Is the processor upgrade on the new 16 MBP worth it?

[wallysb01]

Few percent if anything, nothing overwhelming. If needed you'd already know, save your $$$. Smart money is on waiting for the reputable reviews before you make a decision.

Also worth noting a delay also allows Apple's providers to settle the manufacturing process, of which there are always issues initially.

Q-6

I'm not sure this is great advice. We saw definite improvements in the mid-2019 models moving from the i7-> 2.3 i9, but less so from the 2.3 i9 -> 2.4 i9. Power draw on those machines was about 45-50 watts in sustained workloads. If Apple is right and this thing can draw ~20% more power, we could see the chips differentiate themselves more, particularly the i9s from the i7s. Many of the "good" benchmarks (likely run after the system stopped indexing itself or doing other background tasks), showed all versions obtaining and maintaining some kind of boost frequency for a cinebench run (often around 3GHz for the i7s and i9s). Which makes a ton of sense, given that cinebench scores were a bit less different than the math (2.4*8 vs 2.6*6) would imply. So the 6 core was boosting a bit higher than the i9s, but not enough to overcome the 2 core disadvantage. The difference between the 2.3 and 2.4 i9s though was very slight.

I suggest people go check out this thread if interested in the same chips in the old body.

As for how it impacts this decision. If you are single core workloads, the top turbo boost of the i9s is a something to be very aware of. A single core work load I'm guessing won't be thermally constrained, so you should be pegging the top turbo of the chip most of the time. That's 4.5, 4.8 or 5.0. So $400 gets you ~11% faster CPU for the life of the machine in single core workflows. For multicore, the benchmarks will tell us if the added 12Ws is both real and how much it impacts the i7 vs i9 choice. Previously, the i7 was a touch closer to the i9 than it should be on paper, but that might change. It should be about a 23% difference (2.4*8 / 2.6*6), but higher available wattage might impact things in either direction, it really could just come down to the efficiency of each chip at maintaining what turbo at that power envelope.

The other practical consideration is that the 2.3 i9 is not available without the 5500M and 1TB of RAM. So value seeking CPU users might be inclined to pick up the 2.4 i9 CPU upgrade on the base model and leave many of the other options unchecked. I'll likely do this for example, because the GPU is largely wasted on me and spec-ing out the base model to match a slightly upgraded top model suggests that 5300M -> 5500M upgrade is costing $100.

tl;dr: The CPU upgrades are almost certainly going to have some benefit, but it might be slightly less than what the frequency math implies. Decide for yourself if its worth it.

 

[ilikewhey]

It would be interesting to see if they are using multiple M2s instead of just one as many of the other Workstation Class Windows laptops are doing today.

who apple? they havent use traditional m.2 since 2013. thats why there are custom m.2 converters for 2014 and 2015 mbps. for 2016 they completely moved away from replaceable socket all together and only use solder storage now.

 

[wallysb01]

It would be interesting to see if they are using multiple M2s instead of just one as many of the other Workstation Class Windows laptops are doing today.

For 8TB don't they kind of have to be using 2? Probably even for 4TB, right?

 

[pshufd]

For 8TB don't they kind of have to be using 2? Probably even for 4TB, right?

I think that you're right as I have never seen a 4 TB M.2 before.

It will be interesting to see a teardown. What would be most cool is if they left the second slot open to add one.

 

[poorcody]

Apple uses a proprietary SSD controller in their T2 now, so they are not as limited in their design choices.

From Anandtech:

While I don’t expect Apple to be alone here for too long, for practical purposes this is a new record for laptop storage; the only other laptops I know that come close are large laptops that are RAIDing together two SSDs. This, I suspect, is Apple flexing its muscles on the chip and integration front; producing their own SSD controller (as part of the T2) means they aren’t reliant on component suppliers in the same way that other vendors are.

 

[keithop]

Is the i7 likely to run cooler than the i9 when running boring old office stuff? low noise more important to me this time than flat out performance

 

[pshufd]

Apple uses a proprietary SSD controller in their T2 now, so they are not as limited in their design choices.

From Anandtech:

The Dell Precision 7740 can accomadate 12 TB I believe. 2 x M.2 and 2 x SATA3. I believe that there are several other workstation-class laptops that can do this as well. Those laptops can typically accommodate 128 GB of RAM as well.

 

[wallysb01]

Is the i7 likely to run cooler than the i9 when running boring old office stuff? low noise more important to me this time than flat out performance

No. In the "hurry up and wait" type workflows now, once some operation is done, CPUs go into idle states and you won't be able to tell the difference. During that bursty work load (say a column sort or something), the power draw might peak out higher, but it should also finish sooner and achieve that wait state sooner.

 

[keithop]

No. In the "hurry up and wait" type workflows now, once some operation is done, CPUs go into idle states and you won't be able to tell the difference. During that bursty work load (say a column sort or something), the power draw might peak out higher, but it should also finish sooner and achieve that wait state sooner.

thanks for the help. I had the "pleasure" of a noisy i9 in the previous gen and even after all the updates, unless I turned the turbo boost off, the fans were on all the time just doing office work whereas people claimed the i7 was quieter. hoping to avoid that this time but of course would prefer the i9 for the tiny upgrade cost if there's really no difference.

 

[Queen6]

I'm not sure this is great advice. We saw definite improvements in the mid-2019 models moving from the i7-> 2.3 i9, but less so from the 2.3 i9 -> 2.4 i9. Power draw on those machines was about 45-50 watts in sustained workloads. If Apple is right and this thing can draw ~20% more power, we could see the chips differentiate themselves more, particularly the i9s from the i7s. Many of the "good" benchmarks (likely run after the system stopped indexing itself or doing other background tasks), showed all versions obtaining and maintaining some kind of boost frequency for a cinebench run (often around 3GHz for the i7s and i9s). Which makes a ton of sense, given that cinebench scores were a bit less different than the math (2.4*8 vs 2.6*6) would imply. So the 6 core was boosting a bit higher than the i9s, but not enough to overcome the 2 core disadvantage. The difference between the 2.3 and 2.4 i9s though was very slight.

I suggest people go check out this thread if interested in the same chips in the old body.

As for how it impacts this decision. If you are single core workloads, the top turbo boost of the i9s is a something to be very aware of. A single core work load I'm guessing won't be thermally constrained, so you should be pegging the top turbo of the chip most of the time. That's 4.5, 4.8 or 5.0. So $400 gets you ~11% faster CPU for the life of the machine in single core workflows. For multicore, the benchmarks will tell us if the added 12Ws is both real and how much it impacts the i7 vs i9 choice. Previously, the i7 was a touch closer to the i9 than it should be on paper, but that might change. It should be about a 23% difference (2.4*8 / 2.6*6), but higher available wattage might impact things in either direction, it really could just come down to the efficiency of each chip at maintaining what turbo at that power envelope.

The other practical consideration is that the 2.3 i9 is not available without the 5500M and 1TB of RAM. So value seeking CPU users might be inclined to pick up the 2.4 i9 CPU upgrade on the base model and leave many of the other options unchecked. I'll likely do this for example, because the GPU is largely wasted on me and spec-ing out the base model to match a slightly upgraded top model suggests that 5300M -> 5500M upgrade is costing $100.

tl;dr: The CPU upgrades are almost certainly going to have some benefit, but it might be slightly less than what the frequency math implies. Decide for yourself if its worth it.

Unless you can monetise the extra percentage or just want bragging rights save the $$$, until the new 16" MBP is reviewed in depth no one really knows if there will be any tangible difference if at all. What is bad advise to advocating throwing money at Apple without any facts on the table. Generations of MBP have only seen a few percent increase when bumping up the CPU, no reason to think the 16" will be any different, until proved otherwise.

OP is asking i9 to i9 not i7 to i9...

Q-6

 

[wallysb01]

Unless you can monetise the extra percentage or just want bragging rights save the $$$, until the new 16" MBP is reviewed in depth no one really knows if there will be any tangible difference if at all. What is bad advise to advocating throwing money at Apple without any facts on the table. Generations of MBP have only seen a few percent increase when bumping up the CPU, no reason to think the 16" will be any different, until proved otherwise.

OP is asking i9 to i9 not i7 to i9...

Q-6

Time is money for everyone. How do you monetize yours?

Assuming your CPU is even just partially rate limiting in your workflow, and if it wasn't you probably wouldn't be asking this question, $200 is peanuts over the life of a machine. If it saves me 1-2 hours over say 4 years, its worth it.... Assuming say 5% of my work hours are CPU limited, that equates to needing just 0.25-0.5% increase in performance... We know that's the case. Waiting for reviews be damned. The mid-2019 machines with poor cooling and the same chips show us as much. If you are worried about being proven wrong, did you forget about returns? If the machine is poorly designed and the i9s don't differentiate themselves (which seems a highly unlikely occurrences given they did so in the mid-2019 machine), you can return it and get your cash back.

Current knowledge, risk, time, money. Value appropriately.

 

[Queen6]

Time is money for everyone. How do you monetize yours?

Assuming your CPU is even just partially rate limiting in your workflow, and if it wasn't you probably wouldn't be asking this question, $200 is peanuts over the life of a machine. If it saves me 1-2 hours over say 4 years, its worth it.... Assuming say 5% of my work hours are CPU limited, that equates to needing just 0.25-0.5% increase in performance... We know that's the case. Waiting for reviews be damned. The mid-2019 machines with poor cooling and the same chips show us as much. If you are worried about being proven wrong, did you forget about returns? If the machine is poorly designed and the i9s don't differentiate themselves (which seems a highly unlikely occurrences given they did so in the mid-2019 machine), you can return it and get your cash back.

Current knowledge, risk, time, money. Value appropriately.

As they say if you need to ask, you likely don't need. Personally my systems are switched out at the 2 year point as the next Gen is likely to be at least 20% faster if not more, and the cost minimal in the bigger picture.

Could care very less about being proved wrong, as don't have dog in the fight. Not my money on the table, nor my business OP can buy what he chooses. I can afford to wait and see as my current W10 Primary notebook is rock solid, can easily wait on the 10th Gen H series CPU. Early adopt by all means, equally don't cry if it doesn't meet the expectation, or is rapidly surpassed, which it will be.

If no urgent need waiting 2-3 months is simply common sense, as allows the manufacturer to workout the initial production kinks, less likelihood of a return. Unless you want to be the "first on the block" which in the real world impresses nobody.

2020 is where it's at for me, if I pick up a 2019 MBP will be for evaluation only to see if the OS can keep up. Ideally want a solid + 4GHZ across all cores on a notebook, currently hitting the wall at 3.9GHz across six cores...

Q-6

 

[wallysb01]

As they say if you need to ask, you likely don't need. Personally my systems are switched out at the 2 year point as the next Gen is likely to be at least 20% faster if not more, and the cost minimal in the bigger picture.

Could care very less about being proved wrong, as don't have dog in the fight. Not my money on the table, nor my business OP can buy what he chooses. I can afford to wait and see as my current W10 Primary notebook is rock solid, can easily wait on the 10th Gen H series CPU. Early adopt by all means, equally don't cry if it doesn't meet the expectation, or is rapidly surpassed, which it will be.

If no urgent need waiting 2-3 months is simply common sense, as allows the manufacturer to workout the initial production kinks, less likelihood of a return. Unless you want to be the "first on the block" which in the real world impresses nobody.

2020 is where it's at for me, if I pick up a 2019 MBP will be for evaluation only to see if the OS can keep up. Ideally want a solid + 4GHZ across all cores on a notebook, currently hitting the wall at 3.9GHz across six cores...

Q-6

Odd priorities you have. 10th gen H isn't going to "rapidly surpass" much of anything.... Most people don't buy a computer unless they need it. 2 year cycle a $3000 machine for 20% gains, but you worry about the ROI on $200 for 4% gains? List kind of goes on. Have a good one.

Edit and FYI: https://browser.geekbench.com/v5/cpu/search?q=MacBookPro16,1
And: https://www.notebookcheck.net/9880H-vs-9980HK_11348_11341.247596.0.html

Take it with a pretty big grain of salt, as we all know geekbench is a really short running benchmark, but 2.4 is doing ~15-20% better. Notebook check has them ~5% apart, as the frequency math would have us assume.

 

[Queen6]

Odd priorities you have. 10th gen H isn't going to "rapidly surpass" much of anything.... Most people don't buy a computer unless they need it. 2 year cycle a $3000 machine for 20% gains, but you worry about the ROI on $200 for 4% gains? List kind of goes on. Have a good one.

Your confusing your needs with others, nor really reading into it; 10th Gen over 8th Gen makes sense. Most companies rotate out their hardware based on the lease cycle, tax and or failure metrics.

New hardware for business is factor of profitability, new hardware for the base user is just often nice to have, and there's nothing wrong with that.

OP is asking will he benefit, unlikely in any tangible manner beyond saving a few seconds on an export. If the OP want's all good, equally it's not going be a significant gain, nor should it be portrayed as such, especially on an unproven platform such as the 16" MBP as it's not like Apple has had a firm handle on the thermals in the past...

Q-6

 

[iBrooker]

Doesn't seem worth it for 1 or 2 second bursts:

 

[Queen6]

Doesn't seem worth it for 1 or 2 second bursts:

It won't be, save your $$$ and put elsewhere. All the BS in the world won't override the laws of physics, the 16" MBP can either adequately power & cool the CPU at max frequency, if not it will roll back simple as that. The difference between the stock i9 and the upgraded i9 will be marginal at best in such a thin & light chassis, around 0.2GHz on a sustained load as the CPU will reduce frequency by default.

If you want the prestige go for it, equally don't be fooled this is a significant upgrade. For my needs I'd put the $$$ into 32GB as that would be more impactful to my workflow, than 0.2GHz for a matter of seconds.

Q-6

 

[wallysb01]

Doesn't seem worth it for 1 or 2 second bursts:

A lot of misconceptions in that video and his 'data' isn't showing what he thinks its showing. No mobile CPU is supposed to maintain the top turbo on all cores for long, nor does it have to in order for there to be performance differences between the models.

Your confusing your needs with others, nor really reading into it; 10th Gen over 8th Gen makes sense. Most companies rotate out their hardware based on the lease cycle, tax and or failure metrics.

New hardware for business is factor of profitability, new hardware for the base user is just often nice to have, and there's nothing wrong with that.

OP is asking will he benefit, unlikely in any tangible manner beyond saving a few seconds on an export. If the OP want's all good, equally it's not going be a significant gain, nor should it be portrayed as such, especially on an unproven platform such as the 16" MBP as it's not like Apple has had a firm handle on the thermals in the past...

Q-6

Hm, you keep going back to the OP, but you seem to be confounding the OP with yourself every bit as much as you claim I am. Logic, FCP were the uses. Not base user where new hardware is "just often nice to have", but presumably a profit driving machine.... Your statement of "a few seconds on an export" seems to imply you assume a light workload. Small videos or tracks, etc. Look, we can all live with a CPU that's 5% slower, question is, is 5% worth $200 to you. You're firmly giving the advice of "It won't be, save your $$$ and put elsewhere." Oh you tacked on an "Unless you can monetise the extra percentage.." exception, but wow, you've bent over backwards saying that's highly unlikely to be true. Saying waiting for reviews is one thing, but again, question is what's the difference? We can say how these chips are likely to preform, but only with the caveat that we haven't seen long-running or repeated benchmarks yet. Given what we know though, that doesn't seem like a big risk. And again, returns.... For the cooperate or business purchase that can be a pain, and I'm giving this a few days to iron out some details because of it, but for the average independent video editor that supposedly still has his old machine as you're testing the new one....meh....bfd dude....
[automerge]1573694989[/automerge]

It won't be, save your $$$ and put elsewhere. All the BS in the world won't override the laws of physics, the 16" MBP can either adequately power & cool the CPU at max frequency, if not it will roll back simple as that. The difference between the stock i9 and the upgraded i9 will be marginal at best in such a thin & light chassis, around 0.2GHz on a sustained load as the CPU will reduce frequency by default.

If you want the prestige go for it, equally don't be fooled this is a significant upgrade. For my needs I'd put the $$$ into 32GB as that would be more impactful to my workflow, than 0.2GHz for a matter of seconds.

Q-6

Unfortunately its not that simple. The silicon is tested and validated at higher clock rates for a reason.

 

[Queen6]

A lot of misconceptions in that video and his 'data' isn't showing what he thinks its showing. No mobile CPU is supposed to maintain the top turbo on all cores for long, nor does it have to in order for there to be performance differences between the models.



Hm, you keep going back to the OP, but you seem to be confounding the OP with yourself every bit as much as you claim I am. Logic, FCP were the uses. Not base user where new hardware is "just often nice to have", but presumably a profit driving machine.... Your statement of "a few seconds on an export" seems to imply you assume a light workload. Small videos or tracks, etc. Look, we can all live with a CPU that's 5% slower, question is, is 5% worth $200 to you. You're firmly giving the advice of "It won't be, save your $$$ and put elsewhere." Oh you tacked on an "Unless you can monetise the extra percentage.." exception, but wow, you've bent over backwards saying that's highly unlikely to be true. Saying waiting for reviews is one thing, but again, question is what's the difference? We can say how these chips are likely to preform, but only with the caveat that we haven't seen long-running or repeated benchmarks yet. Given what we know though, that doesn't seem like a big risk. And again, returns.... For the cooperate or business purchase that can be a pain, and I'm giving this a few days to iron out some details because of it, but for the average independent video editor that supposedly still has his old machine as you're testing the new one....meh....bfd dude....
[automerge]1573694989[/automerge]


Unfortunately its not that simple. The silicon is tested and validated at higher clock rates for a reason.

Simple average private user isn't going to see much out it, no matter how much of a meal you make of it, end of story...

Q-6

 

[wallysb01]

Simple average private user isn't going to see much out it, no matter how much of a meal you make of it, end of story...

Q-6

ha, ok man. $200 over 4 work-years is 2.5 cents an hour. It doesn’t have to be much of a meal...

 

[Strangedream]

Personally I would invest in more RAM before a CPU upgrade. A +0.1GHz increment doesn't seem worth it.

 

[wallysb01]

Personally I would invest in more RAM before a CPU upgrade. A +0.1GHz increment doesn't seem worth it.

For a casual user, the 16 -> 32GB upgrade is probably the most noticeable upgrade you can make. At least as time moves on and all the webpages, apps, etc., keep demanding more resources. At a certain point though, RAM upgrades won't change anything for your work. Where that level is, is user specific. If you are sitting around waiting for the CPU to run though, more processor upgrades are never wasted.