theveterans
Member
Looks normal to me since the ave voltage is 1.16v. Your CPU speed is too high for web browsing IMO. I'd keep the profile to "better performance" instead of "best performance"
CPU Throttling on 15 inch Surface Book 2 after 1709?
- Thread starter Bill Dolan
- Start date
Just got my 15 inch SB2 Friday and am trying to determine if I have a problem. Using Best Performance on intensive games, I'm getting ~84c - 93c. Are those good temps? What's strange is that sometimes my temps actually increase -after- I exit the game. I've seen them hit 100c afterward when I am back on the desktop. Is this normal? It's almost as if the CPU fan ramps down before it has properly cooled off the CPU
theveterans
Member
^ It might be that at the moment you close the game, the CPU is at maximum compared to when you're playing games. Gaming temps are normal IMO, but during exiting is not.
wynand32
Well-Known Member
Other people know more about this than me, but I'll note that my SB2 15" has never hit 100c and has maxed out at 92c in all of my testing. I'd be suspicious of a machine that hits 100c at any point, that seems to be an indication with some other people here of a problematic machine.
Just wanted to ask a follow-up question. Any thoughts on whether gaming on the SB2 15" will cause long term damage, particularly to the screen/tablet section? With CPU temps reaching the low 90's the back does get quite warm, and these temps would remain fairly consistent over a 1-2 hour gaming session.
theveterans
Member
^ I wouldn't worry about the screen though, if you can find an X-ray image or a ifixit image of SB2's internal on the tablet side, look at where the heatsinks are laid out on the batteries. If they're are next to them, there's a possibility for a premature battery degradation due to heat.
chrisoutwright
New Member
Well, it seems SB2 15" has indeed batches with bad CPUs. I myself got 1.41V in max VID, outrageous!
Have called up MS Support in Switzerland, and guess what? They hung up as soon as I mentioned benchmarks and VID. I think they did not even understand what these terms and numbers meant, nor were they interested.
In Cinebench a got scores between 500 and 600.
In ThrottleShop with 1024M computation time was 426s.
Is this bad? IntelProcessor Diagnostic Tool showed pass everywhere, so I am not sure if I should return my SB2, as temps stay OK apart form the outliers.
It would be my 3rd exchange so far for SB2.
It should be noted, that undervolting is possible in my SB2. So far I managed -90mv for some time. I am sure that -70mV are achievable for a longer term under full usage. This is a good sign, right?
There is a reddit post by TickleMyElmos, stating Cinebench scores over 700 with changes on the TDP via XTU. Is this even possible, considering more realistic score are within 500-600?
Via XTU I got a bench somewhat below 1000, about middle ground in the XTU ranking that can be seen in HWBOT for the same CPU. These scores are for several devices with the same CPU.
What do you think about TickleMyElmos's score?
Btw: on a second try with ThrottleStop, I tried to adjust the TDP values via the XTU tool for both Turbo Boost Power Max and Turbo Boost Short Power Max to the same value of 33W, also using an undervolt of -80mV. Funny enough ThrottleStop needed only 300 sec this time while being on battery.
A huge increase, what gives?
Have called up MS Support in Switzerland, and guess what? They hung up as soon as I mentioned benchmarks and VID. I think they did not even understand what these terms and numbers meant, nor were they interested.
In Cinebench a got scores between 500 and 600.
In ThrottleShop with 1024M computation time was 426s.
Is this bad? IntelProcessor Diagnostic Tool showed pass everywhere, so I am not sure if I should return my SB2, as temps stay OK apart form the outliers.
It would be my 3rd exchange so far for SB2.
It should be noted, that undervolting is possible in my SB2. So far I managed -90mv for some time. I am sure that -70mV are achievable for a longer term under full usage. This is a good sign, right?
There is a reddit post by TickleMyElmos, stating Cinebench scores over 700 with changes on the TDP via XTU. Is this even possible, considering more realistic score are within 500-600?
Via XTU I got a bench somewhat below 1000, about middle ground in the XTU ranking that can be seen in HWBOT for the same CPU. These scores are for several devices with the same CPU.
What do you think about TickleMyElmos's score?
Btw: on a second try with ThrottleStop, I tried to adjust the TDP values via the XTU tool for both Turbo Boost Power Max and Turbo Boost Short Power Max to the same value of 33W, also using an undervolt of -80mV. Funny enough ThrottleStop needed only 300 sec this time while being on battery.
A huge increase, what gives?
Attachments
Last edited:
chrisoutwright
New Member
Well it has been some time, and since nobody responded on my findings here, I returned my SB2 15" for another one.
As it turns out, the voltage max VID is about 1.26-1.34v and thus lower than on my last unit.
I therefore confirm that some batches have flaky CPUs, replace immediately as long as you can.
As it turns out, the voltage max VID is about 1.26-1.34v and thus lower than on my last unit.
I therefore confirm that some batches have flaky CPUs, replace immediately as long as you can.
Last edited:
Something to think about - It might or might not be part of the isses(s). Let me know what you think if time permits.
Here's how much the Meltdown and Spectre fix hurt my Surface Book performance
