The Ice Lake Benchmark Preview: Inside Intel's 10nm
by Dr. Ian Cutress on August 1, 2019 9:00 AM EST- Posted in
- CPUs
- Intel
- GPUs
- 10nm
- Core
- Ice Lake
- Cannon Lake
- Sunny Cove
- 10th Gen Core
Power Results (15W and 25W)
Based on the SKU table, Intel was very keen to point out that all of the Y-series processors for Ice Lake and all the 15W U-series processors have cTDP up modes. This means that OEMs, if they build for it, can take advantage of a higher base power of a processor which leads to longer turbo periods and a higher frequency during sustained performance levels.
While cTDP is a good idea, one of the issues we have with the concept is that Intel’s OEM partners that design the laptops and notebooks for these processors don’t ever advertise or publicise if they’re using a CPU in cTDP up or down mode. I could understand why a vendor might not want to advertise using a down mode, but an up mode means extra performance, and it’s hard to tell from the outside what is going on.
For what it is worth, most users cannot change between these modes anyway. They are baked into the firmware and the operating system. However there are a few systems that do expose this to the user, as I recently found out with my Whiskey Lake-U platform, where the OS power plan has advanced options to set the TDP levels. Very interesting indeed.
Also, for Ice Lake-U, Intel is introducing a feature called Intel Dynamic Tuning 2.0.
We covered this in our architecture disclosure article, but the short and simple of it is that it allows OEMs to implement a system whereby the PL1/TDP of a system can change based on an algorithm over time. So it allows for higher strict turbo, and then adjusts the turbo budget over time.
This feature will be branded under Intel’s Adaptix brand, which covers all these CPU optimizations. However, it should be noted, that this feature is optional for the OEM. It requires the OEM to actually do the work to characterize the thermal profile of the system. We suspect that it will be mostly on premium devices, but as the chips roll out into cheaper systems, this will not be there. Intel is not making this feature standard.
Testing Power
Based on the time available, we weren’t able to do much power testing. What I was able to do was run a power profile during the start of our 3DPM AVX512 test in both 15W and 25W modes for the Core i7-1065G7.
The test here runs for 20 seconds, then rests for 10 seconds. Here are the first four sub-tests, and there are a lot of interesting points to note.
The peak power in these systems is clearly the PL2 mode, which on the Intel SDS platform seems to be around the 50W mode. Given that the functional test system is a bit of a chonk, with a strong thermal profile and the fan on all the time, this is perhaps to be expected. The suggested PL2 for Kaby Lake-R was 44W, so this might indicate a small jump in strategy. Of course, with the Kaby Lake-R designs, we never saw many devices that actually had a PL2 of 44W – most OEMs chose something smaller, like 22W or 35W.
The fact that the CPU can sustain a 50W PL2 means that Intel could easily release Ice Lake into the desktop market at the 35W range. Easy. Please do this Intel.
Second to note is the AVX-512 frequency. Not listed here, but under the 15W mode we saw the AVX-512 frequency around 1.0-1.1 GHz, while at 25W it was around 1.4-1.5 GHz. That’s quite a drop from non AVX-512 code, for sure.
Third, we come to the turbo window. Increasing the base TDP means that the turbo window has more budget to turbo, and we can see that this equates to more than 2x on all the sub-tests. In the 15W mode, on the first test, we blow through the budget within 5 seconds, but on the 25W mode, we can actually turbo all the way through the 20 seconds of the first test. This means that there is still technically budget on the table by the time we start the second test under the 25W mode.
Also, that third test – if you are wondering why that graph looks a little light on the data points compared to the others, it is because the AVX-512 instructions took so much of the time on the CPU, that our power software didn’t get any for itself to update the power values. We still got enough to make a graph, but that just goes to show what hammering the CPU can do.
For the base power consumption, we actually have an issue here with the observer effect. Our polling software is polling too often and spiking up the power a little bit. However, if we take the average power consumption between 25-30 seconds, under 25W this is 2.96W, and under 15W this is 2.87W, which is similar.
For users interested in the score differential between the two:
For 3DPM without AVX instructions, the 15W mode scored 816, and 25W mode scored 1020 (+25%).
For 3DPM with AVX-512, the 15W mode scored 7204, and 25W mode scored 9242 (+28%).
Facebook
Twitter
RSS
261 Comments
View All Comments
Phynaz - Friday, August 2, 2019 - link
What? TDP doesn’t mean what you think it does.Alexvrb - Monday, August 5, 2019 - link
I didn't feel like quoting the entire paragraph. But please DO elaborate. Then tell me how useful TDP is when they let OEMs set PL2 and Tau to... anything, really. You can take two "95W" processors and their power and thermals under load are radically different across a range of mainboards. The is reflected in mobile as well, where they let OEMs do pretty much whatever - the results aren't constrained by the processor no matter what the claimed TDP is. That doesn't even COUNT overclocking.Meanwhile AMD chips don't hand over control to mainboards unless you ARE overclocking, which is how it SHOULD be.
Alistair - Friday, August 2, 2019 - link
I didn't see any discussion or comparison vs. the i7-9850H. Let's see a 28W TDP version of the 6 core i7-9850H put against these new chips. Same money, 50 percent more cores. Anyone in their right mind should be looking for an i7-9850H or 9750H laptop instead over these 10nm products. Where is the 6 core 10nm CPU? Don't buy a 4 cores laptop if you're looking for good performance in 2019-2020 imo.If you want a 4 core laptop get a cheaper 14nm based laptop. If you want performance get a 6 core. I really really don't see the point in these products.
Alexvrb - Friday, August 2, 2019 - link
They gotta do *something* with all those 10nm wafers. Ian can't eat them all, and China said they don't want any more half-baked 10nm products after the last go-around. Maybe in 2020 we'll see 10nm++ and it will be as good as phase one 10nm was supposed to be.But yeah, their current 10nm products are a bit disappointing outside of the fatter GPUs and better memory speeds. If you're using something with a dGPU there's little point vs their own 14++, it only starts to make sense if you want AMD-like iGPU performance with the latest Core processor design. Even then that's only limited to models with a high EU count (48+) as the 32 EU models just look meh.
They're going to have some stiff competition when 7nm Zen 2 APUs launch. I guess that's why they're attacking the low-power first, as AMD is still stuck on 12nm rehash Zen+ products for now.
InvidiousIgnoramus - Friday, August 2, 2019 - link
I still find it amusing that the architecture with "Ice" in it's name has low clock speeds presumably from power/heat issues.abufrejoval - Friday, August 2, 2019 - link
Great work! And kudos to AMD to make Intel work so much harder to get good news out!Two die carrier layouts but the chips looking identical:
First of all, I assume that the bigger and square chip is essentially the North-Bridge in 14nm?
And the smaller rectangular one the CPU+iGPU?
And I guess at 64EU we are talking about more than 60% of die area going to iGPU while even at quad core and AES-512 the CPU + cache will be perhaps 30%?
Is there any HSA or GPGPU compute to 'pay' for that iGPU surface and power in professional workloads?
Or is it really just for gaming?
Am I also correct to assume that of the extra thermal budget in the 28Watt parts, none really goes to the CPU, only allows it to stay within the 15 Watt envelope while the iGPU is also running?
Are we talking different die layouts and sizes for dual/quad CPUs and 64/32 iGPU EUs or is it really all just binning, meaning that an Core i3-1000G1 is a chip where 70% surface area of an Core i7-1060G7 failed to make it?
Why am I thinking they are heading down a path without consumer value returns?
I got a Lenovo S730 i7-8565U or Whisky Lake recently for a little over €1000 and I got a couple of J5005 Atoms recently for a little over €100 (admittedly complete notebook vs. RAM less Mini-ITX mainboard). The difference in power is 15 vs 10 Watts.
Both are fairly competent 2D machines even at 4k. Both are terrible gaming machines, but I don't really think that ultrabook portable gaming performance is a selling point.
If I were free to choose CPU vs. GPU real-estate, I'd definitely go left, say 6 or 8 CPU cores or just higher sustained turbos and make do with the J5005's 18 iGPU EUs, because CPU power is what I profit from professionally.
For GPU, every € I spend gets me vastly more gaming experience in less mobile form factors, which is fine: I don't see how I could run in a game and outside without breaking my newest toy.
Sahrin - Friday, August 2, 2019 - link
$426 for a quad core in 2019. What a time to be alive.eva02langley - Friday, August 2, 2019 - link
So basically... expensive, low yield, 4 cores, low frequency.Outside of better IGPU, barely matching AMD offering, and AVX512, which is not even a matter for a 4 cores CPU, 10nm is an abysmal failure.
Phynaz - Friday, August 2, 2019 - link
So basically....you’re an imbecileKorguz - Friday, August 2, 2019 - link
your one to talk phynaz, i guess you want to be stuck on quad cores in notebooks for ever ???