Mass performance benchmark


#21

My first benchmarks with the i7-7y75:

Balanced Performance Bios Setting (Default):
PCMark10 2936, 2954
CinebenchR15 GPU 32.47
CinebenchR15 MultiCPU 241, 233
CinebenchR15 SingleCPU 123

Max Performance Bios Setting:
PCMark10 2786, 2802
CinebenchR15 GPU 32.26, 34.28, 31.59
CinebenchR15 MultiCPU 237, 226
CinebenchR15 SingleCPU 127, 125

Blank out of the box plugged-in V, latest Windows updates, disabled antivirus, standard drivers. Can’t say why the max performance setting is actually worse than the balanced one. Can’t say I’m super exited about these results, either. However, V feels fast enough so far.

I’ll update as far as I get new stuff, but it seems I can validate the test by notebookcheck (https://www.notebookcheck.net/Eve-V-m3-i5-i7-Convertible-Review.264092.0.html)

What impresses me most is that you don’t hear anything - at least most of the time. Interestingly, while Cinebench OpenGL test loads (at the very beginning, some images are fading), you can hear something ^^ Same when you boot up. However, I wouldn’t say it is coil whine (it is not a high pitching tone). Checked again, sound is also appearing when opening Edge for example.
It seems that this sound is coming from the USC-C power cable when it is plugged in (at the side of the V). I guess somehow, the V needs more power and fetches is then over the cable? After I removed the cable, the sound is gone :open_mouth: Same with both ports (USB-C and TB3).

What I also experience is a kind of “prickling” or “tingling” when moving your fingers over the back of the V as long as it is plugged in. Have the same problem with the body of the HP Spectre x360 (old version) when I use a cheap charger without the grounding.


[WIKI] - Fantastically outdated collection of answers to questions about the V
#22

Anything you can compare these to to get a frame of reference pls?


#23

Unfortunately not yet, seems that the PCMark results cannot be compared (with Futuremarks own benchmark) without a good set of reference benchmarks, for my 4year old desktop pc at home I get a nice comparison and actually a much higher score (which is not a surprise) :wink:

It seems I cannot even compare these results with the V reviews currently out there, as they primarily used PCMark 8 … doh!


#24

Well as it is it’s a starting point, we hadnt chosen anything else so PC Mark 10 seems as good as any :slight_smile:

I’ll use it on mine if it ever arrives :wink: circa January hopefully


#27

You are right to say that benchmarks should not be the decision factor for the end-consumer.
In my case I use the computer to compile programs of small to medium complexity, so for me those benchmarks can tell me if that specific CPU would gain me that time advantage over another. (even 10s split is a great deal for me).
Also, this is a 7W TDP Y-series Intel CPU, there are almost none sintetic benchmarks on the internet for such thing, so it will be nice to see more of these. (besides the old ones on this forum)

I’d request from the comunity an m3 vs i5-Y comparison if possible, or simply benchmarks for both. I might buy the m3 but I’m not quite sure about the performance and bottlenecks with eGPU (not gaming)


#28

Do you have a step by step guide for these - I’ll replicate them with my i5.

I did dl cinebench earlier and ran the one called CPU which did some rendering square by square and got 224. I had no idea what this meant and moved on to more fun things.


#30

Thanks, I sent you a PM!


#31

Quick test. With i5 version, no overclock. Stable at 2.1GHz.

My score was 230

EDIT: Benchmark was Cinebench r15! Sorry forgot to say it


#32

I’m down for helping benchmark my V (when it comes), but I’m confused as to the procedures to do so. As far as I know, it takes the V a while to “warm up” before it starts operating at peak efficiency.

Perhaps we should take two benchmarks, one on day one and another a week later?


#33

Well, the guys from notebookcheck.net get ahead of me with the same idea.They benchmarked all 3 processor models, even by rising the TDP from 7W to 12W. You can check them here: m3, i5, i7 benchmark Eve V

TL;DR : Eve has kept their word and made Eve with Y-series CPUs rocking 7W TDP that are somewhat close to U-series counterparts in terms of performance, and not only burst but maintained performance! Good Job there!

As for the difference between the 3 chips, that’s debatable !
The guys had a really weird anomaly where the m3 with the same Intel HD 615 outperformed the other 2 in a graphics test. They blamed it all on “silicon lottery”, which might be true but to me it seems that the m3 with it’s lower frequencies is able to sustain longer and actually thermal-throttle the GPU less (need a couple more sources on that before being sure).
Besides this weird one test, they are 7-8% apart one from the other, benchmark-wise. The i7 brings enterprise-grade features like Intel vPro and also (strangely only hereIntel TSX-NI which helps with multi-threaded workflows, where the i5 has none of those but offers higher boost clocks compared to m3.

In my opinion, strangely as it seems, the m3 is the winner as it offers longer sustained performance (even at max freq of 2.60 instead of 3.20 GHz of the i5) for the lowest cost. A true bang for the buck.
Obviously i5 is the go-to option for those who can’t stand 2.6 GHz or need the extra storage option!


#34

Yep they got a magic m3!


#35

I’ve been watching the behavior of my i5 processor and it looks like it can easily pass notebookcheck’s m3 score in cinebench. I’ve run Throttlestop and undervolted the CPU by -70mv. When I run cinebench I get the following:

2.8Ghz at 12-14watts and temps at 80°C
2.4Ghz on all cores at 8-9 watts and temp at 67°C
2.2Ghz on all cores at 7 watts and temp at 63°C ( this is when PL1 kicks in )

Is there any way to bump the 7w limit to like 11w? (maybe not 11, maybe like 9w) I’ve been playing Fortnite and Overwatch with egpu. I didn’t post my gaming benchmark of those games yet because my FPS have been all over the place. If PL1 can get me to a stable 2.4Ghz the V would be a pleasant egpu gaming machine because right now PL1 kinda kills it.


#36

Intel xtu can do it. But currently has issues with suspend. (at least on mine. )


#37

You can check this post. Talks about TDP adjustments and performances

https://eve.community/t/ikirins-i5-performance-report/6126


#38

I’ve been following this thread for my tests and the processor keeps hitting PL1 and throttles down to 1.6Ghz


#39

Oh gotcha.
202020202020


#40

This is an issue that someone with a U processor is having. Maybe it applies to the Y.

https://communities.intel.com/thread/110446


#41

This really bugs me. My friend gave me her tablet with an M3 6Y30 processor and the tablet stayed at a steady 2Ghz during the whole cinebench test. PL1 or PL2 never kicked in. No thermal throttling also. If my V behaved like that, gaming would be stellar.

I went into the BIOS and set the system performance from standart to max. During load it stays steady at 2.4Ghz before PL1 kicks in again.


#42

@iKirin @Konstantinos Hopefully, Eve can make some BIOS fixes for the higher performances settings.


#43

After further testing, my cpu stayed steady at 2.3Ghz during the cinebench test and steady at 2.6Ghz for the XTU stress test. PL1 kicked in for both but it did nothing to the clock speeds. Thermal throttling kicked in only during the XTU test and it went down to 2.4, 2.5Ghz. Both test had an undervolt of -70mv.

I don’t know yet, but it seems like the CPU is doing better on battery than wired. Needs further testing just to be sure.

I’ll keep these settings for now and I’ll test some games and post benchmarks in my thread if needed.