Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
Points of "relative performance" is not a thing! If it's a benchmark then say what it is. Quite simple. The stupid thing is that Apple do seem to have facts on their side, but they're bizarrely still using smoke and mirrors tactics to obfuscate and confuse.
I don't disagree that the general format of the chart is a good one, but the complete nonsense of what the Y-axis actually represents totally undercuts the point Apple are trying to make. If you're being genuine, then using disingenuous methods is not the best way to convince anyone of your truthfulness.
There’s another difference the article doesn’t point out: you can play hundreds of top tier games on the MSI.
I sort of get the point of doing a comparison like this, but there’s a reason Apple stopped at the chip: their comparison was to show how powerful their silicon is compared to Intel. The MSI was a useful example because gaming is typically a demanding activity.
Comparing the features of actual laptops however is less helpful in this context. MSI is a gaming laptop. Gaming is literally why it exists. Saying the MacBook is more “useful” ignores the fact that, despite the power of its chip, it can’t come close to competing with the MSI on its turf.
Apple has key hardware (ASI) and software (Metal) pieces in place to substantially improve the state of Mac gaming.
Now they just need to spend some $ to get top tier games, including some exclusive (or semi-exclusive) content. Maybe a partnership with Sony to get PS exclusives on the Mac and/or buy/build a in-house game studio.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
Points of "relative performance" is not a thing! If it's a benchmark then say what it is. Quite simple. The stupid thing is that Apple do seem to have facts on their side, but they're bizarrely still using smoke and mirrors tactics to obfuscate and confuse.
I don't disagree that the general format of the chart is a good one, but the complete nonsense of what the Y-axis actually represents totally undercuts the point Apple are trying to make. If you're being genuine, then using disingenuous methods is not the best way to convince anyone of your truthfulness.
Annoying, but think of all the determined nerd interest it creates in finding out just what those actual benchmarks are, and to be the very first to spray those benchmarks about everywhere! it’s genius marketing. That is what it is.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
Points of "relative performance" is not a thing! If it's a benchmark then say what it is. Quite simple. The stupid thing is that Apple do seem to have facts on their side, but they're bizarrely still using smoke and mirrors tactics to obfuscate and confuse.
I don't disagree that the general format of the chart is a good one, but the complete nonsense of what the Y-axis actually represents totally undercuts the point Apple are trying to make. If you're being genuine, then using disingenuous methods is not the best way to convince anyone of your truthfulness.
Annoying, but think of all the determined nerd interest it creates in finding out just what those actual benchmarks are, and to be the very first to spray those benchmarks about everywhere! it’s genius marketing. That is what it is.
Not a bad idea, maybe that's Apple's angle, but I still find it cringeworthy.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
"My M1Max is capable of 170 relative performances" doesn't really roll off the tongue, nor does it actually mean anything unless you know to what it's relative. If all you're doing is qualifying top line performance against power consumption, then the processor may not be all that capable.
If you're going to aggregate performance metrics, you might want to say how and what you're aggregating.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
"My M1Max is capable of 170 relative performances" doesn't really roll off the tongue, nor does it actually mean anything unless you know to what it's relative. If all you're doing is qualifying top line performance against power consumption, then the processor may not be all that capable.
If you're going to aggregate performance metrics, you might want to say how and what you're aggregating.
Every time I see one of those graphs I cringe. Can't imagine why Apple think they're ok.
Totally disagree. Those plots are awesome.
In what way are they awesome? The Y axis is a mystery, and completely unverifiable.
The y-axis is just relative performance like you see in bar charts. Apple says it is based on industry benchmarks, but they surely use the most advantageous ones, just like everyone else does.
These xy charts are often better at displaying multi-variate information. Apple surely wants to show that they have both an absolute perf advantage and a power efficiency advantage.
This xy perf vs power plot is the way to do it. You want to be in the upper left part of the plot. They explained this during WWDC20 when they announced the transition.
You can pretty much count on Intel or MS never presenting it in this way. Intel will say they win the absolute performance crown, but won’t mention how much power it uses. Or MS says that their Surface device lasts 20 hours, but in the small print say it includes idle times.
"My M1Max is capable of 170 relative performances" doesn't really roll off the tongue, nor does it actually mean anything unless you know to what it's relative. If all you're doing is qualifying top line performance against power consumption, then the processor may not be all that capable.
If you're going to aggregate performance metrics, you might want to say how and what you're aggregating.
They do plot a comparative machine in all the plots. They are named in the small print, but the machines are named. Once you see that, the relative performance to that machine can be judged. It's all relative distance or displacement along both axes, between the two machines. If they use bar charts, they then will have to rely on people to read and listen about the parameter that was not plotted, which is usually always power consumption. This typically means nobody will think about power consumption because the information wasn't conveyed.
With the XY plots, the information is all there. The essential message is there. The CPU in the Jade die (SoC codename) will have about 20% more performance than a late model 8core Intel i9 at about half the Watts, so on and so forth. What they can always be faulted for are not outright naming the competitor hardware instead of leaving it in the small print, and not naming the industry benchmarks they used. The plot style? I definitely think it is better than the bar charts, at least for this type of messaging.
It really outlines how far into the exponential part of the power curve that competitor chips are at. They are only gaining 5%, 10% for rather huge amounts of power, while Apple is getting about the same or better for much less power. Definitely a message Apple wants to get across. Thankfully, Anandtech is finally moving to perf vs power XY plots too. The bar chart rendition of it was basically unreadable. Sixcolor's autoscaling bar charts could really use some work too. No better than just reading off a table.
It seems some of you want to know what specific tasks or processes, like C4D or compiling or transcoding. That's not an issue because of the plot. They can plot GB5 multi-core or Metal on the y-axis and you should be happy right? I'd be too. That's not an issue with the plot style. Everyone has everyone seen hundreds of bar charts with some unknown relative performance as the metric. So plot abuse will occur in all plot styles.
Heck, most benchmarks themselves are relative performance. GB5 is a relative performance benchmark where the number is scaled to some Intel processor whose score is baselined as 1000 or whatever it is. The score itself is meaningless without knowing what other systems or chips score on the benchmark. That's relative performance. You are judging performance based on the difference in GB5 numbers, not the numbers themselves.
If games are what you want on your laptop, you will be getting a Windows gaming laptop as there effectively are no games for the Mac (Apple having systematically and intentionally having killed them off by ejecting NVIDIA and not supporting industry standards like Vulkan or 32 bit backward compatibility).
Try a ray tracing benchmark and you would see that RTX 3070 destroy the M1 Max integrated GPU. This may become a moot point when the M2 Max is tested late next year but NVIDIA will be releasing their next GPUs at the same time which are rumored to have far better ray tracing performance than the 30 series.
Serious question: the graphs on graphics performance have MSI more than twice as fast. Isn't this one of the core elements of gaming performance (I don't play games but thought it was all about having a good graphics card)? I would think everything else is secondary to this and MSI seems to deliver substantially better figures (so long as it is plugged into an outlet, obviously).
Serious question: the graphs on graphics performance have MSI more than twice as fast. Isn't this one of the core elements of gaming performance (I don't play games but thought it was all about having a good graphics card)? I would think everything else is secondary to this and MSI seems to deliver substantially better figures (so long as it is plugged into an outlet, obviously).
As has been said so many times already. If someone is buying a computer to place games, you recommend a PC, be it desktop or laptop. Apple has only fallen further behind on this with the ARM transition. It's really a games availability issue, not a hardware issue. If Apple puts some real effort to having a games ecosystem for macOS, a discussion between PC and Mac gaming machines could be had, elsewise, get a PC until there is good catalog of Mac games. Maybe cloud gaming will level the playing field, but not for another year maybe.
On the other hand, the are a lot of GPU compute users, likely more than there are gamers. And more of them will be willing to spend $5000+ on a laptop, with many of them using their own custom code that could be written specifically to Apple's TBDR GPU design assumptions and get equivalent performance. These MBP machines are quite competitive with a lot of great intangibles. So, they can easily compare this MSI laptop and a MBP and come away thinking that the MBP is a much much better option for their GPU compute needs.
There’s another difference the article doesn’t point out: you can play hundreds of top tier games on the MSI.
Well be that as it may, I don't think the intent of the article was to compare the number of Wintel games to Mac games. It seem to be comparing hardware performance and features. On that front, it's not a particularly in-depth hardware comparison.
FWIW, Apple has never marketed the MacBook Pro as a gamer laptop, but I think comparing game performance (where it can be done) is certainly fair. I would like to have seen more of that in this article, even if Apple made no claims about game performance.
Comments
I don't disagree that the general format of the chart is a good one, but the complete nonsense of what the Y-axis actually represents totally undercuts the point Apple are trying to make. If you're being genuine, then using disingenuous methods is not the best way to convince anyone of your truthfulness.
it’s genius marketing. That is what it is.
With the XY plots, the information is all there. The essential message is there. The CPU in the Jade die (SoC codename) will have about 20% more performance than a late model 8core Intel i9 at about half the Watts, so on and so forth. What they can always be faulted for are not outright naming the competitor hardware instead of leaving it in the small print, and not naming the industry benchmarks they used. The plot style? I definitely think it is better than the bar charts, at least for this type of messaging.
It really outlines how far into the exponential part of the power curve that competitor chips are at. They are only gaining 5%, 10% for rather huge amounts of power, while Apple is getting about the same or better for much less power. Definitely a message Apple wants to get across. Thankfully, Anandtech is finally moving to perf vs power XY plots too. The bar chart rendition of it was basically unreadable. Sixcolor's autoscaling bar charts could really use some work too. No better than just reading off a table.
It seems some of you want to know what specific tasks or processes, like C4D or compiling or transcoding. That's not an issue because of the plot. They can plot GB5 multi-core or Metal on the y-axis and you should be happy right? I'd be too. That's not an issue with the plot style. Everyone has everyone seen hundreds of bar charts with some unknown relative performance as the metric. So plot abuse will occur in all plot styles.
Heck, most benchmarks themselves are relative performance. GB5 is a relative performance benchmark where the number is scaled to some Intel processor whose score is baselined as 1000 or whatever it is. The score itself is meaningless without knowing what other systems or chips score on the benchmark. That's relative performance. You are judging performance based on the difference in GB5 numbers, not the numbers themselves.
This may become a moot point when the M2 Max is tested late next year but NVIDIA will be releasing their next GPUs at the same time which are rumored to have far better ray tracing performance than the 30 series.
On the other hand, the are a lot of GPU compute users, likely more than there are gamers. And more of them will be willing to spend $5000+ on a laptop, with many of them using their own custom code that could be written specifically to Apple's TBDR GPU design assumptions and get equivalent performance. These MBP machines are quite competitive with a lot of great intangibles. So, they can easily compare this MSI laptop and a MBP and come away thinking that the MBP is a much much better option for their GPU compute needs.
Well be that as it may, I don't think the intent of the article was to compare the number of Wintel games to Mac games. It seem to be comparing hardware performance and features. On that front, it's not a particularly in-depth hardware comparison.
FWIW, Apple has never marketed the MacBook Pro as a gamer laptop, but I think comparing game performance (where it can be done) is certainly fair. I would like to have seen more of that in this article, even if Apple made no claims about game performance.
14" MacBook Pro vs RTX Surface Laptop Studio: RIP Intel?
https://www.youtube.com/watch?v=E4Dov6TJMgg