***Gigabyte P35X owner's lounge***

Check in the BIOS if the laptop is in AHCI or IDE mode,whatever its set to , change it to the other one, save and try to reboot. If your main drive is a HDD check if it has bad sectors.

 

I only have the AHCI option and the RAID option, no IDE. My main drive is a brand spanking new Samsung 850 EVO mSATA and it's healthy according to the Samsung software, as I would hope.

Sent from my Nexus 6P using Tapatalk

 

Hello all! P35x V3 here and I've got a few questions.
Can I upgrade to 32GB RAM with the DD3RL 1600Mhz 2x16GB sticks available?
Where can I get a bigger PSU? The stock 180 is not enough when overclocked.
For the future are there any means of installing external GPU without a thunderbolt?
Thank you all very much.
As my contribution I will bestow my perfect cooling solution. With this I run 3.3Ghz without throttling under maximum load gaming at ultra 3k and my GPU stays 65-69c.
But how you ask? How may one so well cool such a thin nuclear powered laptop? Why with this of course! http://scene7.samsclub.com/is/image/samsclub/0004601344280_A?$img_size_380x380$
Sitting on its back pointing upward and the laptop sitting with the hottest part right on the blower output.
Is it ridiculous? You bet it is. Is it loud? Quite. (but not with headphones). Does it eliminate throttling and even allow overclocking? Yes and very well. Im repasting this week, and I have a cheap cooling pad that didnt do much. maximum stable undervolt is -65mv core and -50mv cache (probably -65mv too but results have gotten confusing and i cant remember). Before while gaming I was lucky if the Cpu got above 2Ghz and the GPU stayed at 83-85c.

 

Haha, that's funny, but why not!

 

Could someone please tell me what's the button that's under the battery logo on the right front side for ?

THANKS !!!

 

when y'r laptop is off, click to see batterie % level on the 5 front led

 

Thanks !!!

 

Did my first ever repaste! GC Extreme. Major improvement!! CPU can hold 2.8-3.3 Ghz where before around 1.2-2.2 Ghz. GPU still gets up to 84-87c though. Is that normal? Should I try again? Tests were done in Crysis 3 ultra 3k, and Fallout 4 3k Ultra.

 

Repasted again this morning for slightly better results. CPU tears along nicely with minimal throttling and the GPU at 135/500 gets to 85 and flirts with 86c but no higher. Getting some serious performance and its worth mentioning I play games at 3k at ultra sometimes even with AA thrown in (not really needed). Quite a load and I am impressed at this little machine. Scored 1247 in unigine heaven using the regulation settings.

 

Ok, I upgraded to 10 from 8.1 last week, almost everything is fine except the Nvidia GPU.

When I was still using 8.1, my P35X would occasionally pop up a notification "graphic drive (nvidia windows kernel-or-something) is not responding and has recovered" (can't remember the exact wording), screens goes black follow by a few flickerings, then everything was fine again apart from closed the file explorer. This happens mostly when I try to play a video from hard drive for the first time after boot up.

But now it will show the notification then freeze (sometimes only the notification sound, everything else already freezed), then either stay frozen or goes BSoD with driver problem (nvlddmkm.sys). Either way, I'll have to force-reset the computer to restart. I uninstalled the geforce experience, googled and tried update geforce driver, uninstalled and reinstalled the driver, nothing works. The only way to use this laptop now is if I disable the Nvidia GPU through driver manager, otherwise it will freeze within hours (sometimes minutes after logged in). This is fine at the moment since I'll be using it to type up essays and reports, etc for the majority of time, but that also means I'll have to play games with intel hd 4600 and absolutely destroyed the purpose of this laptop.

Has anyone else experienced this issue when they upgrade? Or in 8.1 as well? What's your solution? I trued almost all the google solutions, I want to know what works for you?

Regards.

Defan


Sent from my D6603 using Tapatalk

 

Hey there. Are you sure you've tried it without geforce experience, when you reinstall drivers it tends to install itself again. Reason i mention is that I had hassles with BSODs linked to the graphics card until i got shut of that software. been stable and fine since then, win 7 and win 10. Just a long shot, from your post it sounds like you've tried it already...

 

Have you tried using DDU to do a full clean of the nvidia driver. Then install just the driver fresh? I've just had a very similar issue with my desktop which I hope has been solved by this. Also think Nvidia drivers are not great at the moment...

Sent from my Sony Z3 using Tapatalk

 

Hey guys I need help. GPU Z seems to be reporting PerfCap Pwr limit. Im running 135+core and 550+memory. This means I dont have enough juice right? Also my battery gets drained down to 96% sometimes while gaming and plugged in.
I saw on here that the Aorus x7 200w plugin would work, can anyone confirm the specific model and possibly provide a link to where I can buy it?
Thanks so much guys! I need MOAR POWAHHH
PS got a new cooling pad, GPU maxs at 84c now with full overclock.

 

What cooling pad did you go with?

 

I'm not entirely sure, but PerfCap Pwr refers to the power limits set in the vBIOS of the GPU and not on the power brick that supplies the laptop - that's my understanding. If that's the case then a bigger power adapter would not get rid of that limit; however, you noticed that your battery is being drained when gaming, if the P35x uses battery power to supplement the power brick, then it could be that getting the higher wattage power brick will improve general performance (perhaps your CPU is also being limited by that). Probably worth a shot getting the power brick, but you might need to flash a modified vBIOS to remove the PerfCap Pwr restriction you're seeing.

 

Thank you so much for the reply Robbo99999! That makes sense. I'm having trouble tracking down the 200w Aorus AC adapter though.
I have the means to flash the vBios but not the knowledge to properly edit it. I had tried before to raise the core higher but for some reason it did nothing once flashed so i reflashed stock vbios.
The cooler is the targus chill mat. Snagged it at best buy while i was there on a whim

 

Have a look here for some vBIOS that you could flash that would totally unlock your GPU (max power consumption raised / voltage & frequency sliders unlocked):
https://biosmods.wordpress.com/
Be mindful that you'll have to make sure you're flashing the right vBIOS that's compatible, just because it's a 970M or 980M vBIOS doesn't mean it's compatible - research it & check in forums, etc if it's compatible with your model. (If you get it wrong you could brick your laptop & make it unusable).

I'm not familiar with the P35x, don't know which vBIOS if any is compatible.

 

None of those bios are compatible but I edited my own and successfully flashed it. It took a metric butt ton of flashes but eventually I got the power table right. 1266 Core 3000 Memory and no power throttling. Temps can get up to 87 but did not throttle.
Next up I try 1300 core >
This may require the aforementioned turbo fan...

 

Anybody know if there's a way to reverse the F1-F12 function keys on these bad boys? So I could have them set so I didn't have to hold the function key to adjust volume, brightness, etc?

Sent from my Nexus 6P using Tapatalk

 

Impressive that you could edit it yourself, you're brave as well! Very good results, surprised you got that high with the overclock, although if 970M that's easier, or you have 980M?

 

Can you post a stable version of your modded vbios pls?

 

Its the 8GB 980m, and I was able to get 1304/3005 but temps were 88c, (63c with turbo fan )
I brought everything back down to stock clocks and am working on temps now, ive lowered the voltage table by 1 step across the board.
snoopy I would but I actually didnt save that bios, and unless you have an asic of 76.5% or higher I can certainly not guarantee its stability even then.
I wish I had saved the 1304 bios, but I have a better understanding of editing vbios now and if I want to push a major overclock again I can.
Right now I'm gonna see how low I can get temps, since Im pretty sure even though my 980m was not throttling, my cpu was power throttling. Now my goal is a perfectly smooth running machine with low temps. Ive actually noticed very slight power throttling with stock clocks on the GPU so im planning on fixing that while lowering voltage a little more.
If I do come up with a stable, cool vbios then I will upload it.
After that maybe I will take another pass at the overclock vbios

 

Well after spending literally all day I finally got the lowest voltage table possible without volt or power throttling.
Let me just say it was a lot of work. But the results are amazing to say the least.
Running stable at stock clocks I get a maximum temp of 83c under maximum load.
As far as the power table is concerned you cant really mess with it unless you are ok with power throttling to around 1066 or 1053 at times. Nvidia atleast did get it right for stock clocks, but not the voltage table.
What is also amazing is that my CPU does not appear to be throttling at all, but I will have to test that more. Currently running -65mV core and cache
Ill be doing more testing on the chill vbios and possibly start work on a overclock vbios tomorrow.
My chill vbios goes to the highest bidder at dawn!
EDIT: How does my temps sound? Should I repaste or is it as good as its gonna get? Ive repasted twice with GC extreme and the second attempt was better than the first.

 

Really would like to know this as well!

 

Sounds brilliant... can't wait till the bidding begins

 

Well. It is with great sorrow that I report that despite all the voltage I was able to stably drop, temps got as high as 85c at times.
I reflashed my stock vbios and once the fans are going good it stays at 81c after another repaste. (I tested the undervolt vbios after repaste too, for 85c)
So thats depressing. All those hours of tweaking the voltage table to the lowest possible voltage and it ends up running hotter than stock. I think its due to it drawing more amps? I dont know.
If you guys want I can post the undervolt vbios still.
The good news is my CPU runs at a solid 3.3 Ghz in games!
For the gpu its back to the drawing board.
EDIT: Running a stable 1288/3151 temps arent great though. Performance is though, stock I got 1127 score in Heaven, now I get 1399! Outruns a stock 970
Just wish cooling was better but what do you expect with such hardware in such little laptop.
I was almost able to get 1329 core, and I tried adding a smidge more voltage and it got slightly more stable but I dont want to add more and settled for 1288 at stock voltage. I can do 1304 core if I dont boost memory, but the +650 memory clocks grants much more performance than a little more on the core.

 

3.3 GHz is great! do you have the v3 or the v4?

My v3 only gets to 2.6-2.8 GHz when gaming and even then it thermally throttles sometimes. Thats with -40mV. Bear in mind it is with stock TIM and I haven't cleaned dust out of the heatsink in over 6 months.

 

What's the best score you've achieved in Fire Strike with a stable mod?

 

What do you all use to monitor temps? Got my P35xV5 in finally and the Samsung 950 Pro installed as the boot.

 

Its the v3, thermally im solid but running 1300/3105 on the 980m results in major CPU power throttling with or without undervolting, down to about 2.4 ghz. If you repaste I recomend either GC extreme, or if youre brave, liquid ultra (considering it).

The only benchmark program I have is Heaven 4.0, Ill download firestrike though. Heaven score is 1399 maxed out 1080p as per regulations. http://www.techpowerup.com/forums/threads/unigine-heaven-4-0-benchmark-scores.198888/ Right there with a lot of GTX 970 and R9 290x

I use a mix of GPU Z, Intel XTU and EVGA Precision X.
I really need to track down that Power supply dang it!

 

Ok im getting 8832 in firestrike with 1278/3105. BUT, I'm getting power throttling like crazy on my poor CPU down to 1.8Ghz. Even the GPU power throttles down to 300, 800 or 1000 in the really demanding scenes.
Ive contacted Aorus to hopefully buy a real power brick from them. My CPU does not thermally throttle at any point with stock or OC GPU clocks and my GPU is maxing out at 83c but when the fans kick in it will lower it to 80-82c after yet another repaste.
Its worth noting that in Heaven 4.0 nothing throttles, so its clear Fire Strike puts a much higher load.
In the mean time I may set the GPU OC vBios power table to stock levels since I dont have the watts to push more anyway. Atleast the CPU shouldnt throttle then.

 

Yes, I'm 100% certain the Geforce Experience is gone.

I let the windows updates (apparently it only apply the updates when specifically restart the computer, not when it's shut down), did a fresh installation of the latest Nvidia driver with DDU (restart in safe mode and everything), the problem still persist.

At this rate, I'll have to either use my gaming laptop without the 980m, or revert back to 8.1 and lose everything...

Sent from my D6603 using Tapatalk

 

If it helps to compare notes...
Checking back to when I last ran that benchmark, my best result was 9075 with +135/+575, -50mV undervolt. (repaste liquid ultra). No change to power brick.
For the Heaven benchmark at the same settings the best score was 1674, although that was before I did the win10 upgrade, so I probably need to ru-run that one!

 

I guess I'll post my results in here since there isn't a V5 thread, nor that much activity for the V5. My firestrike is at P9135. CPU gets up to 85-87, GPU to 75. That's with +135/+241 using MSI Afterburner. Stock CPU Clocks after my repaste with IC Diamond 7.

 

Wondering how you guys have such good results! Quite impressive indeed.
I have the Aorus 200w AC brick and Liquid Ultra on the way. Things are about to get interesting...

 

Ok, I think I solved my "TDR failure BSoD" problem after upgrade from 8.1 to 10:

I've tried to repair the system using Windows 10 installation medium's (iso on usb) "install with latest update" option, not working;

Then I change the fan setting to "normal" instead of "quiet" in Smart Manager because the back of laptop (where the heatpipes are) is really hot, the BSoD might have been cause by overheating...lasted for a day before the same BSoD occurred again;

Then a thread in Microsoft's forum caught my attention: one of it's replies suggest to go to "Troubleshooting" in the control panel (press "Start", type in "troubleshooting"), click "View All" on the left, follow by "Hardware & Device". The system then told me it has detected and fixed the hardware issues. This lasted little longer than a day.

The another reply said something interesting: the BSoD may have been caused by the changing power state in the dGPU (in both AMD and Nvidia) when the computer goes in and out of idle state. So I change the power plan back to "Balanced" (previously on "High Performance") and the prefered graphics processor in Nvidia control panel to "Auto-select".

It has been running for 3 days now without graphic driver crashing, no freezes (file explorer closes itself from time to time - once 2 days or so), no BSoD.

I have not got the chance to testing with intense gaming, but I can open and close a game with no problem, and I can't subject to more graphic-intense testings until after coursework deadlines.

Regards.

Sent from my D6603 using Tapatalk

 

@DelFang Glad things have improved for you. Having not had any issues for many months I found, to my dismay, that BSOD's were not a thing of the past for my machine... power state failure - so I think you're on to something there, I was foolish enough to invoke a windows upgrade whilst the HDMI out was being used, the damn thing decided to blue screen mid-way through the process, completely screwing my windows installation. Ended up having to restore a backup image. Most annoying!

@Key7795 Yes, will be interesting to see how you get on - am using a V3, I'm interested to see how the later versions perform.

 

I can't seem to get smart manager to disable winlock. I click on it and it grays out like its disabled until I release the mouse button at which point is goes back to on. It's driving me nuts not being able to use the windows key. Anyone got any ideas?

 

Guys... I repasted with liquid ultra and now I have Very Bad artifacting or whatever its called. Flickering lines all across the screen like crazy when I game.
I heeded the warnings and I didnt see any LU get where it shouldnt be but something has gone horribly wrong.
I tried cleaning it all off and putting GC extreme back but the problem remains and makes games unplayable.
So my GPU is fried?
Should I try and send it back to gigabyte for warranty?
What would I be able to sell it for on ebay? (of course explaining whats going on)
This is so terrible...
EDIT just sent an email to newegg which is where I bought it, hope they can help. Customer chat is down I assume because its sunday but Ill try that in the morning.
EDIT2 just loaded the undervolt vbios and that seems to have fixed it! Atleast so far. Still gonna see about warranty.

 

@Mr. Fox is the most experienced person I know in applying Liquid Ultra, and he may know if the symptoms you're seeing are a common side effect of getting a bead of Liquid Ultra somewhere on the MXM board where it shouldn't be. I would have said to remove the GPU from the laptop & then examine it closely everywhere with a magnifying glass to spot beads of Liquid Ultra & remove them, but I think your GPU is soldered onto the motherboard - guess you can still try the magnifying glass thing. Not sure of the best method of removing escaped minute beads of Liquid Ultra either, again Mr Fox might have some advice. I'd also check your heatsink is mounting flat and properly on all your VRAM chips (& other chips).

EDIT: ah, didn't read your post properly, you say it's fixed with the undervolted vBIOS - maybe your heatsink isn't mounting properly and the undervolted vBIOS is creating less heat which is getting round the problem. Or maybe you could try flashing back your previous vBIOS again (if you've got a fresh copy) - perhaps something got corrupted in it.

 

Well, I don't know what happens when people put it where it does not belong because I have never done it on purpose or by accident. I'm pretty sure it WON'T end well if you do, and I think it would be a lot worse than what the problem is here. @Key7795 it sounds like you've got something going on besides Liquid Ultra being on something it should not be on. If you were careful and did noy spill or squirt CLU on something other than the GPU die and heat sink copper plate, there is nothing to be concerned about. What you are describing sounds like a defective GPU that may be covered under warranty. The folks at Gigabyte may not approve of customers taking apart their machines. I'm not sure what their policy is on that.

It should never be necessary to undervolt a GPU to stop artifacts. That fact that it helps, as @Robbo99999 suggests, is a sure indication of a hardware problem. Your machine should work fine stock or overclocked, with stock voltage or extra voltage. It should not require abnormally low voltage to work right, so I think your GPU is headed for the last round-up. If it's a BGA system, sadly, that means super-expensive mobo replacement time. I hope they take care of it under warranty for you.

 

@Key7795 FWIW, I doubt you'll have any issues getting it repaired, provided you haven't removed or disturbed the warranty stickers inside. The GPU on mine went tits up and they didn't bat an eyelash at repairing it. Their process was a little tedious and they don't really keep you very well updated unless you have an advocate from their tech support side, but for what they did, I think it went fairly quick. End to end, it took me about 2.5 weeks from the day I sent it in until I had it back in my hands. I had to pay to ship it to them (with insurance it was a bit pricey), but they shipped it back on their dime. I had mine open several times replacing the WLAN card, drives, and memory so i don't think they care about that part, but like I said, if the warranty stickers are disturbed, then that may be a different thing. I wouldn't volunteer that you repasted it tho...just sayin...lol.

 

Thank you for the reply Mr. Fox.
I am curious if it may be because of the thermal pads? I did have a couple get slightly messed up, but I fixed them as well as I could.
I'm going to contact Gigabyte and see what they say about fixing it. I am reallyy not looking forward to not having my laptop as I am currently disabled and dont have a lot to do beside it.
Also it is worth noting that the artifacting starts once the GPU gets around 79c, but as I said undervolt stops it and it gets up to 84-85c. However as you have said it should not require undervolting to function.

That is very useful information, thank you

 

SUCCESS!!!
Running standard volts I am getting no artifacts.
What I did was go get some copper sheet, and cut pieces to size and replaced each thermal pad with them (the pads were rough)
I feel accomplished
My 200w Brick came today too so testing will begin. Currently running a solid 3.3ghz in BF4

 

Well, congrats, but perhaps flashing back to a non-corrupted version of the vBIOS solved it. Also, isn't it very difficult to replace the thermal pads with copper, I wouldn't do it for the following reasons:
1) The copper plates are conductive & if they move out of place they could short out the board.
2) Copper plates aren't compressible, so I don't know how you can guarantee that all the chips are contacted properly - it would have to be done more accurately than mm perfect, plus you would probably need to use thermal paste on them to get a good interface. If it's not done mm perfect you either get poor contact or place uneven stresses on the board which could cause damage.

I see lots of possible problems with replacing the thermal pads with copper sheets, and I wouldn't recommend anyone to try it. But, it is good that your machine is working properly now.

 

Thanks, however I was using the stock VBios when the problem struck.
It was time consuming yes, but this was a diagnostic sort of thing. I read that VRam overheating could cause this so I thought Id try and see if it helped or even fixed it.
I used a small bit of superglue between the copper pieces and heatsink to hold them in place.
This is the biggest issue and I think what happened last night. After atleast an hour of BF4 the problem struck again though milder.
The copper is 0.4mm thick so I think I may be able to add another piece to each along with thermal paste. I would like to think that properly done, copper would be better than thermal padding but I cannot know.
I flattened out the pieces before application so hopefully a bit more thickness will do it with paste.
What is the Thermal Pads thickness? It looks about 1 MM but accounting for squish maybe the 0.80 of copper will work.
Thanks

 

I think it would be best to not use super glue at all - use thermal paste between copper and the plates if must. It'll be tricky to get just right copper piece thickness - too thin, then you'll have to fill the gaps with thermal paste, which results in less efficient thermal conductivity; too thick, then it'll risk physical damage to the CPU/GPU modules and the motherboard.

Alternatively, have you try to buy some thermal tab/tape from places like ebay?

Regards.


Sent from my D6603 using Tapatalk

 

You guys are right, so I ordered some 5mw 1mm thermal pads from amazon. They'll be here tomorrow.
Heres hoping!

 

Definitely don't use superglue between the heatsink & the copper shims - superglue will prevent good heat transfer, so you're not getting any benefit from that mod. I still wouldn't bother with the copper shims for all the reasons I mentioned in my previous post.

Good, I think that's better than using the copper shims - I'd remove all your shims & just use correct thickness thermal pads. You had the problem occur last night again - sounds like you have an issue with that GPU, I'd just get it fixed on warranty like some others have mentioned, so I'd put it all back to stock & then send it back in for warranty.

 

Fresh thermal pads and all's well. Crysis 3, BF4, Fallout 4, Rise of the Tomb Raider. No artifacts. 3k and cranked up.
Now then, Mr Fox would you be so kind as to share your experience and results with liquid ultra?
I need to use the thinnest layer possible? How thin is too thin?
What should it look like?
Best way to remove the residue?
Only on die, or on the copper heatsink as well?
Only on top, or the sides of the die as well?
Sorry for all the questions, but I am every curious and hope to learn from the experienced.
Thank you.
PS nobody use copper or superglue... very poor contact and heat transfer as the good folks here have said.