Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

anandtech.com

Post reply on HN
Print view View original

61–70 of 79 posts

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#61

Post by kllrnohj »

agloeregrets wrote: post #40
Macha wrote: post #24

Earlier quoted context omitted.

Currently (and for laptop CPUs specifically) 1. AMD Ryzen 4900H 2. Apple M1, though the Intel 10875H is pretty much the same perf (though at 3x the power usage) 3. Only Intel uses this, so whatever Intel's fastest chip on this process is. That said, it varies by use cases, and M1 leans heavily on accelerators so has big wins in some and big losses in others when compared the 4900H or 10875H, which are closer to each…

> 2. Apple M1, though the Intel 10875H is pretty much the same perf (though at 3x the power usage) In single core? At least by Geekbench (clearly not the best test) it is shown to be more than 25% slower than M1.

Assuming Intel is sticking to Specint2017 for that claim, the M1's lead isn't that big: https://www.anandtech.com/show/16252/mac-mini-apple-m1-teste...

So per that the M1 is 8% faster in single-core than the 28W i7-1185G7 (side note: Intel fix your naming yeesh). It would not at all be surprising if Intel claws that back, that's well within "typical" generational IPC gains.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#62

Post by reader_mode »

AmVess wrote: post #54
reader_mode wrote: post #52

Earlier quoted context omitted.

You can't under-volt on latest machines (mine included). I need to limit TDP to get the machine usable. I've thought about opening it up and reapplying paste but I don't want to risk bricking my work device - I don't have the time to deal with it. I will have a few weeks off soon so might consider it if the expected gains are significant but from what I saw back when I looked into it (this[1] mostly) it's not worth i…

It may drop the temp a few C or delay throttling, but small heatsinks and tiny fans can only dump so much heat out of the chassis. I did all kinds of things to my XPS, but the largest gain came from undervolting and it still sounded like a leafblower when it was being worked moderately hard. My M1 Macbook on the other hand, is dead quiet on top of absolutely destroying the XPS in CPU performance while having vastly s…

Yeah no doubt M1s are amazing, I'm kind of sad Apple is the only one with it because Apple isn't big on providing options. My ideal form factor is a Yoga or Surface Pro style - and iOS is not a real option.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#63

Post by kllrnohj »

ithrow wrote: post #58

The X1 Titanium Yoga will start at $1,899 If someone is not tied to Windows, it's hard to justify this over a $999 MBA.

Bold claim to make when you don't know the starting specs nor upgrade prices. Remember the $999 MBA is only 8GB RAM & 256GB SSD. It's not exactly a good configuration, and Apple's upgrade pricing means it quickly becomes a $1400-$1600 MBA when you give it modern specs on those two aspects.

The X1 Titanium Yoga then offers a touch screen, a better aspect ratio, convertible usage, and significantly faster multi-core performance, at least in bursts. That can easily be worth another $400 depending on the use case & personal preferences. It's also fully possible the X1 Titanium Yoga will be overpriced, of course.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#64

Post by ksec »

giantrobot wrote: post #59
ksec wrote: post #30

Earlier quoted context omitted.

1. Hamstring might not be fair to "Intel Compiler" , as it was designed for Intel in the first place. Unfair Comparison would be a more valid description. i.e The problem is not with "Intel's compiler", the problem is with the benchmarks and marketing. 2. Keep in mind fastest Single Threaded benchmark is still somewhat true considering Intel can turbo up to 5Ghz. 3. TDP comparison between AMD and Intel are meaningles…

The Intel compiler has been doing function multi-versioning for many years now. If you annotate a function for multi-versioning the compiler will build several optimized versions of it, including a generic version, running the most optimal one for the host CPU based on a CPUID check. Intel purposefully misidentifies AMD chips with the CPUID check so always runs the generic version of a function on AMD. It does this e…

>Intel is free to optimize the shit out of their compiler but it's bullshit that they also actively sabotage AMD chips at the same time. It's very much a vendor lock-in tactic.

I would argue they actively make sure their compiler optimisation works only on Intel CPU. Which is actually a fair deal. Those optimisation work isn't free.

It is BS tactics? Yes. Do I Like it ? No. It is wrong to do it? That is up to debate.

The point is using an Intel Compiler on Non Intel CPU comparison makes the benchmarks pretty much useless. Something like GCC / LLVM should have been used to be vendor neutral.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#65

Post by bserge »

reader_mode wrote: post #52
bserge wrote: post #50

Earlier quoted context omitted.

It's the heatsink. And Apple isn't the only offender. All major manufacturers use terrible cooling systems, even on thick workstation class laptops. Their testing is seriously flawed, they design the heatsink for "normal workload" which means jack shit. Everyone's got different uses for their machines, so just test the heatsink at maximum load/TDP for several hours at least, make sure it doesn't overheat. Is that too…

You can't under-volt on latest machines (mine included). I need to limit TDP to get the machine usable. I've thought about opening it up and reapplying paste but I don't want to risk bricking my work device - I don't have the time to deal with it. I will have a few weeks off soon so might consider it if the expected gains are significant but from what I saw back when I looked into it (this[1] mostly) it's not worth i…

That's unfortunate. Is it because of Plundervolt? Intel's answer to that was basically "lol, just disable manual voltage control", even older machines got BIOS updates disabling it, which made a tiny number of people angry.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#66

Post by jeffbee »

reader_mode wrote: post #22
Jonnax wrote: post #16

From this article: https://wccftech.com/intel-launches-tiger-lake-h-tgl-h-mobil... Intel say they have the fastest single threaded performance of any laptop processor. They show a graph of a benchmark of something called SPECRATE * 2017_int_base compiled with the Intel compiler. Does the Intel compiler still hamstring AMD? Also is this benchmark useful in real world usage?

Does it really matter in a 45W power range ? I have a 2018 15 MBP with an i9-8950HK that has the same power range - this thing is unusable at full load for anything but burst - and the way it taxes the thermals makes it a terrible laptop (people in the office start turning their heads). I'm not convinced any portable laptop can handle that TDP efficiently - so maybe this is interesting for gaming laptops - but to me…

The 45W rating is hardly relevant. Excluding Apple, Intel 11th Gen Core CPUs are the fastest on most single-threaded benchmarks, whether they are configured at 15W, 27W, or 45W. In fact there is barely any difference in peak performance between the 15W and the 27W config.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#67

Post by giantrobot »

ksec wrote: post #64
giantrobot wrote: post #59

Earlier quoted context omitted.

The Intel compiler has been doing function multi-versioning for many years now. If you annotate a function for multi-versioning the compiler will build several optimized versions of it, including a generic version, running the most optimal one for the host CPU based on a CPUID check. Intel purposefully misidentifies AMD chips with the CPUID check so always runs the generic version of a function on AMD. It does this e…

>Intel is free to optimize the shit out of their compiler but it's bullshit that they also actively sabotage AMD chips at the same time. It's very much a vendor lock-in tactic. I would argue they actively make sure their compiler optimisation works only on Intel CPU. Which is actually a fair deal. Those optimisation work isn't free. It is BS tactics? Yes. Do I Like it ? No. It is wrong to do it? That is up to debate.…

I don't find fault in Intel using insider knowledge to make their compilers work better on their own CPUs. If they make optimizations they know work on their CPUs but can't guarantee AMD CPUs will work exactly the same, I also don't find fault in them only supporting their CPUs with those optimizations.

Where I find fault is bullshit marketing saying their CPUs are x% faster than AMD and then point to code compiled with their compiler. I'd be fine if they wanted to say ICC produced faster results than GCC or LLVM. There's no problem selling your CPU with your compiler to get the best performance possible.

Instead they're saying our chips are faster than AMD. Full stop. Once you wade into the bullshit you see they're using ICC with FMV enabled. This means even if you know what compiler flags they used you still can't be sure what code is running on each CPU. It's bullshit marketing and I think it needs to be called out whenever it happens.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#68

Post by tutfbhuf »

mwcampbell wrote: post #33

45W hardly strikes me as "mobile". That's in the realm of laptops with fans and big batteries, right? I figured a "mobile" processor would need a maximum TDP of 15W or less.

I find that 25W CPUs are quite a good middle ground. They usually offer better perf than 15W and run much cooler than 45W.

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#69

Post by reader_mode »

jeffbee wrote: post #66
reader_mode wrote: post #22

Earlier quoted context omitted.

Does it really matter in a 45W power range ? I have a 2018 15 MBP with an i9-8950HK that has the same power range - this thing is unusable at full load for anything but burst - and the way it taxes the thermals makes it a terrible laptop (people in the office start turning their heads). I'm not convinced any portable laptop can handle that TDP efficiently - so maybe this is interesting for gaming laptops - but to me…

The 45W rating is hardly relevant. Excluding Apple, Intel 11th Gen Core CPUs are the fastest on most single-threaded benchmarks, whether they are configured at 15W, 27W, or 45W. In fact there is barely any difference in peak performance between the 15W and the 27W config.

Compared to last gen AMD cores ?

Re: Intel’s 8-Core Mobile Tiger Lake-H, at 45 W, to Ship in Q1

#70

Post by jeffbee »

reader_mode wrote: post #69
jeffbee wrote: post #66

Earlier quoted context omitted.

The 45W rating is hardly relevant. Excluding Apple, Intel 11th Gen Core CPUs are the fastest on most single-threaded benchmarks, whether they are configured at 15W, 27W, or 45W. In fact there is barely any difference in peak performance between the 15W and the 27W config.

Compared to last gen AMD cores ?

11th Gen Core are 10-20% faster than Zen 2 “Renoir” on the workloads I’ve tried.
Post reply on HN