• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

3070 Ti - Big performance issues after driver update

Is this something like an issue with the shunts on a regular desktop card?
Yeah it's the same thing, makes the firmware think it's constantly drawing some absurd amount of power and it downclocks to the lowest frequency.
 
Laptops are definitely not fine imho. Not going any further than the lack of options in the BIOS, options that are commonplace on a regular PC, that right there is a deal breaker for me.
You want to diagnose a video card on a PC? Remove the card, place it in another PC (or place a different video card in your own PC), see if it fixes the problem. On laptops, that's a no-go. And I could go on.

You're a power user, that's understandable. But 99% of people don't care, and to at least 0.5 of the remaining percentile, it's an acceptable tradeoff for mobility and/or cost.

Firstly, it's a gaming laptop, which is a problematic combination; I suspect there has to be something wrong with the power delivery circuit for the nv GPU. It's probably gotten too hot too many times over the years and degraded (designed with not enough spare amperage so to speak).

It's a 3070 Ti, I don't think it has "years" of mileage on it yet. Either way the culprit seems to be the processor that's issuing a BD PROCHOT signal, since it's bidirectional, it will throttle the GPU as well.

This could be:

1. Thermal issue (although it doesn't seem to be the case)
2. Power delivery issue (bad power brick, or damaged DC jack sensors), perhaps the aforementioned problem @Vya Domus mentioned
3. Assembly error (this would cause the laptop to sporadically malfunction as the heat makes the material contract or expand)

Either way this laptop needs to be professionally serviced if the OP does not know how to fix it themselves
 
It's a 3070 Ti, I don't think it has "years" of mileage on it yet. Either way the culprit seems to be the processor that's issuing a BD PROCHOT signal, since it's bidirectional, it will throttle the GPU as well.

This could be:

1. Thermal issue (although it doesn't seem to be the case)
2. Power delivery issue (bad power brick, or damaged DC jack sensors), perhaps the aforementioned problem @Vya Domus mentioned
3. Assembly error (this would cause the laptop to sporadically malfunction as the heat makes the material contract or expand)

Either way this laptop needs to be professionally serviced if the OP does not know how to fix it themselves
It's 1.5 years old, but I'm a casual gamer. On average it's probably 30 minutes a day.

I've been thinking of trying a new PSU. I just can't get one in Vietnam, so I would need to import one. But honestly if this would be the case then why does it work for 1.5 month without any issues.

The issues always start after updating GPU drivers. But downgrading does not make it go away. The only thing that seems to make it go away is a fresh Windows install.
 
It's 1.5 years old, but I'm a casual gamer. On average it's probably 30 minutes a day.

I've been thinking of trying a new PSU. I just can't get one in Vietnam, so I would need to import one. But honestly if this would be the case then why does it work for 1.5 month without any issues.

The issues always start after updating GPU drivers. But downgrading does not make it go away. The only thing that seems to make it go away is a fresh Windows install.
Honestly, this is a question for the manufacturer's support. But if they messed up something, they might not be able to admit it.
 
It's a 3070 Ti, I don't think it has "years" of mileage on it yet. Either way the culprit seems to be the processor that's issuing a BD PROCHOT signal, since it's bidirectional, it will throttle the GPU as well.

This could be:

1. Thermal issue (although it doesn't seem to be the case)
2. Power delivery issue (bad power brick, or damaged DC jack sensors), perhaps the aforementioned problem @Vya Domus mentioned
3. Assembly error (this would cause the laptop to sporadically malfunction as the heat makes the material contract or expand)

Either way this laptop needs to be professionally serviced if the OP does not know how to fix it themselves
I don't want it to be broken. @haste18 maybe try cleaning the heatsinks of the computer - that is right after the blower; in notebooks a "carpet" of dust tends to build up there.
 

Attachments

  • notebook_carpet_of_dust.png
    notebook_carpet_of_dust.png
    2.8 KB · Views: 75
I don't want it to be broken. @haste18 maybe try cleaning the heatsinks of the computer - that is right after the blower; in notebooks a "carpet" of dust tends to build up there.

I've done this before and they were full of dust, but it didn't make a difference.

An update on what happened in the last 24 hours:
I got an EC update from the laptop supplier. Flashed it, rebooted, but got back in 0.39 GHz, so no difference. He also wanted me to put BIOS settings back to default, which enabled MsHybrid instead of dGPU. There was an update for the Intel onboard card so I updated the driver. After reboot the screen started flickering (a lot) and I had to change it back to dGPU. Everything looked fine after that, so I decided to start a benchmark test with OCCT. All previous once in the last few days ended after 5 seconds with black screen. Now it all ran fine. Did multiple tests and decided to try to play Cyberpunk and this also ran fine for at least 30 minutes. Good FPS (80 - 90) and no issues.

After lunch I came back and again stuck at 0.39 GHz. I'm just baffled. How is this possible?
 
I've done this before and they were full of dust, but it didn't make a difference.

An update on what happened in the last 24 hours:
I got an EC update from the laptop supplier. Flashed it, rebooted, but got back in 0.39 GHz, so no difference. He also wanted me to put BIOS settings back to default, which enabled MsHybrid instead of dGPU. There was an update for the Intel onboard card so I updated the driver. After reboot the screen started flickering (a lot) and I had to change it back to dGPU. Everything looked fine after that, so I decided to start a benchmark test with OCCT. All previous once in the last few days ended after 5 seconds with black screen. Now it all ran fine. Did multiple tests and decided to try to play Cyberpunk and this also ran fine for at least 30 minutes. Good FPS (80 - 90) and no issues.

After lunch I came back and again stuck at 0.39 GHz. I'm just baffled. How is this possible?
If you got corrupted image using the IGP, I'd say you have memory errors. Try to put Memtest on a USB stick, boot and run it from there.

Edit: Simpler yet, try one stick at a time. They usually don't go bad in pairs.
 
Last edited:
If you got corrupted image using the IGP, I'd say you have memory errors. Try to put Memtest on a USB stick, boot and run it from there.

Edit: Simpler yet, try one stick at a time. They usually don't go bad in pairs.
So I did a test for almost 11 hours and nothing wrong. You think I should still do a test with one each?
1711112549624.jpeg
 
So I did a test for almost 11 hours and nothing wrong. You think I should still do a test with one each?
View attachment 340147
Yes, but not memtest. Just use one stick, run Cinebench, play whatever games you play, see if it's stable. Then try the other stick.
You had image corruption while using the IGP, that usually a sign of bad VRAM. Since the IGP doesn't have VRAM and uses the system RAM, I'm guessing that's where the problem is (even if memtest can't find it).
 
Yes, but not memtest. Just use one stick, run Cinebench, play whatever games you play, see if it's stable. Then try the other stick.
You had image corruption while using the IGP, that usually a sign of bad VRAM. Since the IGP doesn't have VRAM and uses the system RAM, I'm guessing that's where the problem is (even if memtest can't find it).
So I took out one stick and everything works fine. Ran some benchmarks and played Cyberpunk for a while. Still only dGPU on.
Decided to activate MsHybrid and rebooted. No issues.

What actually started the flickering of the screen with IGP enabled was when I shutdown my laptop and held the power button for 30 seconds. This also causes my laptop to start charging, like the battery isn't fully charged (99%).

Anyways. After that I turned the laptop back on, still one stick of 16GB inside and noticed a few times short screen flickerings. Then started a OCCT benchmark and it only got worse.
video_2024_03_22_21_01_23.gif


Could this have anything to do with my battery / PSU? I've tested the PSU with a multi-meter and it shows no issue.
 
So I took out one stick and everything works fine. Ran some benchmarks and played Cyberpunk for a while. Still only dGPU on.
Decided to activate MsHybrid and rebooted. No issues.

What actually started the flickering of the screen with IGP enabled was when I shutdown my laptop and held the power button for 30 seconds. This also causes my laptop to start charging, like the battery isn't fully charged (99%).

Anyways. After that I turned the laptop back on, still one stick of 16GB inside and noticed a few times short screen flickerings. Then started a OCCT benchmark and it only got worse.
View attachment 340165

Could this have anything to do with my battery / PSU? I've tested the PSU with a multi-meter and it shows no issue.
I'd say we're looking at a faulty CPU. You can try the other stick, you'll probably see the same thing.
One other thing you could try, is use the other RAM slots.

But seriously, send that thing in for service, it's not working at default values, that's not normal.
 
Maybe some ant(s) crawled in there or a cockroach decided it made a good toilet and some circuitry has been compromised. Your first pic shows an input of over 40V?? IMHO best to send it back under warranty. If you want to risk it and take responsibility for possibly making things worse you could try disassembly and cleaning board with CRC contact cleaner (no residue type) but if it's something else like a dry joint it will not help.
 
Back
Top