Lenovo y510p Heat/Fan question

Hey guys I've had my y510p for a few weeks now and today was the first time that I decided to actually play games on it for a long period of time and I noticed some things and I'm not too sure if it's normal.

So I was playing on it for about 4-5 hours and I had my CPUHW monitor up the entire time and I was ok with the temps that I saw except for the SLI 750m in the ultrabay. It reached a maximum of 96 degrees C and it definitely felt hot because my hand on my mouse was on fire and so was the number pad. So this occurred while playing World of Warcraft on the "High" preset. I would think that there are more intensive games that would push the card and the temps further than what I saw today and I am wondering is that normal?

 

Nope this is not normal. Your Ultrabay 750M is overheating. It starts throttling at 97C, and you'll notice a huge performance drop at that point because the core clock speed will be cut in half until temperatures drop. You're correct about WoW not being a super demanding game, so something which pushes your system harder is gonna result in even higher temperatures.

For reference, here are my maximum CPU and GPU temperatures after playing BF3 nonstop for several hours:



This is about as hot as my Y500 ever gets when gaming, even though I'm running a pretty hefty GPU overclock. Compared to the Y510p, the Y500 is an all-around cooler-running machine, but your temperatures can certainly be improved.

There are many ways to reduce temperatures. Reset clock speeds if you are overclocking anything. Decrease ambient temperature with AC if you are in a hot room. Elevate the bottom of the laptop to increase airflow, or get a good cooling pad. If those don't fix it, then the last resort is to open up the laptop and reapply thermal paste to the CPU and GPU. I've been through a couple of these laptops myself and poor factory paste jobs seem to be the norm. I did a repaste and use a cooling pad, and the combination dropped my temperatures under load by a fair amount.

 

Hi,

nope this is normal for y510p

650m max voltage is 1.025, 750m default voltage(ultrabay) is 1.087, and memory speed is higher too.
Bottom cooling does nothing, thermal paste replace is a good start, but still gonna see 90+ temps.

Reason is, springs too weak, and thermap gap pads of memory pushes heatsink away. Lenovo came around that issue with 4x+ amount of thermal paste than neccessary.

What I did is changing bolts in ultrabay heatsink, no springs, but rubber spacers, and with that came greatly increased pressure.
Corner of the heatsink is slightly bent, but thermal paste footprint is much much better.
Internal gpu heatsink was modified too, with 0.5mm spacers under springs.
Don't need to mention, this is really against warranty, and don't do it, unless you know what you're doing.

With games like BL2, where general gpu usage is lower, i can maintain <83 temps on gpus. (1189/2800) But this is with undervolted cpu, and disabled HT via unlocked bios.
Where gpu is constantly in the upper 80s, 90s, like Serious Sam3, temps reaching 90+. Heatsink and fan speed is way too weak for this kinda heat.
I would fire all the engineers, worked on that machines cooling....

Lastly, you can lower throttling temp of ultrabay, but it will lower speed as well, or make a custom heatsink.

 

Yeah, the 750M does run hotter and at higher voltage, 1.125 V to be exact, not 1.087 V. Kind of unnecessary if you ask me, as I am running my 650M well past 750M speeds with 100 mV lower Vcore. VRAM doesn't affect temperatures of course.

Still, I don't recall the 750M Y500 ever having overheating issues this bad. I mean, I only see the kinds of temperatures OP is getting if I run my GPU core way past spec, like 1260 MHz @ 1.137 V. If anything, I would think the Haswell CPU is the biggest culprit in the Y510p's thermal problems, as it is indisputable that Haswell runs hotter than Ivy Bridge due to Intel moving the VRM onto the CPU itself. The increase in heat is why overclockers are having a hard time hitting even the same frequencies they achieved on Ivy Bridge. Intel is putting the VRM back onto the mobo with Broadwell but this means your current Haswell mobo will no longer support Broadwell CPU's.

 

It is 1.087 to be exact. Integrated is 1.056 to be even more exact. Probably it is connected to ASIC quality, but it is not confirmed.
You don't know what the exact VRAM voltage is in these cards, so 750M memory could be overvolted compared to 650M, and frequency in itself is affecting temperatures.

Haswell is really hotter, but with undervolting, it is not a real isse, integrated gpu is never reaching throttling temp(after mod), even with forced cpu speed.
Ultrabay is a different story, that heatsink is not enough...

edit: wrong picture attached

Correction, max voltage is sometimes 1 notch higher than those.

 

Agreed. Anyways, because the main GPU is so close to the CPU, that is probably one of the main heat sources. And the Y510p is not the only laptop that has heat issues. Any laptop under 1.5 inches thick running a quadcore Haswell runs hot.

Imho, the undervolted 4702MQ should have been standard in all Y510p and Y410p laptops. But is it possible to undervolt our own cpu's now that the unlocked bios for Y510p/Y410p is out?

 

CPU and GPU distance doesn't matter IMHO, the common heatsink is.
Cpu heat individually is no problem with undervolt, cpu and gpu heat together, with single heatsink is problematic.
This is why Lenovo implemented cpu throttle at 65+ gpu temp.

When your cpu heatsink is attached to motherboard vrm, then it still adds to cpu heat. When not, motherboard becomes hot.
I think it was a good move from Intel, because this way vrm heat can be extracted faster. I go forward, with this, vrm efficiency is better, so vrm+cpu heat together compared to previous architecture is better.

Do we have information regarding 4702mq default vid, so it can be compared to 4700mq? I dont think it is undervolted by factory, just TDP limit and core speed is lowered.
When i set 4702 speed on my 4700, its tdp is lowered to 4702 level. With undervolting it can be even lower.

Undervolting can be achived without unlocked bios, You just have to install Intel XTU utility. Mine does ~ -100mV, so try somewhere there first.
You can undervolt core, L3 cache, and iGPU too.
With unlocked bios, and proper bios settings, you can even adjust TDP limits, turn off short power max, which is 57W !

 

Hey thanks for the replies I know now that my ultrabay graphics card overheats then throttles, but I noticed you guys talking about the other gpu and the CPU and on mine I didn't have any problems with heat on them. My CPU maxed at 77 and the first gpu maxed at 68.

 

Wow, that looks awesome! I wonder if this will improve my battery life, since i do carry mine on campus a lot. Thanks!

EDIT: What settings do i put in to achieve 4702?

 

load them properly( even with forced max cpu speed) and wonder

 

Install XTU

set 1 core to 3.2,
2 core to 3.1,
3 core to 3.0
4 core to 2.9

With unlocked bios, and setting MSR lock to unlocked ( attachment) , you can set TDP limit lower.
You can also enable flex ratio to 2.2 ( max non turbo multi)
( You could also increase multi by 2, so 3.4 4 core load, but it is untested)

 

Hmmm everything I've seen seems to indicate the 750M stabilizing at 1.125 V: http://forum.notebookreview.com/ideapad-essential/716619-calling-all-y400-y500-gt-750m-owners.html

My 650M maxes out at 1.025 V but stabilizes at 1.000 V most of the time.

Take apart the Ultrabay and repaste with a good TIM. My money is on a bad factory past job. It's very easy to do: http://forum.notebookreview.com/ide...-pictures-disassembled-y500-ultrabay-gpu.html

 

Is undervolting the GPU advisable? I downvolted the CPU by 109 mV (no instability so far thankfully), and the temperature on average has dropped 8C. It'd be great if it could go lower, of course.

 

yes, but to achive that you have to set higher boost clock, but lower tdp, so gpu throttles back speed and voltage. google kepler undervolt.
so far unlocking tdp settings on mobile bios is no success( or no one tried )

 

@octiceps

That was 2 chip, almost 6 months ago, and we dont know their ASIC quality.

Probably nv achived better yields, better manufacturing process, that is why came 750m, current batches even better, lets do 755m...
Or just a simple overvolt, but with lower quality ( high leakage) IC you cant do that too much, without sky rocketing temps.
But these are just simple scpeculations

 

By the way, why is the CPU frequency locked at 2.10 GHz when it can go higher? I only noticed that when monitoring the temperature and frequency while running prime95. The odd thing is that Intel XTU indicated 0% CPU throttling when the frequency is locked there. Is there a way to remove this limit?

 

Actually, I'm pretty sure your 750M runs at lower voltage because of the updated BIOS and vBIOS in the Y510p. Lenovo probably figured the Y510p would run too hot if they didn't do something about it. Anyway, 1.125 V isn't necessary to reach 1058 MHz core unless it's a really low ASIC chip with lots of leakage or something. Those 1.125 V 750M's were from the refreshed Y500 model that was released before the Y510p.

You need to use ThrottleStop. The lack of Turbo Boost is a well-known issue with the Y510p. Just be aware that CPU temperatures will increase drastically with ThrottleStop. You could easily see a 20C or more increase. And since the CPU shares a fan with the first GPU the GPU temperature will rise as well.

 

Bios has nothing to do with default voltage. It is, like with CPUs the VID, is fused, and can be override by bios.
These cards doing 1189 with 1.06V, and 1.08V, and when they throttle back to 1058, voltage is around 1V, @940 somewhere around 0.95
You could ask: but where is the proof?
My ultrabay is running with bios from y500 unlock topic from techinferno, and default voltage stayed the same.

So these chips are really cherry picked compared to 650m, or like i said, improved.
(Cant say the same about 755m, we should get some info about them, ASIC quality, max clock, and voltage.)

 

It seems that you don't have to use ThrottleStop to remove the block on Turbo Boost. I set the Maximum Processor State back to 100% (it was 90% originally for some reason) and now the boost is re-enabled. Unfortunately, the temperature gets a boost to 90C+ as well on prime95. What is a safe range of temperature for this CPU again?

 

load the gpu, so it is 66C or higher, cpu speed will be 2.4

 

Oh I see what you mean now. Regardless, what would be a safe range of temperature? I know the maximum is 100C, but still, running around 90C makes me uncomfortable.

 

I don't think the 750M chips are cherry-picked, they're just running higher voltage and clocks compared to 650M. I saw some 750M with lower ASIC Quality than 650M. It would've been really helpful in this system if the 750M had been binned for heat output and power consumption. Here are the numbers on my 650M SLI for instance:

 

My desktop GTX 560 has an ASIC quality of 90% and my Y410p's 750m has an ASIC quality of 86.4%, both overclocked

Not too shabby

 

Hi ghoula! Would you please send me a copy of the unlocked Y410P/Y510P BIOS? I can not download that from the tech inferno website. My email is [email protected] thanks~

 

Hi ghoula! Would you please send me a copy of the unlocked Y410P/Y510P BIOS? I can not download that from the tech inferno website. My email is [email protected] thanks~

 

You can donate 5 USD to the site, and gain elite membership and then you will be able to download everything.
Or if you dont want to pay, make it to 10 approved comments in the forums.

 

Actually, it's 5 approved posts. Not very hard.

 

Even easier

 

I can send it to you if you like. But i do recommend donating to svl7 for his work.

 

Alright, i sent it.

Remember to follow the Instructions on techinferno and donate (i donated 2$).

 

Thanks a lot~