Intel sent me an early laptop sample and it's……………

[Deli]

9 minutes ago, danwat1234 said:

Why no mention of single core performance differences between tiger Lake and AMD? Not allowed to show IPC difference?

Dave Lee also posts his preview, with Cinebench R20 single core performance. In which the Tigerlake scores 576. Yes 576, which is even higher than 10900k at 5.3GHz can manage. I find it hard to believe. But is is what it is. We will see when Tiger Lake lands in real laptops, not just engineering sample.

[Results45]

Daaaaang. Now that's some screaming segue benchmark scores:

 

 

But can Linus segue a segway segue into a segue segway of another segue?

[Diazonium]

Yeah, I am calling bullshit on the "100 °C is fine" argument.

1. If your server is hitting 100 °C, you need to call the server room to have them kill power to all machines because the AC broke down. Tj(max) on Cascade Lake X (last gen X299) is 86 °C according to Intel spec, and based on the Tcase figure regular 28 core Xeons should have a Tj(max) of about 90 °C(+/- 2°C). Not 100 °C.
2. If the CPU is constantly hitting 95-105°C, what sort of temperatures do you think the VRM is hitting? Even if the CPU really can take the heat for 5 years, can every component on the motherboard take it as well? Most notably, capacitors near the CPU and the VRM FETs. With typical laptop warranties being at most 2 or 3 years, do you really have the trust in the manufacturers to use more expensive capacitors to tolerate high temperatures? Or are they just going to go with caps rated for 3000h at 85°C? We see desktop motherboards being proud to have excellent capacitors (ie. 10 000h @ 105°C), when was the last time a laptop manufacturer bragged about the quality of the capacitors used?
3. Recent Macbooks are in fact dying a whole lot, often with shorts developing in VRM capacitors and BGA/FCBGA failures killing CPUs.

 

[hishnash]

1 hour ago, Diazonium said:

Recent Macbooks are in fact dying a whole lot

No if even 0.1% of Macbooks were failing it would be massive number of devices. I expect the failures you are seeing are just regular failure rate failures. remember the number of macBooks apple selese massively out-numbers any other single laptop line form any other brand due to the fact that apple just have very few different product lines.
 

 

1 hour ago, Diazonium said:

If the CPU is constantly hitting 95-105°C, what sort of temperatures do you think the VRM is hitting?

If your concerned about VRM temperatures (and other board temperatures, and your are right to be conserned about these) look at the heat sensors from these components (most devices will have them). A VRM can also hit 100C while your cpu is at 30C. The cooling of your cpu and your VRM are very different and the max internal temp of your CPU is not at all a good proxy to the temp of your VRM or even the temp of the socket. Uset the temp sensors on these rather than the max internal temp of the cpu. 
 

[danwat1234]

100C is fine!!! I have an old Asus g50VT laptop core 2 duo X9100. It red lines at 104C. for almost its entire life it has lived near this level running distributed computing in the background at a slight overclock of 3.3GHZ and slight over-volt. Hasn't failed yet. So unless 10 nanometer is real delicate with regards to temperatures it's no big deal at all.

[Egg-Roll]

Still a better Tiger Lake than what is in the video:

https://goo.gl/maps/xUFgoMB3ANwzy2jk7

[BlueChinchillaEatingDorito]

7 hours ago, eeeee1 said:

what if i dont have time at the moment? you could just tell me. idc about spoilers 

Your first comment was at 9:53 AM PST and the one I quoted you was posted at 10:05 AM. Yet you had time to have this that back and forth conversation. Try harder next time. 

[hishnash]

1 hour ago, danwat1234 said:

100C is fine!!! I have an old Asus g50VT laptop core 2 duo X9100. It red lines at 104C. for almost its entire life it has lived near this level running distributed computing in the background at a slight overclock of 3.3GHZ and slight over-volt. Hasn't failed yet. So unless 10 nanometer is real delicate with regards to temperatures it's no big deal at all.

One of the things a lot of people miss is the max CPU temp is never going to be an issue. what you should consider however is the VRM and socket temp. 

Modern CPUs will not overheat and damage themselves so the max temp at the center of your cpu die is not realy important at all as a metric. Since you cant even use it as a good proxy for your VRM or socket temps as the cooling of the VRM might be very different to that of the cpu you can have a cpu at only 50C while your VRM might be overloaded at 105C. 

[danwat1234]

16 minutes ago, hishnash said:

One of the things a lot of people miss is the max CPU temp is never going to be an issue. what you should consider however is the VRM and socket temp. 

Modern CPUs will not overheat and damage themselves so the max temp at the center of your cpu die is not realy important at all as a metric. Since you cant even use it as a good proxy for your VRM or socket temps as the cooling of the VRM might be very different to that of the cpu you can have a cpu at only 50C while your VRM might be overloaded at 105C. 

Agreed I'm not sure how hot the VRMs in that laptop run but it's the original motherboard.

[TetraSky]

So the only reason why these Intel mobile CPUs will even sell, is because AMD has supply issues. Ok.

[BlueChinchillaEatingDorito]

29 minutes ago, TetraSky said:

So the only reason why these Intel mobile CPUs will even sell, is because AMD has supply issues. Ok.

And not the trust issues mainstream buyers have had with AMD laptops in the past? I don't think I've seen a single Ryzen powered laptop at my university even though they've been on the market for how many years? 

 

AMD: APUs are going to revolutionize everything. 

 

Buyers when APU laptops drop:

 

[saltycaramel]

Takeaways

- raw power is not as important as the whole package, look at the bigger picture of the build quality of the laptop

- 100C is fine

- supply chain is king

 

good to know as an apple user

 

also, is it customary for intel to send these super early samples to influencers (dave lee, linus, etc.) or they’re in scramble mode just this year because they’re about to look really bad in the inevitable apple silicon vs intel slide when in mid october the new 12” Macbook is unveiled at the iPhone event?

 

[TheJesus]

I don't know if anybody at LMG will or can answer this, but what RAM speed was the Intel laptop running in the gaming benches? Their specs support LPDDR4-4266 and with an iGPU, that's going to increase FPS compared to a Ryzen laptop running say 3200.

[hishnash]

3 hours ago, TheJesus said:

I don't know if anybody at LMG will or can answer this, but what RAM speed was the Intel laptop running in the gaming benches? Their specs support LPDDR4-4266 and with an iGPU, that's going to increase FPS compared to a Ryzen laptop running say 3200.

Yep supporting faster memory is a key feature of these SoC. one of the sad bits is they will not support LPDDR5 so they will be painfully slow compared to the Apple SoCs that will for sure being using LPDDR5.

[Results45]

Honestly I think Hardware Canuck's video was more streamlined and to the point.

 

 

This LMG video seemed to jump around from the design to benchmarks to criticizing Intel marketing to availability to positives against Ryzen availability then back to hardware design.

 

Great video overall content, needs a smidge more work on the delivery & editing.