I was searching for a laptop to buy and came across some interesting stuff that I don’t understand:
How does a mobile laptop cpu works?
Because you see, I’m no expert on the matter, but no complete noob also.
The clock speed is about cycles per second, the higher the clock speed the more cycles per second. I also understand that clock isn’t an universal benchmark, it depends on a lot of stuff (even deeper stuff than the quantity of cores, like the set of instructions and the architecture itself).
But I stumbled on the laptop i3 and i5 and was like: wait a second… what?
Intel says (on its ark page) that the clock for the i3 7020u is 2.3 GHz, ok, and the i5 8250u runs on 1.6 GHz with a turbo boost of 3.4 GHz (that seems high for a laptop, sounds nice). And then it hit me…
How the hell does a clock speed work?
Does the i3 stays clocked on 2.3 GHz even in idle? Does that mean it will be there going nuts over my standing still desktop environment? Does this means it will drain more battery?
What clock speed actually means? Does the cpu stay always ready warming up or does it stay put and wait for its time?
And by the way I don’t need lots of processing power, will just be doing some coding for a college course and studying. Even after graduating I won’t be doing much graphically intensive stuff, mostly web development, so an i5 will probably be overkill, that’s why it’s the main example here.
To be honest, neither of those values will give you any decent comparison on those CPUs. Those can be helpfull when comparing a i7 7600 to a i7 7700k, but other than that, Cores, Clock and such say basically nothing (at least not in detail).
For the “easy” solution you can go to the Passmark website and compare up to 3 CPUs: Like that
This isn’t a assesment of actual real world performance, but it can tell you a bit about how those CPUs might perform. But this also depends heavily on your workload.
Both are 15W TDP. Anything else, again would be workload dependend. I’m assuming Brightness of the Screen would be more of a factor. As well as battery size. The difference in Powerdraw of hose CPU’s should be negligeble.
Well that’s the trick isn’t it (?); you’re going to have to compromise somewhere. Personally I always choose performance in a workstation - not that I don’t see the potential value in being able to use in places where power options are scarce - it’s just not something that affects me (if I want to spend time in the woods, for instance, it’s not because I want to sit in front of a computer).
+1 for not worrying about I-whatever too much/just use cpubenchmark/passmark for power (shows typical range from idle to max) and performance info/comparison… while there are general takeaways from denotations (like the u-series you mentioned generally using less power), but things change generation to generation and I suspect it suits Intel to confuse the average consumer.
I haven’t seen any specs yet for the new AMD 12nm processors, but they should do more on less, so to speak; so if your’e not in a hurry to make a purchase you might consider holding off a couple months to see if any will meet your requirements.
Edit: just checked and I’m totally wrong here/looks like 12nm has been out. I had seen somewhere that AMD next gen mobile/laptop would be 12nm alongside 7nm desktop - weird/is that true/anyone know why that would be the case/seems counter-intuitive??
Edit 2: double-wrong, here’s a comparison:
Since battery life is an important factor for laptop use, CPU TDP is a much better indicator of anticipated battery consumption, than is CPU clock speed. As @domsch mentions, with a 15W CPU, the display is likely drawing far more battery power, that the CPU. Ever notice how the battery life on your cell phone evaporates, just by reading text? The display sucks a lot of power! Incidentally, a conventional hard disk will also draw far more battery power, than a SSD.
not really the TDP is a measure of “power” used Thermal Design something or another.
I have no idea of the exact CPU’s in question but given Intels track record id say there is ~10% difference in the IPC(instructions pr clock, which is optimization).
A good rule of thumb is that the higher TDP the higher clock just compare a regular desktop CPU to a mobile CPU TDP, and clock speeds.
I think what is really confusing you is Intels segmentation of their products, which is just a method of extracting more money from the users.
Again im just pulling this out of my Ar… but id wager the i3 is 2 core hyperthreaded, so 4 threads ~2300/2Mhz , and the i5 is 4 core non hyper threaded, so 4 threads @~1600Mhz
Both Intel and AMD chips are able to slow down to a couple hundred MHZ when idle. Additionally, even if they are running at full speed, it does not consume as much power/produce at much heat if the chips are idle even if they are running at full speed.
This tech is marketed as SpeedStep for Intel, and Cool’N’Quiet for AMD. It should be on all Intel and AMD chips made in the past 10-15 years, new chips still have it even if it is not heavily marketed anymore.
You should be able to turn it off in the UEFI, unless you buy a super locked down brand such as HP, although there is no point to turning it off unless you are doing serious overclocking.
This is separate from throttling where the CPUs slow down under heavy load because they are getting too hot, or are drawing too much power.
The opposite, sort of, is Turbo Boost. If the CPU has some temperature and power headroom, it can clock higher then normal, and a lot higher then normal but only if one-two cores are being used.
TDP matters a great deal when it comes to sustained performance. A low TDP chip may clock up high for a very brief period, but can’t stay there. If you just do office stuff that may be fine, but not if you’re running longer tasks requiring sustained performance.
The way modern CPUs work is they scale their clockspeed based on demand, power utilization, and temperature. So if you don’t need it, your CPU will clock way down to a couple hundred megahertz. Then as your demand increases it will clock all the way up to its boost clock, until it either hits its top temperature or power utilization, then it will throttle back.
So basically, don’t worry about base clock. Mobile chips often have a very low base clock, with a high boost. Look at the boost clock, TDP, and number of cores.
Regarding your question regarding the i3 7020u vs i5 8250u: first of all the i3 processors don’t have “turbo boost” so the ability of the processor to modulate the clocks is more limited compared to i5 and up processors. That means that Intel can use more cores from a single wafer because they just need to be “good enough”. The base clock is the frequency Intel is guaranteeing 100% of the time if you’re below the maximum temperature available. Since both CPUs have the same TDP to keep the i5 inside the TDP with double the cores and threads Intel just lowers the clocks. In real workloads the base clock doesen’t matter that much since the cooling systems are efficient enough to keep temps lower than tj max so that i5 will run always run above base clock and above 2.3GHz for sure.
Regarding the power consumption: the i3 processor won’t keep the clocks at 2.3GHz all the time, it will lower the clocks when the load doesen’t need much power. What Intel is saying is that 2.3GHz is all you’ll get out of that processor. While with the i5 you’ll get 3.4GHz for single core short boost and below that as you increase the number of cores/thread involved in the operations.
For your use I’d surely suggest the i5 8250u. 4 cores and 8 threads can do much more than just 2 cores and 4 threads. Also, if the laptop you buy is configured with a lower TDP, you can install the Intel XTU and increase the TDP to get more performance out of it (more TDP = more clockspeed for longer periods). I’m also of the opinion that buying something a bit more “overkill” in the beginning allows to not having to upgrade soon.
