Intel Amber Lake

[LTC8K6]

Probably not much different than an 8500Y
https://en.wikichip.org/wiki/intel/core_i7/i7-8500y

 

[mikk]

Probably not much different than an 8500Y
https://en.wikichip.org/wiki/intel/core_i7/i7-8500y

1800 Mhz base vs 1500 Mhz base

"Not much" ....20% higher base.

 

[LTC8K6]

1800 Mhz base vs 1500 Mhz base

"Not much" ....20% higher base.

40% more power...?

 

[mikk]

40% more power...?

Not much different according to you.

 

[LTC8K6]

Not much different according to you.

Well, you said it was 7W and not 5W. I don't know where you got that number, though.
So that's technically your claim, not mine.

It's all preliminary info anyway.

I doubt there will be much difference between the two in reality.

 

[LTC8K6]

How often do mobile chips run at base clock under load?

I know desktop chips hardly ever run at base clock under load.

 

[mikk]

Well, you said it was 7W and not 5W. I don't know where you got that number, though.
So that's technically your claim, not mine.

I said 7W because 8510Y is a 7W SKU and this number comes from Intels graphics driver. What is your problem exactly?

 

[LTC8K6]

I only questioned the 40% more power number. I don't know where you went.

 

[mikk]

Driver inf 24.20.100.6222, it's accurate therefore and explains the high base clock.

 

[Eug]

Has anyone seen a review of an Amber Lake machine yet. I know there are a few models like the Dells out there, but it seems just about nobody actually has tested ones for a real review.

 

[Dayman1225]

New Amber Lake machine from HP, 8200Y or 8500Y, leather design which is pretty nice.

Few other notes; It has 1w 1080P panels as an option, perhaps the ones Intel showed at computex? and the inclusion of the Intel XMM 7560 Gigabit LTE modem but the antenna setup is only 2x2 MiMO.

 

[Dayman1225]

Bit more news on that HP Machine, it uses a customized Multi Chip Package from Intel that has the CPU, PCH and LTE Modem all on the same package

Intel's Customized SoC for HP: Amber Lake-Y with On-Package LTE Modem

Ian Cutress has asked Intel some additional questions about the product and is waiting on some answers now

 

[Eug]

Is this real? If so, why is the performance so underwhelming?

https://www.digitaltrends.com/computing/google-pixel-slate-intel-processor-leak/

 

[Dayman1225]

Is this real? If so, why is the performance so underwhelming?

https://www.digitaltrends.com/computing/google-pixel-slate-intel-processor-leak/

Perhaps its throttling due to it being in a tablet? or that its running Android.

 

[LTC8K6]

Is there a proper comparison to go by?

 

[Dayman1225]

Is there a proper comparison to go by?

Not really.

 

[Eug]

Is there a proper comparison to go by?

What do you mean by proper?

The MacBook Kaby Lake i7-7Y75 gets around 8700/4500 in OS X, but that's with TDP up with 1.4 GHz base clock.

There is this Windows test of an i7-7Y75 with a normal 1.3 GHz base clock that scores 7156/4086:

https://browser.geekbench.com/v4/cpu/6135564

So, I guess the Google one for the i7-8500Y, if using the normal TDP, would represent an improvement of maybe 12% multi-core. The score is worse single-core though for whatever reason.

 

[LTC8K6]

Well, it's Android 9. So I guess a proper comparison would be to another Android bench.

 

[Eug]

Here we go:

Celeron 3965Y Google Nocturne: 3476/2012 <-- Ouch!
m3-8100Y Google Nocturne: 7264/3885
i5-8200Y Google Nocturne: 8164/3909
i7-8500Y Google Nocturne: 8064/3764

https://browser.geekbench.com/v4/cpu/search?utf8=✓&q=nocturne

The best score is actually the i5-8200Y. It seems they ran the i7-8500Y benchmarks repeatedly and they always came up short compared to the single i5-8200Y score. Maybe throttling is an issue.

Peak speeds for the m3, i5, and i7 are 3.4 GHz, 3.9 GHz, and 4.2 GHz respectively.

 

[LTC8K6]

Yeah, the 8500Y and 8200Y are essentially the same chip when you put them in a tablet.
Cooling makes all the difference.

 

[IntelUser2000]

Perhaps its throttling due to it being in a tablet? or that its running Android.

Not really the same device. Geekbench page shows "Google nocturne", which is the codename for Google's next generation Pixel device, Pixel Slate.

Performance gain nonexistent in single thread is a big issue. The nocturne and the XPS 13 is a pre-release device though, so it still has a chance to get better. We'll see.

It has 1w 1080P panels as an option, perhaps the ones Intel showed at computex?

Yea it is. The regular 1080p display is 2W. Having an option means the 1W display is more expensive.

 

[ZGR]

How often do mobile chips run at base clock under load?

I know desktop chips hardly ever run at base clock under load.

Hardly ever as well. It is kind of annoying actually. My laptops love to maintain max turbo even if my CPU is hot. I usually end up turning off turbo or lowering turbo since I like to stay cool and quiet.

 

[jpiniero]

Clock speeds reported for one of the 8500Y (on the MT test?) was a median of 3053 and max of 3082. The max clock speed might be set to 3.1.

Meanwhile a Pixelbook with the 7y75 I saw had a max of 3558, which makes sense given the max turbo of 3.6.

 

[IntelUser2000]

Clock speeds reported for one of the 8500Y (on the MT test?) was a median of 3053 and max of 3082. The max clock speed might be set to 3.1.

Meanwhile a Pixelbook with the 7y75 I saw had a max of 3558, which makes sense given the max turbo of 3.6.

How did you get the clock speed number? Its not from Geekbench is it?

Sidenote: Intel never gave up on x86 Android development even after cancelling Broxton. Their efforts continued elsewhere. Brief search shows Android Oreo as the latest branch. Makes sense Google Nocturne is on Android 9.

Also, performance on Android should be pretty good for Geekbench. Its generally better than on Windows for Intel chips.

 

[jpiniero]

How did you get the clock speed number? Its not from Geekbench is it?

Put .gb4 at the end of the link url and you will get the raw data.