Intel Ice Lake benchmarked - i7 performs better with lower TDP than Ryzen 7 3750H-APU
Click here to post a comment for Intel Ice Lake benchmarked - i7 performs better with lower TDP than Ryzen 7 3750H-APU on our message forum
Andy Watson
As Intel is currently constrained on 10nm it will be interesting to see how these are priced compared to AMD.
Lots of Intel news recently so I guess the new AMD processor will not be a dud.
sverek
Wait, are those benchmark has any hint on GPU performance? Cause I don't want APU that performance like i7, yet can't handle minecraft on lowest settings.
schmidtbag
A bit misleading... We all know that Intel's TDP only comes from base clocks. That CPU could boost to 4.5GHz, absolutely obliterate AMD in single-threaded burst calculations, and it'd still be called a 15W part.
Since that chart only shows single-threaded, that makes it very clear that the boost clocks are doing all the heavy lifting.
Keep in mind, I don't necessarily have a problem with boost clocks, but my point is these parts can't really be compared since the i7 isn't really a 15W part. Granted, I call BS that the Ryzen is a 15W part too, especially if the GPU is also under heavy load.
I guess what I'm saying is comparing laptop performance is utterly useless. The only thing that matters is who has better performance-per-watt (from the wall, not claimed by the manufacturer...) and performance-per-, while being able to keep up with your workload.
Fox2232
^ That's fair. Intel makes "honest mistakes" way too often with SDP vs. TDP.
moo100times
Agreeing with all the above re: Intel, AMD and appropriate scepticism. I would also like to weight in re: price. If AMD's offering is a sub 600 GBP laptop offering APU, then also looking at very different performance brackets. I have recently got both AMD Ryzen 5 laptop and Ryzen 3xxx laptop and both were sub 600GBP and significantly better value offerings than Intel equivalents even on same brands (Lenovo, ASUS). Not saying Intel doesn't have a winner, but I will reserve judgement until it is released...as well as wonder how they will change their cpu to cover for all existing security flaws whilst maintaining performance increase...
TieSKey
And don't forget most laptop CPUs/GPUs are throttled for total system power limits and ofc, thermals. A good bulk and noisy one can get double the scores of an ultra thin on the same hardware and even turning off the screen during the benchmark can get u some points.
As u already said, comparing laptop performance is useless. (unless u have the exact same design, memory, total system power draw and limit)
JamesSneed
We have to assume boost clocks were used for these results and if that is the case AMD isn't all that far behind considering the 3500U is still on the old 12nm process. AMD's 7nm Zen2 replacement of the 3500U should compete quite well and possibly even be better if you count the 15% IPC increase plus a little more boost clock. Intel is in damage control mode as of late.
JamesSneed
No its not. They are both 4c/8t.
schmidtbag
Not like it makes a difference since the test was single-threaded anyway.
TLD LARS
I would be very surprised if this benchmark is true, until now Intel 10nm laptop CPU´es does not even perform well compared to Intel 14nm laptop CPU´es.
The huge difference between base and boost suggests that this CPU is made for very short bursts and not sustained speeds, the bursts could be close to 45W TDP no problem even with a 15W rated TDP.
MaCk0y
This was expected, no? The Ryzen is a 12nm Zen+ 1Q19 released chip. It would be worrying for Intel if they didn't beat AMD with a 10nm yet to be released mobile CPU.
Aura89
Not sure what the relevance of this information is.
It's testing Ice Lake against Zen+ mobile, not Zen 2, so it's not testing 7nm power consumption, which is what everyone is interested in right now.
Not entirely sure anyone technically expects Zen+ to best Intel in mobile markets for power, though i do believe this is wrong either way as Intels TDPs and power restrictions are simply wrong and have been tested time and time again. i3-8130u 15 watt processor can use up to around 25 watts if i recall correctly, just as an example.
3700u is a 15w processor but it is also configurable by the manufacturer up to 35 watts, so if it's using 30 watts, your laptop manufacturer must have set that power limit instead of 15.
Gets a little confusing with laptops since each manufacturer can have the same CPU in their laptop but a very different overall performance if they mess with the configurable wattage.
Srsbsns
"I'll believe it when I see independent benchmarks"
Megastitious
Yeah ok and as soon as AMD gives them a single moment to breath, they won't hesitate to screw customers back with the same old "barely improved" CPU's of theirs, cause of no competition.That's how treacherous Intel is.They don't care about their customers.They care about their competition statuses and how to milk money with barely the same product every year.If they wanted to deliver great mainstream CPU's like what AMD does, they would have done that long ago.
Mundosold
Numbers are useless - laptop performance is about thermal/power management. Many laptops do not have adequate cooling for even 15W and will thermal throttle after a few seconds, and instead are designed for short term 'bursts' - like rendering a web page or opening a program. My Dell Inspiron with an 8250U will hit 97C within 5 seconds of load, even with undervolt and repaste.
To get useful numbers, you would have to run a sustained benchmark (10 minutes or so), and ignore the first 3 minutes or so of performance. You would also need to have each laptop be almost identical other than the CPU, with identical cooling.
This is why people undervolt laptops - which reduces temperatures and power usage (which helps stay under TDP limits). Enough undervolting and you can significantly increased sustainable performance, the risk being system instability if you are too low.
That said I am sure Intel is still ahead with laptops because they are putting next-gen up against Zen+. I'd be really interested in a 7nm AMD APU especially if it had 7nm Vega too..
H83
This^ Laptops are to different for a direct comparison. The cooling solution alone plays a big role on the performance of laptop CPUs/GPUs. This is just another pointless marketing stunt...
reb0rn
The leaked info about icelake is know from long time ago... some 15% IPC is expected its only shame ice lake is 2+ years old tech that never get fabs to produce right due 10n issue
The damage is even worst when we knew they sell us old tech because they had no fail plan to move ice lake to 14nm...
The next full redesign sapphire rapid should be even a lot larger IPC gain, if they are smart they will try to speed it up as fast they can
coth
Can't even imaging how is that running. I have rather simple notebook Lenovo 330s series with 2-core Ryzen 2200U. Well it's good for business apps like browsers, office etc, got NVMe and HDD for storage, but, 2200U hits 80 C in load. It's hot. And that are just 2 cores. Can't even imagine how hot are 4-core mobile Zens.
JamesSneed
I'm just envisioning someone at Intel running this test with the chip on a test bench with a giant desktop class cooler sitting on this laptop chip so it can run at boost non stop and pull down 25+watts.
user1
Sounds like a laptop design problem, than the chip itself, they tend to skimp on the heatsinks these days, trying to use as little copper and alluminum as possible. Thermal paste is also atrocious on most of the really cheap laptops.
