Microsoft Surface Pro 2017 (i5-7300U, 256 GB) Convertible Review

[Redaktion]

Olden but golden. Without USB Type-C or Thunderbolt 3, the 2017 model of Microsoft's Surface Pro seems kind of dowdy. Fanless operation we've already seen on Acer's Alpha 12, and unfortunately the device throttles noticeably. Is the only benefit left really its decent battery life?

https://www.notebookcheck.net/Microsoft-Surface-Pro-2017-i5-7300U-256-GB-Convertible-Review.230716.0.html

[Anon]

Is PWM detected at all levels of brightness? Or below certain %?

[dthrp]

Battery life is finally decent, and that PM971's 4K performance is truly incredible. Almost makes me want to forgive the lack of ports.

[Digitalguy]

Will you be testing the i7 model?

[GreyFox]

Why would you calculate throttling from max turbo boost frequency, it's designed as a short temporary power boost above TDP.  This is nonsense, throttling wouldn't occur until it drops below base frequency. 

[Danelaw]

Battery life is finally decent, and that PM971's 4K performance is truly incredible. Almost makes me want to forgive the lack of ports.

say what??

i can not see what is so impressiv. on that NVMe drive? (please explain)

here is what im getting on a el cheapo ultrabook laptop 600USD and its OEM PM961and that is the old mi13 model.
http://i.imgur.com/cHt2KpF.png

[dthrp]

I could be wrong, but there's a mix up here between NBC's 4K and 4K QD32 results. Still, considering the subject's 256 GB capacity (with SSDs, larger capacity = better performance) and physical size, QD32 read/write speeds around 800 MBps is a noticeable improvement from PM961, or even the higher-end SM961. Benchmarks in other 4K areas aren't bad either.

Random 4K, especially QD1 or just "4K" mainly determines the "snappiness" of UI experience, and QD1~4 is what matters most for a typical user. Which explains why "Real life performance was excellent – application start times or file copy transfer speeds were very fast." I doubt majority of people actually need 4 digit sequential speeds on the SP, though it's nice to know the potential. As for QD32, it's rather relevant to server level performance, but I guess I'm excited anyway.

I'm no expert, so here's a much better explanation:
http://www.overclock.net/t/1231707/can-someone-explain-the-different-crystaldiskmark-tests
http://www.violin-memory.com/blog/understanding-io-random-vs-sequential/

[Julien Z]

Please switch the display mode from "Enhanced" to "SRGB" and the results will be better. ^ ^

[volodya]

Personally, besides the odd video or audio conversion, I need acceptable responsiveness when juggling big MS Office files, such as big MS Word documents containing MS Excel objects and graphs, possibly with track changes on, and the m3 of the previous generation proved a bit unpleasant. This year's i5 / 8Gb is good enough and it never gets hot under such scenarios

[Howard Cadwell]

Unless I missed it, the comments on throttling are silent on whether the unit was tested with max battery life selected or max performance as the alternative.

Other reviewers have, by contrast with this article, pointed to very good performance on throttling when the performance cpu selection is enabled, and explained the throttling in other modes as by-design.

[Emerald]

I've been reading threads on systemic backlight bleed present on New Surface Pro's from 2017. While your unit appeared not to have it, off the shelf or even display units apparently display this defect consistently, and users on Reddit seem not to have had success finding improved models through replacements.

[robgoren]

I did my own testing on the i5 model and, like the Core M, could not detect any PWM. I wonder if the detected PWM is coming from the power supply?

[Steven Wong]

Hi, I recently brought an i5 SP2017, 8GB RAM. Most of the time it performed quite smooth. However, I found that when I am going to install large program such as Adobe CC, or working on major update (eg. 1703 to 1709). The installation progress seems slow down significantly. Can you guys help us to verify such problem?