Isn't Extreme just going to be VSync limited on most devices? It is unfortunate that they don't have a 1080p Unlimited version, but it does make Unlimited GT1 an... interesting geometry test
Surface 3 powered by cherry trail
- Thread starter eastmen
- Start date
Hm, Extreme is more than 1080, it offers more visuals. My Acer W4-820 with Z3740 reaches ~ 8.200 3DMarks in Extreme, but 15.000 in Unlimited. My iPhone 6 is too powerfull for Extreme (maxed out), but its result in Unlimited is only 17273.
Silent_Buddha
Legend
Well, the first mention I've seen on battery life.
http://gizmodo.com/microsoft-surface-3-review-the-tablet-i-want-at-the-pr-1696598172
It's a glowing review for the most part as are most of the articles mentioning the Surface 3 that I've seen thus far.
It has the battery life at a bit over 6 hours in their standard test. Which uses Chrome. So if you use IE or Firefox that would likely jump up to 7.5-8 hours in their standard test. I sometimes wonder if Google deliberately makes the Windows version of Chrome use an arseload more power than any other browser just to make Windows tablets battery life look worse than they are in reviews.
So I could see more typical users potentially getting close to 10 hours battery life assuming they don't use Chrome for their browser and use WMP or another power efficient media player for video.
Waiting to see what some sites have to say about battery life that benchmark both with IE and with Chrome. Looks promising though. Tempting to just go ahead and pre-order it.
Loved what they had to say about the Surface Pen on the Surface 3. I still hate that it uses a battery, but that's probably a trade-off I can live with. I'll just have to invest in some extremely hard to find rechargeable AAAA batteries.
Regards,
SB
http://gizmodo.com/microsoft-surface-3-review-the-tablet-i-want-at-the-pr-1696598172
It's a glowing review for the most part as are most of the articles mentioning the Surface 3 that I've seen thus far.
It has the battery life at a bit over 6 hours in their standard test. Which uses Chrome. So if you use IE or Firefox that would likely jump up to 7.5-8 hours in their standard test. I sometimes wonder if Google deliberately makes the Windows version of Chrome use an arseload more power than any other browser just to make Windows tablets battery life look worse than they are in reviews.
So I could see more typical users potentially getting close to 10 hours battery life assuming they don't use Chrome for their browser and use WMP or another power efficient media player for video.
Waiting to see what some sites have to say about battery life that benchmark both with IE and with Chrome. Looks promising though. Tempting to just go ahead and pre-order it.
Loved what they had to say about the Surface Pen on the Surface 3. I still hate that it uses a battery, but that's probably a trade-off I can live with. I'll just have to invest in some extremely hard to find rechargeable AAAA batteries.
Regards,
SB
I think, Microsoft should name the Surface 3 "Surface 3C". Everyone expected 2013 a cheap platic iPhone, but iPhone 5C wasn't cheap.
@Silent_Buddha the AAAA battery lasted like eons for me. Cant you simply buy a bunch of the one-use AAAA available on stores?
D
Deleted member 13524
Guest
Wow, I'd forgotten how terrible Gizmodo's reviews are.
My brain hurts from reading that.
My brain hurts from reading that.
More numbers: http://www.neowin.net/news/surface-3-benchmarking-the-younger-brother
"To no surprise, the Pro 3 performed exceptionally better than the Surface 3 but the Surface 2 and 3 scores were closer together than I originally anticipated. The gap between the Pro 3 and the Surface 3 clearly shows that the younger sibling is by no means a powerhouse when it comes to the GPU."
Something is wrong here.
"To no surprise, the Pro 3 performed exceptionally better than the Surface 3 but the Surface 2 and 3 scores were closer together than I originally anticipated. The gap between the Pro 3 and the Surface 3 clearly shows that the younger sibling is by no means a powerhouse when it comes to the GPU."
Something is wrong here.
D
Deleted member 13524
Guest
AFAICT, Ice Storm Unlimited uses a very simple scene with hardly any compute/shader demands, so it ends up being a bandwidth and fillrate-limited test.
3dmark is getting really old and that android<->windows crossover test needs to be revamped from the current DX9/OpenGL ES 2.0 lowest-common denominators.
GFXBench's T-Rex should be a much better comparison. In fact, Manhattan should be an even better comparison for real-life usage, but Tegra 4's GPU on Surface 2 is still stuck with OpenGL ES 2.0...
Summarizing: let's wait for actual game tests.
3dmark is getting really old and that android<->windows crossover test needs to be revamped from the current DX9/OpenGL ES 2.0 lowest-common denominators.
GFXBench's T-Rex should be a much better comparison. In fact, Manhattan should be an even better comparison for real-life usage, but Tegra 4's GPU on Surface 2 is still stuck with OpenGL ES 2.0...
Summarizing: let's wait for actual game tests.
The graphics is all the same as Broadwell, though (Gen 8). Just 16 EU (2x8) instead of 24 (3x8). There is just one (rather important) difference, namely no access to LLC which does affect driver a bit but really just a tiny portion of it. And in probably the same way as it affected Bay Trail vs. Ivy Bridge.
That said I'm wondering if all these tested Surface 3 were the same 4GB memory version? Just wondering if the 2GB version could be single channel, which could definitely affect the scores, possibly even more so than on Broadwell/Haswell - while the alu / mem bandwidth is lower due to less EU and lower clock the lack of LLC could potentially make this worse, not sure though. Bay Trail didn't really suffer from it IIRC but it was very ALU limited in the first place.
Ike Turner
Veteran
Looks like there's a bug with the IGP drivers...rebooting the Surface 3 and re-running the benchmark results in a 80% increases...wth..drivers are not finalized yet..
https://www.thurrott.com/mobile/mic...ons?utm_medium=twitter&utm_source=twitterfeed
https://www.thurrott.com/mobile/mic...ons?utm_medium=twitter&utm_source=twitterfeed
Last edited:
Such a huge difference clearly isn't just "IGP drivers" but points to an issue with power management imho. I wouldn't know though why it would run at twice the clock after a reboot. Maybe there simply was some background process going crazy when doing the first benchmark run (even if the benchmark isn't cpu limited due to TDP limits this likely would make quite some difference).
Ike Turner
Veteran
And this is not an isolated issue..Neowin's results are around 15k which are what Thurrot was getting before the reboot..
Looks like there's a bug with the IGP drivers...rebooting the Surface 3 and re-running the benchmark results in a 80% increases...wth..drivers are not finalized yet..
https://www.thurrott.com/mobile/microsoft-surface/2849/surface-3-desktop-applications?utm_medium=twitter&utm_source=twitterfeed
Thanks for this link.
iTunes runs good enough, but encrypted 1080p videos are too much for my Acer W4-820 with the Z3740, 720p videos run fine. DRM free videos run with HW accelleration.
It's a bit schitzophrenia: Xbox Video is a horror, but it uses HW accelleration for enrypted videos from the Xbox Video store, which means better battery life. iTunes is much better, but without HW accelleration it consumes a lot of energy.
Ike Turner
Veteran
I wouldn't say that it's a horror especially on a touch enabled device...but it still lacks basic features...like the ability to delete a friggin local video file! Wtf MS?! Performance is top notch though..
Silent_Buddha
Legend
Part of it might be due to using the docking station. One preview I saw mentioned that when docked, their Surface 3 had a habit of heavily throttling itself, leading to low benchmark results. This didn't appear to happen when undocked. They conjecture that this is because the dock encompasses a part of the rear of the Surface tablet. The surface that is used to cool the device. So when docked, it isn't able to cool itself as well as when it isn't docked.
https://www.thurrott.com/mobile/microsoft-surface/2837/surface-3-docking-station
Third picture down shows this quite well. There's very little of the chassis (used for passive cooling) exposed.
This isn't a problem for the actively cooled Surface 3 pro, but appears to be a potential problem for the passively cooled Surface 3.
Regards,
SB
Last edited:
Realistically there should be a difference due to a much slower CPU and lower thermal envelope even compared to Core M. The only real LLC access off vs. on is the Phoronix review which shows ~20% difference.
Also, Intel must really skimp on driver development. Architectures may look similar, but often needed new drivers to get up to full speed. Remember the low clock bug for Ivy Bridge? I think someone said it wasn't fixed when Haswell came out. 2 years and you are out.
That makes no sense. Core M has problems with passive cooling. Atom? How did the 22nm generation get by in 8-inch Tablets but this one can't in a 10 inch one? Problems with 14nm process?
Silent_Buddha
Legend
Cover up all the passive cooling surfaces and even ARM CPUs are going to throttle like mad.
Regards,
SB
I was stupid enough to buy TBBT Season 6, The Walking Dead Season 1 over Xbox Video. There are huge DRM problems. I cannot download some episodes, downloaded episodes cannot be played and have to be re-downloaded, what doesn't work.