows that consist of mostly short-bursts of CPU-taxing tasks that already end before (or just as) the machine reaches its' thermal limit. In those, performance gains may very well be comparable to those of the Geekbench benchmark and other "short" benchmarks. In other, more continuously taxing workloads, not so much. That's why it's important to find benchmarks that are comparable (in length and in many other factors) to the tasks you intend to do on the machine. But to dismiss them as completely useless is equally dangerous as taking them as an irrefutable one-size-fits-all measure for how much faster the machine is.
Last edited:
Nope...spend the money on the Vega 16 GPU or the Vega 20 GPU. Apple offers it to complete the build out instead of hearing from people that they aren't offering it, but the extra $ is not going to make for a substantially faster system if you already have the 2.3GH Core i9.
Spend it on GPU, SSD, DRAM or AC+ before the CPU, unless you just have to have bragging rights. Just my 2¢.
[doublepost=1558639372][/doublepost]
Which has zero to do with Apple, this MacBook Pro, Apple's other computers or ANY PC OEM's computers that are going to take a similar hit in performance if the user opts to implement the sort of "fix" or mitigation for Intel's CPU vulnerabilities that Apple, Microsoft, et al. suggest. Thanks, Intel!If you implement the fixes for Intel’s most recent vulnerability then is it really more of an up to 30% performance loss on single-core and up to 11% loss on multi-core performance?
Run GeekBench while watching intel Power Gadget. It does not push the CPU to a continuous full load anyway.
My general opinion on this is that $200 is such a trivial amount of money compared to the work done on a machine for 3+ years that, what the heck, why not (its very, very unlikely to hurt you)? That said, we don't really know how much of difference this will bring. Right now, based on just one video, we have the possibility of a bit of extra thermal headroom on the 2.3 compared to the previous model. Likely enough to get some benefit from that extra 2.4-5.0GHz. But until we see a few more professional reviews and lots more user reviews with both models, we won't know for sure.
I use mainly FCPx / Logic and CC / Wondering if I'd notice any 'real world' differences between the 2.3 and 2.4 - or if I should put the extra $200 towards an eGPU setup instead...... I tend to keep these machines for about 3 years.
If someone needs an ultra-fast 1-4 TB SSD, then go for it, but that seems like a limited use case. Why would anyone use an SSD that has 3.2GB/s sequential read speeds and Up to 2.2GB/s sequential write speed for archival storage (as comparison, the Razer has a max read speed of 1GB/sec, and write speeds around 25% of the MacBook Pro)? And if you say you need that much for workflows, why are you even talking about a laptop?
Hey I'm not a fan of the high price SSDs (and less so for RAM), but at least the SSDs are fast and somewhat worth it.
Not that you are totally wrong, you are just a bit too extreme for a real MBP. The base $2800 gets you the 8 core i9, 16GB, and 512 of fast SSD. (I have only seen limited use cases for 32GB of ram, but that costs 400), and that's at list price - you can get lower prices
The surface pro you quote (which is weird because they aren't even in the same performance class). You can get a quad core i7, 512 of much slower SSD, 16 GB (32 GB not even offered) for $1900. Overall about half the performance of the MBP. So no one would realistically equate these two. Oh and MS charges the same $400 to upgrade to a 1TB drive, but theirs is nowhere near as fast as the MBP
The bottom line is that the proof will be in the pudding when additional reviews are released onto YouTube with various benchmarks, more varied tests, etc. The YouTube video posted by @fokmik (Thanks!) was helpful enough to me to draw some basic conclusions and to be hopeful that additional testing bears out the results. If the Core i9-9880H ends up being able to run for sustained periods at 2.7-2.9GHz while keeping around 85-88c I think Apple has a real winner. Anyone expecting it to boost to 3.8-4.0GHz for sustained periods is simply not dealing with reality...and anyone expecting it to run 4.8GHz for more than the a few seconds with the most rudimentary single threaded task had better brace themselves for a gut check. If the CPUs cannot maintain the base clock for sustained periods, then Apple has a problem.
Cautiously optimistic that both the new CPUs and Apple's changes to the keyboard are both what I hoped for them to be, which is a more reliable and performant MacBook Pro.
Cautiously optimistic that both the new CPUs and Apple's changes to the keyboard are both what I hoped for them to be, which is a more reliable and performant MacBook Pro.
Apple already has the 2019 MBP on its keyboard replacement program? How new could the keyboard be? Takes guts to release a product you already know has a high chance of failure. https://www.apple.com/support/keyboard-service-program-for-mac-notebooks/
Perfectly reasonable opinion even though it runs counter to mine. I think what is implied here though is a fixed budget. Ultimately, that's unlikely to be the case, and instead you should think of each choice compared to the margin opportunity cost, which may even extend outside the computer and into unrelated items.
Say the 2.3 vs 2.4 perform about as expected, single and multicore are all 4% faster. That's not likely to be felt very often, true. So a 1 hr job finishes in in like 57 or 58 minutes... meh. That's won't be particularly well felt. But it does have the potential to add up. During a day of wait for this to finish, do set this job up, wait for it to finish, repeat, it might mean saving 20 minutes on your day...? Sure, there is the aspect of just doing something else while you wait (like commenting here!), but sometimes that's not particularly fair or possible. Anyway, for $200, you don't have to save ~20min per day that very often to make it worth it.
2007 15" MacBook Pro 2.2GHz Core 2 Duo T7500 - Single-Core: 1364, Multi-Core: 2315
2019 15" MacBook Pro 2.3GHZ Core i9-9880H - Single Core: 5879, Multi-Core: 29,184
Your point?
MHZ is not the be-all end-all number in regards to performance.
If computers simply needed more GHz to compute faster, we would be at 10GHz+ by now. But they don't so we are not.
Well, that was the original intention by Intel. Back in the Pentium 4 days, Intel believed that to make their CPU's faster, they just needed to pump more MHZ through.
problem was, the more MHZ, the more energy needed. the more energy produced more heat. By the time they hit 5ghz, the CPU's were portable thermal generators.
it was AMD that came in from out of nowhere and proved you didn't necessarily need the most MHZ to make your CPU fast. more cores, more threads and smaller pipelines for more efficiency that forced intel to go back to the drawing board. Throw out the tech that was in the pentium 4's (Netburst) and focus on efficiency.
Above the Geekbench results will be the real world YouTube videos that can give us some thermal stats.
The Vega 16/20 makes it run a bit hotter, the fans are a bit louder, and the energy consumption is increased slightly, reducing the battery life slightly. For folks who really need a high performance GPU, an eGPU would be a much better option.
As far as storage, I fully agree that Apple’s prices are insanely high. NVMe storage is currently $150 for 1TB, $300 for 2TB, Apple charges several times that. The MacBook Pro doesn’t even have upgradable storage since 2015, forcing users to go with external storage to avoid Apple’s insane prices for internal SSD.
They’ll probably release the 16” with a redesign, OLED screen, super efficient cooling, better graphics and history’s best keyboard in like 3 months for the same amount. Seems to be the way it’s going these days.
That SSD upgrade price is absolutely insane. And extra 256GB for $200, that's just nuts. As you said, 1TB NVMe can be had for less. Its also just superbly lame to have to check an upgrade box to 512GB on a laptop that's north of $2200.