Acer Predator (Vega 56+Ryzen 2) Helios 500

I took the machine to one of the PC World shops nearby (as I bought it online from their website).
Since its still under warranty (its only 6 moths old), Acer customer service told me that I should take it to PC World because I bought it off them, so I need to deal with them.
The technical support guy at PC World was much better to talk to though as the line was also clearer and I could actually understand what he said.

I was told that for repairs like this they need to send it off to a proper lab for reflashing, and that I could have it back in a few days, or up a week (so, next Tuesday) - which was confirmed more or less by the guy at PC World... but I won't hold my breadth in case they need to order a replacement motherboard - though realistically, this shouldn't happen because its just a faulty BIOS updata.

Obviously, it would be great if I can get the machine back in 7 days, but we'll see what happens.

What I don't understand is this:
Why did Acer release this BIOS update in the first place?
Aren't they supposed to check if the update will actually WORK?
Plus, why doesn't the BIOS/UEFI have a backup section of a latest functional version in cases of faulty BIOS updates such as this one so that recovery can be done at home automatically?
Its' 2019, you'd think the laptop would have a fail-safe mechanism in case the BIOS update isn't working as it should - and its not like its impossible to implement.

Anyway, my (sadly, painfully learned) advice would be, STAY AWAY from 1.09 BIOS update.

Did anyone else try updating the BIOS to 1.09 only for the machine not post?

Btw, all other BIOS releases including 1.08 version seem to work fine.

 

Did you overclocks the machine? If yes, did you reset the overclocks before flashing the BIOS?

My point is, if you OC and flash BIOS chance for bricking increases greatly, I done this mistake in the past thankfully only on desktop though.

Sent fra min SM-G970F via Tapatalk

 

I never touched the clocks.
The machine was/is running on stock clocks on both the CPU and GPU.

 

Something interesting.
Someone on Acer's forums mentioned that they were able to update their bios to 1.09 version, and they hadn't mentioned anything about it bricking their BIOS.

So... what could have gone wrong?
I may have had the installation of the newest chipset drivers in the que, but I thought that was already installed BEFORE I started updating the BIOS (as I recall the machine rebooting prior to me initiating the BIOS update).

If it wasn't, could that have bricked the BIOS?

Or could something else have happened?
I mean, if PC World manages to reflash the BIOS, they'll probably use 1.06 version that came with the unit originally...
and now I'm a bit weary of updating the BIOS to ANY version, even though the 1.07 and 1.08 updates went through fine before.
Plus, there's the whole issue of Acer releasing a BIOS 1.07 to ensure functionality with Windows 10... so for safety reasons, I kinda do need at least 1.08v.

Is there a safer way to update the BIOS per chance or a way to recover it?
Like I said, I did what I could to minimize errors, and the laptop was plugged into the AC, as well as didn't interrupt the update... so I don't understand.

Could the UK website have a corrupt BIOS file?

P.S. I sent an email to Acer about this and asked them to check their UK files for potential corruption.

 

Chipset drivers are used by Windows after given control by the BIOS, they could cause trouble booting but wouldn't stop POST.
There's no reason to think PC World would be incompetent and flash an old BIOS. Well I sure hope they know what they are doing.
The BIOS file is cryptographically signed and verified before flashing so no that shouldn't be possible. Verified that the filed is identical to the one I used without any problem.

This is the first computer I've heard of where reading the ROM chip with a hardware device by someone experienced have caused a failure to POST, that's the reason I'd like to see the schematics before trying some BIOS hacking. Could be some strange thing going on...

 

Well, its good to know that the installation of chipset drivers (even if it was in the que) would not prevent POST (Which DID occur to me since the chipset drivers are basically software that should technically load up with Windows - as far as I know, they make no change to the BIOS).

As for PC World installing an old BIOS... think of it like this:
They will know that a new BIOS install failed and will probably not mess about with that. Their job is to restore it to a functional state.
I may be able to contact them via phone and ask them to make sure to try installing the 1.09v BIOS, and if that fails, at least 1.08v.

Every time I had my GL702ZC (Asus) laptop in for RMA, the motherboard was replaced every time and it ended up with a stock BIOS loaded up (Asus never took the time to upgrade the BIOS to the latest version, I had to do so myself - but it wasn't apparently an issue to wipe the SSD and HDD and do a clean Windows install).

I don't know what kind of procedure PC World will implement, or if they might end up giving it to an Acer authorized lab to re-flashing... but I'm thinking they will simply try returning the laptop to STOCK state after a re-flash (though I hope they don't)...
And to be quite fair, I told the guy at the shop they don't need access to my Windows, or HDD as the HDD is completely separate of the OS. He said they need access to everything for potential tests.. I kept insisting that since the HDD doesn't contain the OS, there is no reason for them to mess about with the drives... in fact, since this is a BIOS issue, its unlikely the drives could have caused this problem, and even if they did, what good would do for them to wipe everything clean?
If the HDD is damaged, the damage won't go away with a data wipe - that's the thing I never understood.
Why wipe the separate HDD which has no OS files on it in an attempt to 'solve a problem'?
What problem does it solve?

Another Helios 500 Ryzen/Vega user managed to upgrade the BIOS to 1.09v without issues.
They suggested that the file I downloaded off Acer UK website might have been corrupted.
I suppose its possible.
I sent an email to Acer about this issue detailing what happened and that I installed the proper BIOS under 'ideal conditions' and asked them to examine their BIOS file on the UK website to make sure its not corrupted and that its functioning properly (otherwise, other users could experience the same problem),

I also asked them if it may be possible to implement a BIOS recovery method which could be done at home by users themselves in case of a failed BIOS update considering its 2019 and that such measures should really exist.

But since you mentioned that a failure to POST occurred with someone trying to read the ROM chip with a hardware device... yes, it looks like an odd thing to happen.
Unless the BIOS encryption was designed to do that... it may have been intentionally done this way by Acer to prevent tampering... which is quite odd to put such a large effort into an all AMD device.

Its like going out of their way to prevent people from hacking/unlocking the BIOS so they could even upgrade the thing with Zen 2 or Zen 3.

So, I take it you managed to upgrade to 1.09v as well without a problem?
If so, could you email me the file you used?
I want to see if I actually downloaded the correct one (but for the life of me, I cannot understand how I could make such an error because I specifically made sure the BIOS I downloaded was for this machine - and you'd think if you tried flashing a BIOS for a different motherboard, the program would say its impossible because the motherboard or anything else doesn't match).

 

I installed 1.09v and it works without problems.
I downloaded form here: https://www.acer.com/ac/en/SI/content/drivers

 

That's weird.
The file seems identical to what I installed.

 

@Deks
I downloaded the file from the UK site and compared (using windiff) with the one I used and they are 100% identical. Can't understand what went wrong with your machine, very strange.

 

I did have an extra 512GB SSD installed (and switched it to the first slot, and put the original 256GB SSD into the second slot), along with a 16GB RAM stick.

I don't see how a simple hw change like that could possibly affect the BIOS though as I previously updated the BIOS twice after I changed the hw.

 

Everytime check first if you have the latest chipset (b450) drivers then install new bios. I belive you forgot about that or samething go wrong with chipset drivers (probaly messing with old acer gpu drivers which have sonething in package i belive) .

 

I had the latest chipset drivers installed before the bios. As for gpu drivers, those were the 19.5.2 drivers from AMD website.
But, both the chipset and gpu drivers are independent of the bios. So I don't understand what could have messed up the BIOS update

 

Wait that shouldn't be possible, wasn't the original 256GB an NVMe drive?

No that shouldn't matter at all. Hope you get it back quickly.

 

I don't know whether the original 256GB SSD was NVMe or not. Sufficed to say, I put my 500GB SSD into the first slot, and moved the 256GB into the second slot (because the laptop kept booting Windows automatically off the first slot and there was no way to change it permanently inside the BIOS - every time a restart of the machine occurred, the laptop would automatically switch the boot to slot 1).

Anyway, the 256GB SSD (original) is in the second slot and is detected by Windows and works... and my 500GB SSD is in slot 1.

I think both are NVMe... so the SSD would probably work in slot 2 even if the slot was not NVMe (not too sure about those standards though).

I also spoke with PC Specialist today and asked them to try and notify the engineers to install the latest 1.09 bios ig they can as others managed to do so, but if they cannot, they should install the version 1.08 at least as the 1.06 is too old and doesn't have security/stability patches of newer versions. - the guy said he will pass it on to them.

Oh and I asked them if possible to not touch the HDD.
I don't mind if they need to reset Windows (on the SSD), but that the HDD has nothing to do with it and as such shouldn't be touched or deleted, because even if there was a fault with it, wiping the data off it wouldn't do any good anyway (With which the guy agreed - but did say they cannot technically guarantee that).

 

Always take out harddrives etc that the OS does not occupy and always expect them to reinstall the machine when you send stuff like that in for repair.
Even if you tell them it usually does not prevent them doing it.
Really annoying but just how it is in my long experience dealing with laptop services.

Sent fra min SM-G970F via Tapatalk

 

Probably should have taken out the HDD... but I was exhausted/tired by that point and what's done is done... but most of it I have backed up anyway (or at least the stuff I REALLY need - everything else is minor and I can download from the Internet again - such as few programs).

Reinstall/reset of the machine (from SSD) is fine... I told them they don't have to touch the HDD as the OS is not on it.
I don't understand OEM fascination with doing that.
It's utterly pointless and doesn't do ANYTHING to solve a potential mechanical problem (if there is one to begin with) which is something I told them.

And besides, if they have time to do that, you'd think they would try to reflash the latest bios themselves, and if that doesn't work, then test to see why isn't it allowing them to do that and address that problem... or at the very least flash the immediate version before that (which ran fine).

 

17.12 have something related to chipset driver inside


19.5.2

 

And?
I did have Acer's original drivers installed, but I express updated them t 19.5.2 in order to preserve Freesync functionality.
If 17.12 had something to do with the chipset, it could have been Freesync related.... because if you just end up installing 19.5.2 drivers themselves, you lose Freesync.

Anyway, I don't think this would have messed up the BIOS as chipset drivers are OS level, and they don't interact with the BIOS itself.... not to the level where they would corrupt the BIOS.

 

Wierd because i put 1tb nvme inside, 32gb ram and even 2700x and i installed my bios without problems and i do everthing on full oced machine (ram timings, cpu, gpu and hbm2 ram timings).

 

I also did update my system with 1.09 bios and everything went well. Hopefully you get back your system faster.

A little off topic though, has anyone tried the new driver?
https://wccftech.com/amd-delivers-radeon-image-sharpening-to-vega-owners-with-adrenalin-19-9-3/

 

I tried the new driver, installed on top of 19.9.2 which worked fine other than the sleep / resume screen static issue is still present. The "Radeon Image Sharpening" option is available in the display settings.

The 1.09 bios update worked fine for me, but after spending some time reading the Acer community forums it seems like very many models (not just ours) have a chance of bricking during an update. My only thought is maybe the Insyde bios update util has some strange bug related to the ITE update util (for the embedded controller) which can randomly cause the update to fail.

 

Interesting.
Could you send an email about this to Acer directly?
They should be notified if there's a chance their installer could cause a random bug resulting in a corrupt BIOS installation (and therefore a bricked machine).

They should release a NEW version of the BIOS (such as 1.10) which has these issues resolved and potentially includes LATEST AGESA update from AMD (the upcoming one with 100 new things/features - which also might give us Zen 2 since those are microcode updates).

And if AMD is releasing security fixes as well, Acer should be obligated to tailor that update for our machines too.

 

Well it seems the repair team said they need to order a new motherboard.
After assessing it for a few days they determined that it wouldn't POST (which is exactly what I told them initially and was determined by the guy who wrote the note for the team when he tried turning the laptop on).

And now they need to order a new motherboard... Which in all probability will arrive with the outdated stock bios from Acer.

I asked them why did they need to order the new motherboard when I was told they would only need to do a BIOS reflash?
The woman I spoke with said she doesn't know and that they ordered a new motherboard (which was a bit more information than what I was given in a text).

Great. So not only do I need to wait for an extra week or more to get the laptop back (she didn't know the time it would take to get the laptop back though or when the part will arrive) I will also in all probability get an outdated BIOS (because I doubt that Acer ships replacement motherboards with an updated BIOS anyway).

Just what I was afraid would happen... and given this fiasco with a failed BIOS update, I'm now reluctant to do ANY BIOS updates (even 1.08 which previously worked) for fear the random error could kick in again and brick the machine all over again.

This is unacceptable. A bad BIOS flash shouldn't need a motherboard replacement in 2019 (that comes with an outdated BIOS to boot).
It should have auto-recovery mechanism in place for situations like this.

Plus I needed to make sure they knew that I wanted the extra ram stick and ssd that were in the laptop reinstalled into the laptop (so she added this information for them).

P.S. When I asked her to make sure if she can relay the message to the repair team to update the BIOS to the latest version, or if not, that they at least put 1.08v in there which I previously had, she said they cannot do that, that they are basically not authorized to do BIOS reprogramming.

 

Doesn't surprise me they ended up ordering a new motherboard. In situations where the system won't post and the process defined in the service manual doesn't help, the next step is just order a replacement motherboard.

A few business-grade laptops do have better bios recovery systems but it seems like for most gaming laptops this kind of thing is an after thought. I wouldn't worry about the random failure chance, just do the update to 1.09 and it will most likely work.

 

'Most likely' is not good enough I'm afraid as I cannot afford to risk another BIOS update failure and being left without a machine to work on for extra 3 weeks or a month considering I have a college course for which I need the laptop (content creation).

If the BIOS failed to update for whatever reason on the original motherboard, I have the same chance of risking a bricked BIOS yet again... and I cannot take that chance.
I might consider updating to 1.08v (considering it worked in the past)... but, again, given the possibility of a BIOS failure, I don't know if I'm willing to take even that chance (even though the 1.07 and 1.08 contain some updates which ensure the hw to work with Windows properly - in which case, I may need those).

It creates ridiculous problems for me in regards to waiting times for the machine to be repaired when I have deadlines to work with and now extra worries whether the laptop will work properly with the 1.06 stock BIOS (unless by chance the mobo replacement I get has latest BIOS installed... which in all likelihood doesn't).

It's not even been a week and the new motherboard wasn't even ordered yet... meaning that they will probably order it on Monday (if not today), and depending on how fast it will get to them, and them installing and checking everything etc... it will just be a huge time issue which will probably take another 2 weeks (at least) until I can get my machine back.

Are there any steps I can take to try and do a BIOS recovery myself in case of a BIOS update failure repeating itself?
Is there anything I may have missed when I was creating a flash drive with the BIOS and pressed FN+ESC upon powering it up which prevented the stick from activating?
I'm not even sure if I know how to make a properly booting flash drive... usually, those are created when a bootable OS flash stick is made (ISO file), but not in the case of transferring extracted BIOS from a file to the stick itself (or at least, the instructions said nothing on how to make a bootable flash drive for that particular situation).

 

Always a chance for bricked machines due firmware updates. A known problem on Dell machines etc.

You should get you an hardware flasher. Then you could reflash the bios chips yourself if similar happens later.

 

Possibly. Only problem is, I don't know where to begin with getting a hw flasher.
Plus there's an extra problem with the fact someone already tried reading this machine bios in an attempt to extract it, resulting in BIOS failure and machine refusing to Post.

So, not sure if a hw flasher would even work.

Plus, bricking the BIOS shouldn't be a problem today (in 2019).
All mobos should come with contingencies for a potential bios update failure so the previous functional version of the BIOS would be loaded/installed automatically in case the new one is corrupted either due to a bad installer code, power cut (unlikely as it may be), etc.
In case you try to install a BIOS even for a completely different mobo (unlikely as that may be), the installer shouldn't proceed anyway, warning you that the BIOS you want to install is NOT for the said computer... and besides, the utility should have the ability to check BIOS code and ensure or let you know if it will work on the motherboard in the first place.

It cannot be that big of a problem to implement with all automation we have at our disposal. If anything, the motherboard manufacturers are lazy for not implementing this feature into all motherboards for well over 10 years now... and its just creating unnecessary problems.

 

One inexpensive programmer that's often sold together with a SOIC 8 clip (required for programming soldered chips) is the CH341A, look on ebay or similar.

That's the reason I'm wary of using mine.

Well they don't want to pay for something generally unnecessary. Most users never flash their BIOS and of those that do almost nobody have a problem. Penny pinching...
We should be glad they at least support a protected boot block so that a recovery is (sometimes) possible.

 

We do live in a day and age of upgradeable CPU's on the same chipset/slot and most people these days DO upgrade their BIOS, so microcode updates should be standard part of BIOS updates, along with security, etc. Security it quite important actually, given that AMD has most of it taken cared of on a hw level, but an occasional patch may be required... plus, if AMD releases a massive microcode update which provides various tune-ups and fixes, then OEM's are obligated to release them for laptops just as they do for desktops and ENSURE that the BIOS update will work, and/or have contingencies for failed BIOS updates.

They support a protected boot block?
Really?
I had issues with my Helios 500 in the sense it couldn't detect a bootable ISO on a USB flash drive when I tried doing a clean Windows install (even though it was set to UEFI beforehand).
For some inexplicable reason, the machine never detected the USB flash drive in the first place.
I had to mess about with unfamiliar Rufus settings for UEFI and another thing for the machine to auto-detect it... something I NEVER had to do on any other machine. One UEFI setup on Rufus doesn't work, while the other one (can't remember which) does.

In fact, this Helios 500 doesn't give me the option in BIOS to boot from a bootable flash drive first at all (it as some kind of weird auto detect for bootable USB, but the UBS flash drive needs to be configured in a ridiculous manner so the machine would recognize it in the first place, otherwise, you can't boot from a USB key at all), and the BIOS named both SSD's exactly the same, making it impossible to distinguish between the two, and if you set to boot from say slot 2, once you restarted windows, BIOS would revert to booting from Slot 1 automatically.

Whoever programmed this Acer BIOS did so poorly.
We have 0 practical options to choose from, you cannot set to boot from the USB drive first at all (the drive itself needs to be weirdly configured for that), and it doesn't even detect SSD drives with their distinct names.
Stupid, stupid, stupid.

 

How do you know/tell if Freesync is working or not? Is it simply greyed out or does it say it's on when it's off?

I'm running the latest GPU drivers from AMD but genuinely can't remember if I ever did a clean install along the way lol.

 

What you can do to 'ensure' Freesync functionality is to install Acer's original GPU drivers first, then Express install the LATEST drivers on top of those (or use 19.5.2 drivers if you wish to avoid the static display issue appearing when you bring the laptop out of sleep mode).

If Freesync is greyed out in drivers, then it means you tried clean installing the GPU drivers from AMD website directly... and in all likelihood, it won't work like that.

You can also test whether Freesync works or not by using AMD's own windmill demo.

The reason why you 'generally' need Acer's drivers first is because they customized this system. Since laptop displays don't have their own personal ID under Device Manager or drivers, they are recognized as 'generic display', and usually, its the OEM that configures that display.

Acer and other OEM's should do more to support their laptops a lot better than before.They can't get away from supporting the units 6 months after release and then dropping all support for it.

 

In case anyone is curious, I tried the 19.10.1 drivers released yesterday, and they still didn't fix the screen distortion on resume issue.

 

Hey guys, having a bit of an odd issue with my Helios, but firstly I thought I'd add some potentially useful info for the guys experiencing the static/distorted display issue. It's not actually anything to do with sleep/resume so if you're telling AMD that they may have trouble locking it down. It's an issue with anything that turns off the power off to the screen, I don't even have sleep enabled but I can reproduce it easily as when I use my Helios at home it sits on the desk with the lid shut and powers a triple monitor setup, if I open the screen then I get the static/distorted issue too. With the older pre-static drivers I could open it and switch to quad screen fine.

Just figured the additional info/insight may be of some use.

Anyway my Issue:

I recently bought a new LG GL27850-B monitor to replace the older Dell UltraSharp U2913WM as the centre monitor in my triple screen setup, it's pretty good however I'm having an odd issue, if I connect it via DisplayPort to the back of the Helios then I get a flickering issue when running at 144Hz with or without Adaptive Sync enabled (looks almost like the backlight is pulsing, sort of like how screens appear flickering in a low FPS video), this also happens if I connect it to one of the other DisplayPort enabled connectors on the Helios, but if I connect it to the HDMI port on the back then it can only do 100Hz+Freesync but it's fine. In addition if I change the aspect setting on the monitor from original or full to "just scan" it works, but that's not available with Adaptive Sync enabled.

In addition when it's being wierd the OSD on the monitor shows it flicking between 72Hz/144Hz, so something is def up with the signal going to the monitor, however the weird thing is if I connect the monitor to my old Dell Precision M6600 laptop then it works find at 144Hz with Adaptive Sync enabled, so it's not the cable either. I did think that the drivers for the Dell would be a bit more mature as it's a 6-8 year old laptop, but then the Vega in the Helios is a couple of gens out at this point so shouldn't be driver related /shrug.

So yeah, anyone else seen this when connecting a 144Hz Adaptive Sync display over DP? (I know it's hardly something everyone does but figured worth asking here).

 

Still no word from acer on an updated bios for this machine that will processor upgrades besides 2700x?

 

I also noticed (before the BIOS update bricked my mobo) that any kind of shut down of display results in a static noise displaying on the screen.

Given that the laptop is effectively a desktop replacement with desktop grade components inside, there's no real reason why this should be happening.
For a moment I thought it might be due to the premise that the driver might be trying to switch the display to the IGP... but since IGP's generally don't exist on the desktop (at least not for 2700), it stands to reason that's not the problem.

So, what could be causing the display to act out like that with drivers post 19.5.2?
We know that this started happening when Navi was released... hence, it could be down to differences in uArchs primarily reserved for Navi that might be getting activated on Vega for whatever reason.

What I'm curious about though is why aren't any desktop users experiencing same problems?
Some reported other issues with latest drivers (apparently), but none specifically mentioned static display problem on the desktop.

 

Has anyone been in contact with amd or acer about this static screen issue ?

 

It could possibly be something to do with the screen being eDP whereas monitors connected to a desktop Vega card are going to be DP/HDMI/DVI, or it could be a quirk of the custom Vega design Acer used for the Helios. However it doesn't affect any of my monitors just the built in display so it wouldn't present itself to desktop users anyway so they may even have the issue too but never encounter it.

 

Only issues I run into with desktop Vega is when you put the desktop into hibernation the GPU driver sometimes crashes, sometimes your global settings do not apply at all in games.
FreeSync not showing up at times requires driver reinstallation for it to detect it.
Or driver crashing from simply the PC turning off screen and sometimes it refuses to turn off the display at all and the driver failing to clear the GPU memory cache requiring a reboot.

I run latest drivers and noticed these issues less frequent but still persistent.
They happen regardless whether I run an overclocks or not.
I have tested both DP and HDMI output but I usually only use DP.
Polaris has way less issues than this.

Sent fra min SM-G970F via Tapatalk

 

It seems AMD has some work to do on Vega a bit more.
Drivers worked fine up to Navi's release, but then started affecting Vega negatively.
Polaris probably works better as its a larger bottom line so its getting more attention.
Still, I suggest everyone in this thread submits a technical support ticket to AMD and Acer about this display noise issue.
The more people report it, the better chances there are they might fix it (even though they should have done it by now).

 

This static screen issue seems to exist on linux also i noticed.

https://forum.manjaro.org/t/suspend...e-screen-acer-predator-helios-500-ryzen/77603

 

I'm a common poster on the Acer Forums, especially on the BIOS topic.

The "static" on wake-up isn't limited to the internal display either, I use mine with an external display, and it won't even engage the display. Opening up the laptop, shows the static. Some combination of AMD's drivers, and a Windows update, caused this to happen, as it began after a windows update, not a driver update for me. Since then, I just leave the machine on as I keep it closed on a vertical laptop stand and reaching back to pull it off the stand, and turn it on, is inconvenient.

The Vega 56 die is desktop-grade. I've used OverdriveNTool to play with its powerplay tables, and got it using 180w at 1,600mhz (ish) on the boost clock, essentially a slight overclock bump that you'd get on a desktop GPU. No power, stability, thermal, or noise issues whatsoever.

However I think the VBIOS is slightly different, which could be causing the resume from sleep issues people are having. If this wasn't my only machine, I'd attempt to flash a desktop VBIOS to the machine to see if that would resolve it. I already tested that the 330w power brick is more than sufficient to run the Vega 56 at desktop level with the powerplay table edit, so as long as its apples to apples (since the V56 in this is essentially on a PCI-E 16x bus), it /could/ work. But again, this is my primary computer so... kinda chicken to try.

I also had a weird flickering issue that was mentioned by ubersonic earlier in this topic, I had a driver crash, and the monitor began that weird low refresh flicker. AMD Control panel showed my monitor only capable of outputting 6-bit color at that point (where its native is 10-bit) and Freesync disabled. I reinstalled the drivers and the monitor has slowly recovered, however text is still not right, and I've been messing with Windows text configuration tool thing to try to get it right. I at first thought my 34" monitor bit the dust due to my prior powerplay table modding, but nope, just a weird feedback from a driver crash.

I highly doubt Acer will support this machine further, just like ASUS abandoned the GL702ZC (Ryzen 7 1700, RX 580), theres no money in continuing support, we're having a hard time getting any official response from the Acer forums on any of my above mentioned topics.

 

I would be up for doing that later as if it bricks the V56 I can always flash it back as my left/right monitors are on an external GPU. I doubt it would work TBH as Vega 56 Mobile is proberbly about as different from reference Vega 56 as a custom Sapphire Nitro is, if not more. However what worries me though is flashing the BIOS from a 210w TDP card to a 120w TDP card, especially as 210w is beyond the MXM spec.

 

This GPU is BGA anyway, not MXM, not to mention the more robust MXM slots in MSI and Clevo DTRs can handle well over 200W without external power and not have issues. If you’re gonna raise the TDP by that much I’d take a look at the GPU VRMs first and try to find out what is within spec, and keep them well-cooled.

 

I mailed amd for my own part. And i suppose its a good idea if others also start to contact both amd and acer to hopefully get a solution to this problem. Its a shame if they can not do anything because its a really nice laptop otherwise.

 

At +75% power target, it'll draw 210w. Complete system power draw under full synthetic load reaches 285w from the wall, still within the 330w power brick spec. Hits 78°C. I've also re-pasted and added new thermal pads to all the mosfets within a month of owning it because I never trust manufacturer applied cooling solutions, so ymmv on thermals.

The soldered on chip is on a PCI-E 16x bus, essentially just skipping the socket and electrically routing it directly to the CPU.

But yeah, as ubersonic points out... Every aftermarket variant desktop card has their own VBIOS, so flashing probably won't work, or be very unstable. I may use ATI flash to compare a desktop reference one to this BGA one to see where they differ. It doesn't make sense to me that this has wake from sleep issues where a desktop one won't because architecturally they're the same.

I'm also curious if Enduro works. If I put in let's say a 2200G, if I brick the dedicated flashing it, if it'll use the on-chip. But I'm probably well into the weeds on this idea.

 

Can you post 3d mark fire strike and time spy score? I would like to know if is worth it to go past 1500mhz.

 

Hi there, how did you change the stock voltage on this card? I tried modifying power tables or change it by wattman/afterburner but always revert to default. How did you able to increase clockspeed? Mine is not stable past 1456 mhz at stock voltage. thanks man

 

Spoiler: Before PPT Mod

Spoiler: After PPT Mod

1200 point increase in performance just in letting it use more juice.

To unlock voltage control, I used MSI afterburner, and I disabled ULPS, and enabled Allow Voltage Control and Allow Voltage monitoring. Due to the VRM limits, the voltage is locked at exactly 1v, or 1,000mv. A V-BIOS mod could push that up but with shared Mosfets between the GPU and CPU dies, I don't recommend it.

I changed the P6 and P7 states. P6 I changed to stock desktop Vega 56 boost clock of 1474mhz. P7 I set it at 1,600mhz. I set the overall allowable maximum to 1,800mhz, and the power limit to +50%. It seems to comfortably hang around 1,570 mhz under gaming load, at a reasonable fan speed and temperature limit. I did try for higher, with a 75% power target, but Assassins Creed Odyssey crashed the driver after about an hour of gameplay and backing it off to the numbers above did make it stable for hours. And gained nearly 8-10fps in game, which at 3440x1440, Ultra, going from 48fps to nearly 60 is night and day.

But just a disclaimer, your mileage may vary, and I'm sure differences in silicon quality, and board manufacturing quality will change your results. Powerplay table modding is relatively safe, as if you bork your settings, just use DDU to remove and reinstall your driver, but I don't know the longevity of the power delivery will be with it pulling 180w-220w (+50% to +75% respectively). I figured theres enough safeties in modern silicon these days that if it didn't like it, it would just not work. But I am well within the power limit of the adapter.

 

Voltage control not working and modifying power tables wont work in any case because is locked in bios.

Voltage control not working, i have better score and gpu runs cooler just with 1480mhz.

http://www.3dmark.com/fs/20699906

Anything beyond 900mv you need a custom bios or stock vega 56 bios.

 

I wasn't saying that it would have issues just pointing out that it would be out of spec for the GPUs MXM design.