Not quite sure, are you referring to my GB5 scores? In which case, no not quite that good: Zen 5 has higher clocks than Zen 3 but even higher Integer scores than expected given those clocks. Which is still pretty good. In the article, since they are comparing to Zen 4 which had slightly higher clocks than the current Zen 5 mobile chips then indeed they may have gotten better scores for lower clocks but I'm not sure which tests are "integer". I haven't looked at the article you linked to closely myself.
M3 core counts and performance
- Thread starter Cmaier
- Start date
So there are multiple interesting parts to this and how the AMD, Intel, Apple, and Qualcomm GPUs measure up.
Data primarily from here: https://www.notebookcheck.net/AMD-Z...le-M3-and-Qualcomm-Adreno-X1-85.868475.0.html
But I'll pulling in other NotebookCheck articles over the course of the writeup.
Summary: Apple machines would be incredible at gaming if they had the library - okay maybe not in terms of perf/$ but still way better than I had thought. The Qualcomm GPUs are underpowered but also quite efficient (again gets to @NotEntirelyConfused 's point that their SOCs would've made even more impact in a fan-less design had they focused on that). The AMD iGPU is not as good as I was thinking it would be and Intel, again if their drivers were up to par, might actually be decent gaming iGPUs.
Since this is primarily an M3 thread, let's start with Apple! Unfortunately in the article itself Apple M4/M3s are only in a small subset of the benchmarks despite that NotebookCheck actually has the data for most of the rest. For the graphics benchmarks, Apple does very well. In Steel Nomad Light we can see the M4 leads the pack of iGPUs and the M4/M3s do really well in GFXBench and 3DMark. What's is waaaay more interesting is how Apple actually performs on games. Now here, this NotebookCheck article doesn't give very many benchmarks, but we can look at some of the articles on the M3/M3 Pro/M3 Max. Now unfortunately there were two M3 Max reviews, the 14-inch showed much better Witcher 3 performance than the 16" and was done a few weeks later, so I'm assuming there were driver/bug fixes. Otherwise they showed nearly identical gaming performance. Another caveat is that a lot of the gaming benchmarks are at 1080p or 1440p rather than 4K - in other words the CPU, depending on the graphics settings and GPU size, has a greater effect on frame rates here. But the main point I'd like to make the main game I'm going to highlight, The Witcher 3, is an x86 game running under WINE. In other words, Apple's performance and efficiency here are expected to be worse than for native games.
For the Witcher 3 with settings 1920x1080 Ultra Graphics & Postprocessing (HBAO+) the M3, M3 Pro, M3 Max got 28, 43, 92 FPS respectively getting 1.05-1.1 FPS/Watt. Which is incredible, again The Witcher 3 is about as non-native and non-optimized as you can get. The M2 Pro does even better here than the M3 Pro - 51 FPS and similar efficiency. The M2 Pro has 1 more GPU core and different P/E core mix and looking at some of the base M2/M3 score it's also possible that not every game/workload responded well to the new cache structure. Discrete GPUs can still outpace the M3 Max which scores at best around the 4060/4070 mobile, so Apple has room to improve in terms of raw performance per dollar. But their GPUs are very, very good for gaming. I would've liked to see have seen comparisons to some Nvidia MaxQ designs which I think are the closest dGPUs to Apple's iGPUs.
By contrast, the new AMD Strix Point iGPU which I thought would be similar to the M3 Pro was only getting 33 FPS and was 0.46-0.67 FPS/Watt. In other words, it gets a worse FPS for more power! Its graphics benchmarks are similarly unimpressive. What is more impressive is its OpenCL compute scores which is probably a function of its dual-issue FP design (which @leman and I couldn't really figure out exactly how it was supposed to work). But it doesn't seem to help too much on graphics benchmarks or in actual games. I'm worried, because AMD GPUs being good at compute but not great at graphics helped defined the pre-RDNA GPUs and is part of how we got to Nvidia's near total dominance today. Hopefully RDNA 4 turns things around. Strix Halo will also be very interesting to see but right now, it looks as though Apple's iGPUs are just unassailable ... or would be if they had a competitive gaming library!
The Qualcomm GPUs underperform almost across the board except in GFXBench and, to be fair, in perf/watt. They have very efficient GPUs but they just don't have the raw power - again except in GFXBench. I think @Jimmyjames and I talked about why Qualcomm does so well in Aztec Ruins, but didn't come to any conclusions. Obviously it is an old benchmark these days, but it's definitely an odd duck relative to all the other results.
Intel GPUs oddly seem to need a lot of power to perform well and unfortunately performance in graphics benchmarks isn't well reflected in games, indicating driver-performance issues outside the hardware itself. Newer hardware and better driver may improve both. Intel is making big claims for its next generation of Lunar Lake processors.
Again, I just can't emphasize enough how much better the Apple GPUs performed here than I was expecting, especially on titles being translated from x86-Windows! It really showcases that Apple is very close to making a really great gaming laptop. We'll see how the M4 generation shapes up, but maybe it'll by the M5 generation the M5 Pro will become a great midrange gaming machine (obviously the competition isn't going to stand still, but the potential is there).
Data primarily from here: https://www.notebookcheck.net/AMD-Z...le-M3-and-Qualcomm-Adreno-X1-85.868475.0.html
But I'll pulling in other NotebookCheck articles over the course of the writeup.
Summary: Apple machines would be incredible at gaming if they had the library - okay maybe not in terms of perf/$ but still way better than I had thought. The Qualcomm GPUs are underpowered but also quite efficient (again gets to @NotEntirelyConfused 's point that their SOCs would've made even more impact in a fan-less design had they focused on that). The AMD iGPU is not as good as I was thinking it would be and Intel, again if their drivers were up to par, might actually be decent gaming iGPUs.
Since this is primarily an M3 thread, let's start with Apple! Unfortunately in the article itself Apple M4/M3s are only in a small subset of the benchmarks despite that NotebookCheck actually has the data for most of the rest. For the graphics benchmarks, Apple does very well. In Steel Nomad Light we can see the M4 leads the pack of iGPUs and the M4/M3s do really well in GFXBench and 3DMark. What's is waaaay more interesting is how Apple actually performs on games. Now here, this NotebookCheck article doesn't give very many benchmarks, but we can look at some of the articles on the M3/M3 Pro/M3 Max. Now unfortunately there were two M3 Max reviews, the 14-inch showed much better Witcher 3 performance than the 16" and was done a few weeks later, so I'm assuming there were driver/bug fixes. Otherwise they showed nearly identical gaming performance. Another caveat is that a lot of the gaming benchmarks are at 1080p or 1440p rather than 4K - in other words the CPU, depending on the graphics settings and GPU size, has a greater effect on frame rates here. But the main point I'd like to make the main game I'm going to highlight, The Witcher 3, is an x86 game running under WINE. In other words, Apple's performance and efficiency here are expected to be worse than for native games.
For the Witcher 3 with settings 1920x1080 Ultra Graphics & Postprocessing (HBAO+) the M3, M3 Pro, M3 Max got 28, 43, 92 FPS respectively getting 1.05-1.1 FPS/Watt. Which is incredible, again The Witcher 3 is about as non-native and non-optimized as you can get. The M2 Pro does even better here than the M3 Pro - 51 FPS and similar efficiency. The M2 Pro has 1 more GPU core and different P/E core mix and looking at some of the base M2/M3 score it's also possible that not every game/workload responded well to the new cache structure. Discrete GPUs can still outpace the M3 Max which scores at best around the 4060/4070 mobile, so Apple has room to improve in terms of raw performance per dollar. But their GPUs are very, very good for gaming. I would've liked to see have seen comparisons to some Nvidia MaxQ designs which I think are the closest dGPUs to Apple's iGPUs.
By contrast, the new AMD Strix Point iGPU which I thought would be similar to the M3 Pro was only getting 33 FPS and was 0.46-0.67 FPS/Watt. In other words, it gets a worse FPS for more power! Its graphics benchmarks are similarly unimpressive. What is more impressive is its OpenCL compute scores which is probably a function of its dual-issue FP design (which @leman and I couldn't really figure out exactly how it was supposed to work). But it doesn't seem to help too much on graphics benchmarks or in actual games. I'm worried, because AMD GPUs being good at compute but not great at graphics helped defined the pre-RDNA GPUs and is part of how we got to Nvidia's near total dominance today. Hopefully RDNA 4 turns things around. Strix Halo will also be very interesting to see but right now, it looks as though Apple's iGPUs are just unassailable ... or would be if they had a competitive gaming library!
The Qualcomm GPUs underperform almost across the board except in GFXBench and, to be fair, in perf/watt. They have very efficient GPUs but they just don't have the raw power - again except in GFXBench. I think @Jimmyjames and I talked about why Qualcomm does so well in Aztec Ruins, but didn't come to any conclusions. Obviously it is an old benchmark these days, but it's definitely an odd duck relative to all the other results.
Intel GPUs oddly seem to need a lot of power to perform well and unfortunately performance in graphics benchmarks isn't well reflected in games, indicating driver-performance issues outside the hardware itself. Newer hardware and better driver may improve both. Intel is making big claims for its next generation of Lunar Lake processors.
Again, I just can't emphasize enough how much better the Apple GPUs performed here than I was expecting, especially on titles being translated from x86-Windows! It really showcases that Apple is very close to making a really great gaming laptop. We'll see how the M4 generation shapes up, but maybe it'll by the M5 generation the M5 Pro will become a great midrange gaming machine (obviously the competition isn't going to stand still, but the potential is there).
Last edited:
NotEntirelyConfused
Power User
- Joined
- May 15, 2024
- Posts
- 71
No, I was talking about comparisons between Z5 and Z4 in the NBC article I linked. But both that and the Z5-Z3 comparisons suggest IPC gains substantially better than 0, unless I'm missing something major.
Still pressed for time here, but, great writeup again on the GPU stuff!
@NotEntirelyConfused (and everyone else of course!) Here's thumbnail of the findings for the new AMD chip, I just added the data to the previous graph (apologies it's a big chart and compression wasn't kind to the fuzziness of the text):
View attachment 30580
In addition to the NotebookCheck article, I'm also going to talk about the Anandtech article.
There are a few things to note here:
1) The new Zen 5 mobile P-cores on N4P represent a significant jump in ST performance compared to Zen 4. They are only 10% more efficient at the clock speed they are rated for (which is 200 MHz less than the max clocks of the Zen 4 mobile chips) however that's at better performance, implying that they could lower clockspeed even more and get more efficiency though they probably still wouldn't match the lower end Qualcomm cores never mind Apple cores. So single core efficiency is still way behind though at least they improved performance.
2) These Strix Point chips are manufactured on a slightly better node (N4P) than the Apple (N5P) or Qualcomm (N4) chips. Judging from this chart, the difference isn't huge: N4 looks like a slightly more dense N5P but with almost identical power/performance and N4P has similar density to N4 but either 6% better performance at the same power or 11% better efficiency at the same performance for the chip TSMC bases there reference calculations on. While that certainly helps the AMD chip here, I don't expect that to predominate especially as these are load-idle power figures for the whole device.
3) The HX 370 is a much bigger chip with 12 cores and 24 threads, 4 more cores than the 8845HS chip it replaces. True, 8 of the cores are now "c" cores but while they are a smaller and more efficient than the their non-c counterparts for multithreaded workloads I suspect the difference is minimal in this respect. In an interview with Tom's Hardware, Mike Clark states that the Zen 5c cores are about 25% smaller than the Zen 5 cores with most of what was taken out and rearranged having to do with silicon that allows the cores to boost to super high frequencies. Thus for the purposes of multithreaded workloads, especially endurance tests like CB, one might be tempted to say that they are effectively P-cores. However, they also have access to dramatically less cache, especially last level cache, than the standard P-cores and have high core-to-core latency especially with the standard P-cores as they have to communicate through the last level cache with the standard P-cores. This will almost certainly limit their performance relative to the standard P-cores and, in some ways, this is similar to what I suspect is limiting the Qualcomm chips in multithreaded workloads (i.e. cache and bandwidth).
4) These extra cores means a significant percentage of the uplift in multithreaded performance simply comes from having more cores available. For instance, jumping down to the 3D Mark CPU Profile Simulation HX 370 only gets a 5% jump over the previous gen chip (the 8845HS and 7840HS I believe have the same CPU) when both are restricted to 8 threads, but doubles its lead in the same benchmark at max threads. As shown in the graph above, for CB R24, its lead over its predecessor is more substantial. At 57 watts, it now gets the roughly same score (1022) as the Apple M2 Pro (1030, 60W) and Qualcomm X Elite 78 (1033, 62.1W) whereas the 8845HS was about 15-16% behind (842, 56.8W). The three newest chips all now have similar efficiencies at this point in their performance/watt curves. As stated in my previous post, it is unfortunate that we don't have X Elite 80s or 84s to compare here at this performance cure which are higher binned chips and might perform better. I think I saw a Techlinked video which stated that these are apparently more difficult to source for review sites. The Apple M3 Pro by contrast score around 1055. Unfortunately NotebookCheck didn't do wall power measurements for it under CB R24 (I suspect they don't have one anymore), but it's in the mid-40Ws for wall power. That would make it somewhere in the range of 25-30% more efficient at ISO-performance. The M3 Pro is on a slightly better node (N3B) but while transistor scaling increased dramatically for practical chips with increasing SRAM again difference in power/performance due to that alone aren't expected to predominate (a few percent compared to N4P). Overall, I suspect the M2 Pro, Qualcomm X Elite, and especially M3 Pro are a good deal more die efficient than the HX 370.
5) The caveats from my previous post are expected to hold here too.
6) Anandtech compared the HX 370 to the base M3. AMD amusingly asked them to compare to the M3 in the Air, which Anandtech demurred and tested the M3 in the MacBook Pro. As I think you can see from the above chart, for multithreaded applications even comparing to the base M3 at all regardless of device, especially without very well controlled power data, is also not a great idea. This is a chip which can be pushed to 120W wall power (although at such little performance gain at that point I can't think why anyone would - a measly 4% increase in performance costs a whopping 44% more power at this point in the curve!). Anandtech measured the power draw of the HX 370's 28W setting to be around 33W. I suspect this is platform power from HW Info but Anandtech didn't specify in the article. Notebookcheck measured a different device but same chip from Asus and found that it drew 48W wall power at the 28W setting and both got the similar CB R24 scores of 927-950 (Notebookcheck-Anandtech) so I suspect we're talking similar power draws. From what I can tell this is much more similar to the M3 Pro in powermetrics and wall power for a CPU-only test, which again scores around 1050 in CB R24 making ~10-14% more performant at ISO power. Meanwhile, the base M3's power metrics/wall power don't come even close to these figures under CPU load. Notebookcheck measured it in the Air and found it only drew 21W at the wall in the fanless design and got a score of about 600pts, Anandtech measured the M3 in the MacBook Pro to get 718, at presumably higher average power draw in an endurance test than the thermally limited Air, but almost certainly not anywhere near the same wattage as the HX 370 in its 28W TDP configuration.
7) The single threaded SPEC results in the Anandtech article are interesting and here you can see the AMD chip losing in INT performance to the M3 P-core but slightly beating it at FP performance. Without power measurements, not even software ones, we can't get a sense of efficiency but we can readily assume that the the M3 is far superior in that respect. Still, shows that indeed in ST, AMD is doing very well. Might be time to update ISO-clock graphs (I keep saying that).
8) Geekbench 6 results tells a similar story for the new AMD chips. Here's an example (2780/15267 ST/MT) for the same ASUS model as tested by NotebookCheck, unclear at what TDP. Geekbench single core for the M2 Pro is about (2663/14568), the M3 Pro is about (3179/18982), and the Qualcomm Elite 80 is (2845/14458). Here the 12 cores + SMT probably isn't helping AMD as much especially relative to the M3 Pro given that GB 6 switched some of its MT workloads to task based parallelism to cut down on the "MoAr cores is better" phenomenon.
Conclusion: The new AMD chips are naturally still far behind the Qualcomm Elite and Apple M2 P-core in ST efficiency, but performance has been much improved. This remains the weakest element to AMD's chips and it will likely take awhile for x86 processors to match it if they ever do. For MT, while AMD had to add more cores, and one suspects use much more silicon, to compete with the M2 Pro and Qualcomm Elite, the end user doesn't likely care how they did it, only that they now match them, even slightly beat them. This is particularly bad for Qualcomm since they need to be much better than AMD as compatibility issues still abound and the rest of the SOC is underwhelming. If Qualcomm's die is cheaper though, then they can compete on price and that advantage shouldn't be discounted! Also, I've seen mixed reports about the iGPU in the AMD chip (haven't had time to look at that in depth, will try to report if there is anything interesting), but the Qualcomm's was I think worse and of course the AMD Strix Point can still be paired with a dGPU (though that sorta defeats the purpose of the APU? I mean it's a sizable iGPU). For Apple, AMD may be matching the M2 Pro, but the current M3 generation remains comfortably ahead and of course the base M4 (3715/14690 GB 6 in the fanless iPad) is expected to come to the Mac at the end of the year and Pro/Max chips as well. Apple remains the ST and MT mobile performance-efficiency kings for now.
But M2 Pro power draw is only about 34W and M3 Pro 26-27W. So it is much more effecient then.
I can measure power draw of M1 Max and it is in CB24 multi-thread 27-28W according to powermetrics and 30-32W wall power. (But it will be power source and RAM that is adding to power draw from wall.)
Powermetrics or wall power? NBC has M2 Pro drawing 60W of wall power with CB R24. That’s what I’m using in my charts. I don’t have either a meter or an M2 Pro to verify. Everything in the NBC article is wall power.
Very interesting … What was your M1 Max score in CB R24? Unfortunately the NBC M1 Max review is too old to use CB R24 to compare what their wall power measure/score is but it’d still be good to have.
Sorry, for M2P and M3P it was powermetrics for CPU according to https://www.notebookcheck.net/Apple...w-Efficiency-before-performance.772025.0.html But I do not believe that 60W is true. Maybe with display with high brigness and with ineffeciency of the charger. (But I do not know how they measured Qualcomm and AMD. So it can be comparable. But the chargers could also be different effeciency wise. Also motherboards,… Hard for me to agree with such figures.)
I have found one mistake in my previous measurement of power draw. (Heavier background app running I forgot about, that gave me higher idle draw and lower estimated CPU maximum power draw.)
This is re-tested with macOS 14.6, MBP16” M1 Max 10/32c/32GB, CB 2024.1.0
My CB24 score was 585 pts.
This is how I measured - I used USB-C power meter. I was measuring wall power between USB-C cable and Macbook's usb - c port. Auto-brightness off display for 2/16 brightness. (Some background apps running but average power draw should minimize their influence.) I was measuring exact power consumption in exact time. Power draw idle was 4.65W. (Average in 14 minutes.)
Power draw with running Cinebench R24 was 38.16W average during 12 minutes.
I would say that in idle CPU is consuming only dozens of mWh. During load there is also load for motherboard, wires, RAM, also higher amount of losses,… So safely power draw of CPU is less than 33.51W (38.16-4.65). When we subtract power for MB, RAM,… I would say we should be near powermetrics figures. (For that reason I would believe more in using SW power draw for CPU on Windows and Mac.)
According to the test setup they attach laptops to an external display so the internal display doesn't affect things and they also subtract idle. Most of that should come out. That said, I do agree that I would prefer powermetrics on top of wall power for power efficiency tests. They do seem to open with powermetrics in the article you linked but in the actual M2 Max/Pro they state that they are discontinuing powermetrics and just using wall power with external monitors from now on:
Apple M2 Pro and M2 Max analysis - GPU is more efficient, the CPU not always
Apple have introduced their new SoCs, the M2 Pro and M2 Max. Following the slight lack of efficiency in the M2 compared to the M1, now we will shine the spotlight onto both of the new chips in the MacBook Pro models: the M2 Pro and the M2 Max.
www.notebookcheck.net
I think as long as we are consistent it is fine ... however, I'll admit that the NBC figures are substantially higher than yours for power draw. They have a very fancy meter but that shouldn't be necessary for this case. Do you have a cutdown M1 Max or full M1 Max? I'm guessing a cutdown one? I see scores of the full M1 Max getting 792 in CB R24.
EDIT: ah I see that the wall power consumption CB R23 was measured at 46W for the full M1 Max by NBC - M2 Pro was 54W and the M2 Max was 64. So there seems to have been a big jump in the M2 Pro/Max generation for CPU load power draw, 16-18W higher!
Full M1 Max. I can measure it again, but I got about 600 points 3 times.According to the test setup they attach laptops to an external display so the internal display doesn't affect things and they also subtract idle. Most of that should come out. That said, I do agree that I would prefer powermetrics on top of wall power for power efficiency tests. They do seem to open with powermetrics in the article you linked but in the actual M2 Max/Pro they state that they are discontinuing powermetrics and just using wall power with external monitors from now on:
Apple M2 Pro and M2 Max analysis - GPU is more efficient, the CPU not always
Apple have introduced their new SoCs, the M2 Pro and M2 Max. Following the slight lack of efficiency in the M2 compared to the M1, now we will shine the spotlight onto both of the new chips in the MacBook Pro models: the M2 Pro and the M2 Max.
I think as long as we are consistent it is fine ... however, I'll admit that the NBC figures are substantially higher than yours for power draw. They have a very fancy meter but that shouldn't be necessary for this case. Do you have a cutdown M1 Max or full M1 Max? I'm guessing a cutdown one? I see scores of the full M1 Max getting 792 in CB R24.
EDIT: ah I see that the wall power consumption CB R23 was measured at 46W for the full M1 Max by NBC - M2 Pro was 54W and the M2 Max was 64. So there seems to have been a big jump in the M2 Pro/Max generation for CPU load power draw, 16-18W higher!
Weird discrepancy then, 14” model? Lower thermal capacity, lower power, lower scores?
No, 16” as stated above. Re-running CB24 after restart and result is 587. I will look into it again tomorrow. (It is even weirder that in Geekbench 6 my results are comparable with other M1 Maxes in 16”MBP)
Difference in power measurement against NBC can be in CB23 vs CB24. I think, it was reported that CB24 does not cause as high power draw as CB23.
I did not realize that NBC is notebookcheck.net Could have looked into it more before writing my post. Sorry. Looks like you are right. (Still would prefer sw figures only for cpu but it should be comparable either way.)According to the test setup they attach laptops to an external display so the internal display doesn't affect things and they also subtract idle. Most of that should come out. That said, I do agree that I would prefer powermetrics on top of wall power for power efficiency tests. They do seem to open with powermetrics in the article you linked but in the actual M2 Max/Pro they state that they are discontinuing powermetrics and just using wall power with external monitors from now on:
Apple M2 Pro and M2 Max analysis - GPU is more efficient, the CPU not always
Apple have introduced their new SoCs, the M2 Pro and M2 Max. Following the slight lack of efficiency in the M2 compared to the M1, now we will shine the spotlight onto both of the new chips in the MacBook Pro models: the M2 Pro and the M2 Max.
I think as long as we are consistent it is fine ... however, I'll admit that the NBC figures are substantially higher than yours for power draw. They have a very fancy meter but that shouldn't be necessary for this case. Do you have a cutdown M1 Max or full M1 Max? I'm guessing a cutdown one? I see scores of the full M1 Max getting 792 in CB R24.
EDIT: ah I see that the wall power consumption CB R23 was measured at 46W for the full M1 Max by NBC - M2 Pro was 54W and the M2 Max was 64. So there seems to have been a big jump in the M2 Pro/Max generation for CPU load power draw, 16-18W higher!
Dang it so you did, I was reaching for a solution!
In single core I think it’s the other way around. Low single core power draw in CB R23 was one of the pointers that CBR23 had a flaw. CB R24 seems to have fixed that flaw and in multicore I would expect R24 to draw higher or at least as much power as R23. But I could be wrong!
This is a mystery! I too would like both software and wall - they sanity check each other. Maybe tomorrow when you get a chance maybe try running R23 instead of R24? This is very odd.
Hmmm ... I'm starting to think the NotebookCheck is NOT subtracting out idle power in their efficiency analyses ... which is not what I thought they were doing. They report idle but it isn't clear that they do anything with them when calculating efficiencies. They do try to normalize everything as far as I can tell - the efficiency graphs they use seem to be with external modules turned off (wifi/bluetooth) and an external monitor (and hopefully in the case of testing iGPU graphics, the dGPU is turned off), but still there is going to be variation here. CC: @Nycturne @Souko
Thus while many of these laptops have similar idle power, some may not and that makes it much trickier to interpret the graphs when trying to isolate the chip from the rest of the system. For some analyses it may make little difference like the ST CB R24 difference between ARM and x86, but for others ... ugh ... not happy. I may have to redo all the analyses. Not sure how much will change.
Still double checking.
Thus while many of these laptops have similar idle power, some may not and that makes it much trickier to interpret the graphs when trying to isolate the chip from the rest of the system. For some analyses it may make little difference like the ST CB R24 difference between ARM and x86, but for others ... ugh ... not happy. I may have to redo all the analyses. Not sure how much will change.
Still double checking.
Last edited:
So I found out that when I lock the screen. MBP has higher power draw and higher score. With load power consumption about 48W (-4.7W idle = 43.3W) I got 670 pts in CB24. (Still low score for M1 Max full.) Like OK, maybe macOS does not utilize CPU fully when there is chance that I will work on something else but still… Did not happen with earlier macOS 14.X releases. And I still do not understand. Can anyone run CB24 on any MX CPU and post the result here?
I’ll try mine - so lock the screen vs unlocked screen? I don’t have a power meter. Could it be entering “high power mode”?
For me high power mode does not matter. Measuring score with locked and unlocked screen is enough to see if there is any diference. Thank you.
No difference here. Cut down M3 Max 14" OS 14.6, lid open/closed, screen unlocked/locked - 1357-1372 with the best score being screen unlocked. Earlier I got 1384 screen unlocked. Dropandrender say 1436 for my machine, CPU Monkey though reports the same as mine, 1373. The difference is minimal, <5%. I believe that, unlike for the M1 Max, "High Power mode" actually does make a slight bit of difference for the M3 Max if memory serves, but I'm not sure how much, especially for a cutdown one like mine. EDIT: High power mode: 1443. so yeah Dropandrender is probably reporting under high powered mode but again <5% difference.
For your system, CPU Monkey agrees with DropandRender that you should be getting close to 800 in CB R24.
Last edited:
Sarah was great and I miss her. This bizarre AMD stock holding troll — among other “leakers” that steadily buy into a much more competent AMD (or Intel) than reality would have it have really degraded that forum. Probably terminally.
Even the mods are bad about this. Defending the leakers and holding out for a great Zen 5 Apple-caliber (or even Qualcomm by ipc) core etc.
Graphs comparing Qualcomm, AMD, and Apple updated with subtracting idle* power from load power:
The new AMD HX 370 chip does look substantially better here. Compared to Apple's M2 Pro** and Qualcomm in the ST it's so far behind still that it doesn't matter too much (still 1.5x - 2.5x behind in efficiency), but it is now more performant AND more efficient than the previous generation AMD chip. And, to be fair, the gap to Apple/Qualcomm ST is closed a little. In multithreaded though it looks much better as things were roughly equal before and now AMD has a more efficient multithreaded chip than either Qualcomm***/Apple M2 Pro. So as a relative change, that matters more. Again though, I would love to see die size/cost estimates for the new AMD chip vs Qualcomm/Apple. I removed the final data point from the graph (the ultra high power AMD setting) to improve readability but the point stands that this chip is really designed with different parameters in mind than say a base Mx chip.
*some idle powers had to be estimated as I couldn't find exact figures but I'm confident that the estimations are close enough for these purposes. The M2 Pro was in a 14" model and had a very small idle power of just over 2 watts, while the highest was the new AMD Asus Pro 16" device which used just over 10 watts idle (estimated based on load power and the Asus Pro 13" device idle).
**It should be noted that Apple's ST efficiency gets better down the product stack - as these are almost package power estimates, base Mx is more efficient than Mx Pro which is more efficient than Mx Max, etc ... as the there is simply less powered on under ST load - smaller, more efficient SOC with less RAM and memory bandwidth and cache, etc ... which doesn't impact ST performance much and lets the smaller SOCs use less power. This is one of the advantages of having multiple SOCs targeted to different platforms.
***It also has to be said that this Qualcomm Elite 78 looks even worse compared to the 80 and 64 under load. The Asus Qualcomm with the 78 chip had a lower idle power than the Microsoft devices with the 64 and 80 which just makes the load power look even worse! It's possible this is badly binned chip or Asus has terrible power delivery under load or something else. Regardless Qualcomm would probably be doing much better efficiency and performance wise if we had curves for the 64/80.
With her account disabled I don’t suppose there’s a way to invite her to these forums?
Of course immediately after I publish this and NotebookCheck releases the P16 review, but it turns out my estimate of idle power was indeed close enough - 10.2W instead of my 10.8W. Thus I'm only overestimating AMD's efficiency in the above graph by < 4% at worst and most of the time less than that.
The comparisons to the full M3 Max are interesting. Overall the M3 Max is much more expensive than the P16 (for some reason in the specs it says 4,000 Euros as the cost of the P16 but in the text and I confirmed online it is ~$2700 for the 64GB of RAM variant while that's $4200 for the M3 Max). However, the M3 Max has a much faster CPU and the GPU is as fast or faster (exceptions when doing small ray tracing workloads and non-native gaming) and, in general, will use less power, create less noise, and get less hot than the P16. But there's no doubt that you pay for all of that. For AMD, the current HX APUs replace the old U and HS chips, its possible the upcoming Zen 5 laptop chips that replace the old HX chips (paired with dGPUs) will be a better price/performance comparison to the M3 Max - though at that point the GPU performance will probably pull away from the Mac at the cost of power/heat/noise (even the P16's mobile 4070's performance is on the lower end of the mobile 4070s). As it always is, it depends on what you value.
Last edited:
Similar threads
