PNY RTX 4070 Ti XLR8 OC - New worse BIOS/version?

[freeagent]

so at the end you can hear some coil whine?

I have to get right up to the case, it is barely audible, and it has to be in the thousands of fps iirc

 

[ZVom]

Well, when coils whine at 600+fps I personaly consider it acceptable.
The problem is when GPU buzzes at 100 fps.

 

[Reverse17]

Hmmm, the weight seems to be slightly different from the piece tested in TPU, the question is to what extent it's an error between the kitchen scales and the actual changes..

TPU
My XLR8

While weighing the card I cleaned the front air filter and there is a noticeable difference

Currently this is how it looks with the case closed and the 140mm vent ~850-900RPM, ambient temperature around ~23-24'c:

• Starfield after about ~20 minutes of gaming, UV
• CP 2077 UV (fluctuates more or less at ~998-1050RPM)
• CP 2077 stock (screen taken when the highest RPMs were noted, often when playing vent oscillate closer to those unfortunate 1450-1500RPM)

In conclusion, the card will be used anyway with UV until it dies or is sold so it looks really good, worry a bit about the results on STOCK though, my previous copy from February performed clearly better, in such CP 2077 in the same PC configuration had about 1100-1200RPM at 62-63'c, that is, it coincided with the TPU test, today? I believe that they modified with bios for a reason (but these are just theories and guesses..)

Spoiler

PS: the silence of the manufacturer on this issue and the lack of taking an official position also gives food for thought

 

[ZVom]

Hmmm, the weight seems to be slightly different from the piece tested in TPU, the question is to what extent it's an error between the kitchen scales and the actual changes..

TPU
My XLR8

While weighing the card I cleaned the front air filter and there is a noticeable difference

Currently this is how it looks with the case closed and the 140mm vent ~850-900RPM, ambient temperature around ~23-24'c:

• Starfield after about ~20 minutes of gaming, UV
• CP 2077 UV (fluctuates more or less at ~998-1050RPM)
• CP 2077 stock (screen taken when the highest RPMs were noted, often when playing vent oscillate closer to those unfortunate 1450-1500RPM)

In conclusion, the card will be used anyway with UV until it dies or is sold so it looks really good, worry a bit about the results on STOCK though, my previous copy from February performed clearly better, in such CP 2077 in the same PC configuration had about 1100-1200RPM at 62-63'c, that is, it coincided with the TPU test, today? I believe that they modified with bios for a reason (but these are just theories and guesses..)

Spoiler

PS: the silence of the manufacturer on this issue and the lack of taking an official position also gives food for thought

Damn bro, could you provide your UV curve from afterburner? That looks really good.
62C with 1000RPM is ridiculously good.
Are you sure your previous sample was an OC one?

Wouldn't be worried by stock temps as you're not going to use stock curve

 

[Reverse17]

Sure! When I'm home I'll shoot a screenshot of the MSI AB curve (unless you can somehow export the profile to a file?).

Yes, I'm sure the February piece was also the OC version, I had it in the chassis for nearly 2 weeks and had time to get to know it well, in short, where my current XLR8 has ~62'c at 1000RPM, the February copy would probably have 55'c at 1000RPM

In my opinion, they messed up something in this model and now they are ashamed to admit it. (If the weight difference is real then maybe something with the heat pipes?)

However, it doesn't change the fact that it's still a very good version (after bios update) of RTX 4070 Ti, but no longer as specatcularly good as from the TPU test.

 

[3DVCash]

Hmmm, the weight seems to be slightly different from the piece tested in TPU, the question is to what extent it's an error between the kitchen scales and the actual changes..

TPU
My XLR8

While weighing the card I cleaned the front air filter and there is a noticeable difference

Currently this is how it looks with the case closed and the 140mm vent ~850-900RPM, ambient temperature around ~23-24'c:

• Starfield after about ~20 minutes of gaming, UV
• CP 2077 UV (fluctuates more or less at ~998-1050RPM)
• CP 2077 stock (screen taken when the highest RPMs were noted, often when playing vent oscillate closer to those unfortunate 1450-1500RPM)

In conclusion, the card will be used anyway with UV until it dies or is sold so it looks really good, worry a bit about the results on STOCK though, my previous copy from February performed clearly better, in such CP 2077 in the same PC configuration had about 1100-1200RPM at 62-63'c, that is, it coincided with the TPU test, today? I believe that they modified with bios for a reason (but these are just theories and guesses..)

Spoiler

PS: the silence of the manufacturer on this issue and the lack of taking an official position also gives food for thought

Really appreciate the detailed info. It seems like maybe they cut back a bit on the heatsink and just ramped up the fans instead to save a few bucks. Kinda scuzzy cutting corners AFTER all the reviews are out.

I think I'm just grabbing an MSI...

 

[ZVom]

Damn bro, could you provide your UV curve from afterburner? That looks really good.
62C with 1000RPM is ridiculously good.
Are you sure your previous sample was an OC one?

Wouldn't be worried by stock temps as you're not going to use stock curve

Can I see your MSI AB Curve?

 

[Reverse17]

Can I see your MSI AB Curve?

CLICK

Spoiler

UV I made with this method (ENG subtitles available)

 

[ZVom]

Ofc I know techmaniac. Did you take his exact setup?
Btw I just laid my hands on the card and I confirm I have the same issue.. once fans have spun, they don't go below 1450...
I have never flashed gpu's bios. Is it risky?

 

[eidairaman1]

Hi,
I recently bought a >>PNY RTX 4070 Ti XLR8 OC<< and I think I got a new version with a slightly worse fan curve, namely the minimum fan speed is: ~1450RPM (30%), there is no way to go below this value even manually (fan stop works, but the minimum speed after turning on the fans is ~1450RPM at 30%)

BIOS from my card (95.04.3C.C0.CD):

PNY RTX 4070 Ti VBIOS

12 GB GDDR6X, 2310 MHz GPU, 1313 MHz Memory

www.techpowerup.com

BIOS of the card from the TPU test (95.04.31.00.6E):

PNY RTX 4070 Ti VBIOS

12 GB GDDR6X, 2310 MHz GPU, 1313 MHz Memory

www.techpowerup.com

It is possible that the manufacturer decided to change the fan curve in the middle of the card's life...?

In more demanding games it reaches around ~1500-1600RPM, while in Time Spy Extreme Stress Test we have these results:

IMG 1
IMG 2

What do you think?

PS: UV 2850MHz / 1.000V - IMG 3
As you can see it would be possible to go down to these ~1000RPM but the card does not have this possibility, ~1450RPM is minimum.

Yes and it was probably to prevent cooking the card.

Go get a waterblock if the fan noise bothers you.

 

[ZVom]

Can someone who has the card run a bench on Unigine Benchmark with the setup and share the outcome?

BTW is your system also VEEEERY slow when the benchmark is running? My setup has problems with running any youtube clip in the background when the benchmar is up

 

[Reverse17]

Go get a waterblock if the fan noise bothers you.

Super approach to quiet changes hidden from the consumer..

If the card needed changes then it should be withdrawn from the market and release a new revised model under a different name or some addendum, not quietly reduce its performance culture with the hope no one will notice .... Someone will see the reviews before the changes and tempted by the performance of its cooling will be quite disappointed, like me in the first post.

@ZVom
If I don't forget, I'll try to see how it looks to me today.

 

[Tatty_Two]

Super approach to quiet changes hidden from the consumer..

If the card needed changes then it should be withdrawn from the market and release a new revised model under a different name or some addendum, not quietly reduce its performance culture with the hope no one will notice .... Someone will see the reviews before the changes and tempted by the performance of its cooling will be quite disappointed, like me in the first post.

@ZVom
If I don't forget, I'll try to see how it looks to me today.

I agree it feels underhand and it does happen fairly often with a few manufacturers although mostly they do tend to implement those changes with a revision, for example I have had a couple of gigabyte cards in the last 5 years that had lower power limit adjustments than their release models/bioses, for me it leaves a bad taste in my mouth (so to speak).

 

[ZVom]

Super approach to quiet changes hidden from the consumer..

If the card needed changes then it should be withdrawn from the market and release a new revised model under a different name or some addendum, not quietly reduce its performance culture with the hope no one will notice .... Someone will see the reviews before the changes and tempted by the performance of its cooling will be quite disappointed, like me in the first post.

@ZVom
If I don't forget, I'll try to see how it looks to me today.

Just a small reminded @Reverse17

 

[kurismasu]

Here my Heavenbenchmark result on I5 13400F. I know there is bottleneck at fullHD. In the future will replace I5 with I7.

OF course on PNY GPU

 

[Reverse17]

Can someone who has the card run a bench on Unigine Benchmark with the setup and share the outcome?

BTW is your system also VEEEERY slow when the benchmark is running? My setup has problems with running any youtube clip in the background when the benchmar is up

When running UHB 4.0, I can sense the slowdown of animation in windows, videos on YT also lose fluidity.... BUT, I think this is perfectly normal

PS: result at factory settings (the only modification is the new bios starting at 1000RPM as you can see in the RivaTuner information)

 

[sk0rp]

I wonder if it is worth to buy after changes. Bios update I guess can be risky?
How this card is loud in the new bios (1450 RPM+)? More than 30db?

 

[Reverse17]

I think updating the bios with a file from the manufacturer is just as "risky" as updating drivers

If it was a big risk then the manufacturer would not post the bios update on its website (several PNY RTX models have such an updater available on the manufacturer's website).

It is worth doing this immediately after purchasing the card.

 

[Deleted member 235585]

Hi, i've created this account just for saying thank you and absolutely not at the same time.
In fact, I was already interested in purchasing this card but as I learned that there were concerns about RPM and higher temperature etc.
I was reluctant, reading that a user had shared a tool (out of nowhere, unless I don't know) and a BIOS which apparently came from PNY (I have difficulty trusting like that on the internet, sorry )

I said to myself : perfect, for the price I have this 4070Ti, I go ahead and buy it, I received it, I did the bios update as indicated here with the files from here and everything goes on well for two days, besides I see that this card does not even accept a small +100 MHz, it's sad... it doesn't overclock easily (one of the worst in Power Limit anyway)

and after three days, I notice that very rarely when I play, suddenly the screen becomes ultra pixelated like I was in 480p, suddenly and even Windows is affected, I don't know if that really comes from the graphics card and I have not changed any components of my PC recently ( 13600k/32GB 6200mhz DDR5, NVME 3.0, Z690 UNIFY etc ) and I have an Acer XV272UX and it's been two years that i have this one and I had 4 in total and I have NEVER had this kind of problem Pixel Shifting / pixelated screen effect, really, so I suspect 99% or even 100% the BIOS update which must have done something to the graphics card, I don't know...

In addition, when I read that they say that when doing a benchmark they have their windows slow in the background, it's very strange.

My first PNY product was bad (a GTX 1650) if I still have a problem with them, I would never come back to them and I would consider it a cheap brand, not for me.

I had an Asus RTX 3070 Noctua Edition, I had no problems, I wanted to upgrade to 1440p, the performance gap is incredible but if there are problems, I replace it with another brand, besides I can have the 4070 Ti TUF OC at the same price...

Here is what it look like ( it is not my screens, i take it from another forum but it's 100% the same thing )

https://imgur.com/a/jyjzb

 

[Deleted member 235585]

Update : sorry apparently it was because of the Nvidia Drivers 546.01.
I've DDU and come back to the 545.92 and i no longer have the issue, anyway very strange !

 

[Klemc]

Better go back to 537.58 btw.

 

[Deleted member 235585]

Better go back to 537.58 btw.

Why ?
There is a specific reason ?
Anyway so weird that 546.01 cause the weird issues that i had !
Plus I hadn't seen anyone talking about it, in the end glad it didn't come from the bios update.
How do you know which Nvidia drivers update is better than another?

 

[Klemc]

I read 3DGuru pages, let's say after 537 they jumped straight to 545..., they never do that (nvidia), it should have been 538 if as habitually:

Videocards - NVIDIA GeForce Drivers Section

In this section, you can discuss everything NVIDIA driver related. GeForce and GeForce Experience drivers are for NVIDIA Quadro and all GeForce based videocards.

forums.guru3d.com

Each version has his own topic btw.

 

[Deleted member 235585]

J'ai lu les pages de 3DGuru, disons qu'après 537 ils sont passés directement à 545..., ils ne font jamais ça (nvidia), ça aurait dû être 538 si comme d'habitude :

Videocards - NVIDIA GeForce Drivers Section

In this section, you can discuss everything NVIDIA driver related. GeForce and GeForce Experience drivers are for NVIDIA Quadro and all GeForce based videocards.

forums.guru3d.com

Chaque version a son propre sujet d'ailleurs.

Thank you ! anyway i don't have any issues with 545.92 but for the love of god avoid 546.01 xD !

 

[hoagis321]

By now, the PNY 4080 XLR8 is also affected. Got one last week and returned it this week, because fan speed cannot be reduced below 1450 rpm, as with the 4070 tis and 4090s described here. Unlike most of the users in this forum, I do consider these fan speeds unpleasantly noisy, since I have a silent build. From my experience with MSI-cards, the difference between 1200 rpm (as the card was supposed to have under load) and said 1450 rmp is significant and makes this card imho useless for silent builds. The card had no fan stop in idle either, but PNY support assured me that this can be fixed. Minimum fan speed can definitely not be reduced below 1450, I was assured of this by PNY support. "Thermal optimization" was the thing that they called it.
For me, getting MSI-like noise levels for somewhat less than MSI itself demands for its cards was the point of this card. Since in my region you can get other 4080s for less (Inno3D or KFA2) than this PNY, I do not really understand what kind of market niche PNY is trying to occupy now that its exceptional noise levels are gone. Neither do I get the reason for this change, from what I found on the internet, the cards were doing just fine on the slower fan speeds up to now.