Tuesday, February 25th 2025

NVIDIA GeForce RTX 50 Cards Spotted with Missing ROPs, NVIDIA Confirms the Issue, Multiple Vendors Affected

TechPowerUp has discovered that there are NVIDIA GeForce RTX 5090 graphics cards in retail circulation that come with too few render units, which lowers performance. Zotac's GeForce RTX 5090 Solid comes with fewer ROPs than it should—168 are enabled, instead of the 176 that are part of the RTX 5090 specifications. This loss of 8 ROPs has a small, but noticeable impact on performance. During recent testing, we noticed our Zotac RTX 5090 Solid sample underperformed slightly, falling behind even the NVIDIA RTX 5090 Founders Edition card. At the time we didn't pay attention to the ROP count that TechPowerUp GPU-Z was reporting, and instead spent time looking for other reasons, like clocks, power, cooling, etc.

Two days ago, one of our readers who goes by "Wuxi Gamer," posted this thread on the TechPowerUp Forums, reporting that his retail Zotac RTX 5090 Solid was showing fewer ROPs in GPU-Z than the RTX 5090 should have. The user tried everything from driver to software re-installs, to switching between the two video BIOSes the card comes with, all to no avail. What a coincidence that we had this card in our labs already, so we then dug out our sample. Lo and behold—our sample is missing ROPs, too! GPU-Z is able to read and report these units counts, in this case through NVIDIA's NVAPI driver interface. The 8 missing ROPs constitute a 4.54% loss in the GPU's raster hardware capability, and to illustrate what this means for performance, we've run a couple of tests.

In the first test, "Elden Ring" at 4K UHD with maxed out settings and native resolution (no DLSS), you can see how the Zotac RTX 5090 Solid falls behind every other RTX 5090 we tested, including the NVIDIA Founders Edition, a de facto reference-design that establishes a performance baseline for the RTX 5090. The Zotac card is 5.6% slower than the FE, and 8.4% slower than the ASUS ROG Astral RTX 5090 OC, the fastest custom design card for this test. Officially, the Solid is clocked at 2407 MHz rated boost frequency, which matches the Founders Edition clocks—it shouldn't be significantly slower in real-life. The interesting thing is that the loss of performance is not visible when monitoring the clock frequencies, because they are as high as expected—there's just fewer units available to take care of the rendering workload.

A ROP (Raster Operations Pipeline) unit in the GPU processes pixel data, handling tasks like blending, antialiasing, render-to-texture, and writing final pixel values to the frame buffer. In contrast, a shading unit, aka "GPU core" is responsible for computing the color, lighting, and material properties of pixels or vertices during the rendering process, without directly interacting with the frame buffer, so the performance hit of the eight missing ROPs depends on how ROP-intensive a game is.
For example, in Starfield, the performance loss is much smaller, and in DOOM Eternal with ray tracing, the card actually ends up close to its expected performance levels.

We've also put the card through a quick 3DMark Time Spy Extreme graphics score run.
  • NVIDIA Founders Edition: 25439
  • Zotac Solid: 22621
  • Gigabyte Gaming OC: 26220
This should be a number that you can test easily for yourself, if you're one of the lucky RTX 5090 owners. The quickest way is definitely to just fire up GPU-Z and look at the ROP count number, it should be "176."

So far, we know only of Zotac 5090 Solid cards that are affected, none of our review samples from ASUS, Gigabyte, MSI, Palit, and NVIDIA exhibit this issue, all 5090 owners should definitely check their cards and report back.

This is an issue with quality assurance at both NVIDIA and Zotac. NVIDIA's add-in card partners (AICs) do not have the ability to configure ROP counts, either physically on the silicon, or in the video BIOS, and yet the GPU, its video BIOS, and the final product, cleared QA testing at both NVIDIA and Zotac.

We are working with Zotac to return the affected card, so they can forward it to NVIDIA for investigation. At this time Zotac was unable to provide a statement, citing the fluidity of the situation. As for possible fixes. We hope the issue is localized to a bug with the driver or the video BIOS, so NVIDIA could release a user-friendly BIOS update tool that can run from within Windows and update the BIOS of the affected cards. If, however, the ROPs were disabled at the hardware-level, then there's little that end-users or even AIC partners can do, except initiating a limited product recall for replacements or refunds. If the ROPs really are disabled through fuses, it seems unlikely that NVIDIA has a way to re-enable those units in the field, because that would potentially provide details to how such units can be reactivated on other cards and SKUs from the company.

Update 14:22 UTC:
Apparently the issue isn't specific to Zotac, HXL posted a screenshot of an MSI RTX 5090D, the China-specific variant of the RTX 5090 with nerfed compute performance, but which is supposed to have 176 ROPs. Much like the Zotac RTX 5090 Solid, it has 8 missing ROPs.

Update 16:38 UTC:
Another card has been found, this time from Manli.

Update 17:30 UTC:
ComputerBase reports that their Zotac RTX 5090 Solid sample is not affected and shows the correct ROP count of 176. This confirms that the issue isn't affecting all cards of this SKU and probably not even all cards in a batch/production run.

Update 17:36 UTC:
Just to clarify, because it has been asked a couple of times. When no driver is installed, GPU-Z will use an internal database as fallback, to show a hardcoded ROP count of 176, instead of "Unknown." This is a reasonable approximation, because all previous cards had a fixed, immutable ROP count. As soon as the driver is installed, GPU-Z will report the "live" ROP counts active on the GPU—this data is read via the NVIDIA drivers.

Update 19:18 UTC:
A card from Gigabyte is affected, too.

Update Feb 22nd, 6:00 UTC:
Palit, Inno3D and MSI found to be affected as well

Update Feb 22nd, 6:30 UTC:
NVIDIA's global PR director Ben Berraondo confirmed this issue. He told The Verge:
NVIDIAWe have identified a rare issue affecting less than 0.5% (half a percent) of GeForce RTX 5090 / 5090D and 5070 Ti GPUs which have one fewer ROP than specified. The average graphical performance impact is 4%, with no impact on AI and Compute workloads. Affected consumers can contact the board manufacturer for a replacement. The production anomaly has been corrected.
Very interesting—NVIDIA confirms that RTX 5070 Ti is affected, too.

While NVIDIA talks about "one ROP unit," this really means "8 ROPs" in our context. Many years ago, marketing decided that higher numbers = better, so they started to report the number of pixels that can be processed per unit, instead of the actual unit counts. So in this case, one hardware unit is disabled, which mean eight pixels per clock less can be processed, resulting in a loss of "8 ROPs".

Update Feb 25th:
In the meantime, some RTX 5080 GPUs with missing ROPs were found, too, NVIDIA provided the following statement to TechPowerUp:
NVIDIAUpon further investigation, we've identified that an early production build of GeForce RTX 5080 GPUs were also affected by the same issue. Affected consumers can contact the board manufacturer for a replacement.
Add your own comment

491 Comments on NVIDIA GeForce RTX 50 Cards Spotted with Missing ROPs, NVIDIA Confirms the Issue, Multiple Vendors Affected

#301
JustBenching
BwazeAre the RMAs sent from the normal stock which is nonexistent now?

“According to a post on X (formerly Twitter) by UK retailer, Overclockers.co.uk, the wait for fresh RTX 5090 stock could be between three and 16 weeks, potentially meaning buyers in the UK at least could be waiting until mid-May to get hold of one of the cards.“
They are supposed to keep a set amount of cards for RMA but during launch they usually don't do that (ie. amd didn't have any during the XTX debacle).
Posted on Reply
#302
Hecate91
Super XPI wouldn't be surprised if Nvidia was actually doing this. lol
Agreed, I definitely wouldn't put it past Nvidia for them to do it on purpose.
BwazeAre the RMAs sent from the normal stock which is nonexistent now?

“According to a post on X (formerly Twitter) by UK retailer, Overclockers.co.uk, the wait for fresh RTX 5090 stock could be between three and 16 weeks, potentially meaning buyers in the UK at least could be waiting until mid-May to get hold of one of the cards.“
Or how could anyone know if an RMA sent in could be replaced with another card with missing ROPs?
Posted on Reply
#303
Denver
Jensen's solution: Mask the hardware defect via software, send the same defective GPU back to the client.

What the eyes can't see, the heart can't feel. :D
Posted on Reply
#304
Quicks
efikkanLet one thing be clear; they knew. It's 100% a conscious decision. It's not a QA issue.
It's not like they assemble graphics cards and then discover how many ROPs are usable. This means these GPUs are a lower bin, and that's how they know exactly how many are made (usually multiple bins go into a single SKU).

Perhaps they thought it would go unnoticed, or that the performance difference would be insignificant.
Yeah, thats what they all think until they get cought with their pants down...

The truth has a way to come out eventually, are they really so arrogant to think people won't notice?

There is something more to this story and why did they decide to remove the hot-spot sensor as well, after it was said blackwell runs hotter then it should?
DenverJensen's solution: Mask the hardware defect via software, send the same defective GPU back to the client.

What the eyes can't see, the heart can't feel. :D
You will still see lower performance?
Posted on Reply
#305
JustBenching
QuicksYeah, thats what they all think until they get cought with their pants down...

The truth has a way to come out eventually, are they really so arrogant to think people won't notice?

There is something more to this story and why did they decide to remove the hot-spot sensor as well, after it was said blackwell runs hotter then it should?
If they knew then they would have send such a card to a reviewer. Especially not as popular as TPU. Come on now....
Posted on Reply
#306
Quicks
JustBenchingIf they knew then they would have send such a card to a reviewer. Especially not as popular as TPU. Come on now....
That's what I am saying doubt they would be so arrogant. But its not Nvidia that send these cards to reviewers but AIB partners. They are also to blame and not checking for things like this.
Posted on Reply
#307
Sound_Card
Vayra86And here you are, again, flame baiting the same people into the same old discussion we've seen a half dozen times the last month. Makes you wonder whether you're getting paid for it at this point.

Get a life
Seriously, I hope he gets paid for it, because it would actually be even worse if he was doing it for free.
Posted on Reply
#308
Hecate91
QuicksYou will still see lower performance?
All Nvidia has to say is the cards will have "up to" the claimed performance.
QuicksThat's what I am saying doubt they would be so arrogant. But its not Nvidia that send these cards to reviewers but AIB partners. They are also to blame and not checking for things like this.
But should the AIBs have to check every single card they sell in GPU-Z? IMO, the AIBs should be able to trust the chips Nvidia is selling them.
Posted on Reply
#309
Vya Domus
Hecate91But should the AIBs have to check every single card they sell in GPU-Z? IMO, the AIBs should be able to trust the chips Nvidia is selling them.
I've already explained it in a previous post, AIBs know that they are shipping these chips because they need a different VBIOS and they have to make sure it's the right one otherwise the GPU wont function correctly. Both Nvidia and their board partners did this on purpose, it wasn't a mistake, they knew exactly what they were selling.
Posted on Reply
#310
MxPhenom 216
ASIC Engineer
Please God may the 9070XT be interesting
Posted on Reply
#311
JustBenching
Vya DomusI've already explained it in a previous post, AIBs know that they are shipping these chips because they need a different VBIOS and they have to make sure it's the right one otherwise the GPU wont function correctly. Both Nvidia and their board partners did this on purpose, it wasn't a mistake, they knew exactly what they were selling.
And they sent one of these cards to TPU to make sure their little scheme stays under the radar. Makes sense.
Posted on Reply
#312
Vya Domus
JustBenchingAnd they sent one of these cards to TPU to make sure their little scheme stays under the radar. Makes sense.
It doesn't matter if it makes sense or not, it's what happened, you don't get a random batch of chips with less ROPs and then miraculously they all end up with a correct VBIOS (the firmware has to come from Nvidia by the way) which supposedly you wouldn't know about ? You realize this sounds orders of magnitude more stupid, right ?
Posted on Reply
#313
Hecate91
Vya DomusI've already explained it in a previous post, AIBs know that they are shipping these chips because they need a different VBIOS and they have to make sure it's the right one otherwise the GPU wont function correctly. Both Nvidia and their board partners did this on purpose, it wasn't a mistake, they knew exactly what they were selling.
The changed VBIOS seems like it was rushed given some cards are crashing or artifacting, but I would expect Nvidia strong armed the AIBs into selling defective chips with the reasoning of being forced to sell the chips they received or they don't get any at all, because of the reputation Nvidia has with their partner brands.
JustBenchingAnd they sent one of these cards to TPU to make sure their little scheme stays under the radar. Makes sense.
Of course not, cards tech reviewers get are hand picked samples.
Posted on Reply
#314
JustBenching
Hecate91Of course not, cards tech reviewers get are hand picked samples.
You do understand that TPU received one of those missing ROPs cards from zotac, right? Handpicked, sure.
Posted on Reply
#315
pavle
Maybe they've (nvidia) mixed up the chips that were supposed to be for RTX5090_D (but it now has same config as regular 5090)... Quite a mess!
Posted on Reply
#316
Darmok N Jalad
This just seems weird to me. I would think that if a card shipped out with disabled hardware, wouldn't the inconsistency in hardware cause driver instability if the driver was trying to use a pool of ROPs that's disabled? Or maybe there's enough self-correcting software to avoid crashes? Are the BIOSes the same between "anomaly" cards and fully-functional ones?
Posted on Reply
#317
Hecate91
JustBenchingYou do understand that TPU received one of those missing ROPs cards from zotac, right? Handpicked, sure.
I don't see news of this in the OP.
Edit- I see it now, they had a Zotac card for review.
I find it weird TPU didn't say anything about the missing ROPs sooner.
Posted on Reply
#318
wNotyarD
Hecate91I don't see news of this in the OP.
Edit- I see it now, they had a Zotac card for review.
I find it weird TPU didn't say anything about the missing ROPs sooner.
Did they publish a review of the Zotac at all?
Posted on Reply
#319
bearClaw5
If people are dumb enough to pay $2000+ for a 70 series card, more power to nvidia to take advantage of them.
Posted on Reply
#320
Vya Domus
Darmok N JaladThis just seems weird to me. I would think that if a card shipped out with disabled hardware, wouldn't the inconsistency in hardware cause driver instability if the driver was trying to use a pool of ROPs that's disabled? Or maybe there's enough self-correcting software to avoid crashes? Are the BIOSes the same between "anomaly" cards and fully-functional ones?
It's not the driver which ensures correct hardware functionality but the firmware, aka VBIOS, these cards have to be shipped with a different firmware for them to work correctly, there is no self correcting software, the firmware needs to know to which units it can issue micro ops.

It works like this :

Shader is compiled by the driver -> instructions are sent to the GPU -> there's a command processor on the GPU which takes these instruction turns them into micro ops and dispatches them to corresponding units. The command processor can't know which units work and which don't, the firmware on the card needs to have that information.
Posted on Reply
#321
cmguigamf
MxPhenom 216Please God may the 9070XT be interesting
At this point if 9070XT ship as announced, it might already be a win for AMD.
Posted on Reply
#322
Hecate91
wNotyarDDid they publish a review of the Zotac at all?
I haven't seen a review for a Zotac RTX 5090 Solid, either TPU didn't publish it yet, or I suspect Nvidia knew about this issue and told reviewers to not say anything about the affected cards.
Posted on Reply
#323
_roman_
Do anyone believe in that random picked number? I really doubt that number.
It is nice to state that it's only a small number to calm down the Haters, Fanboys and those who do not belong in any of those two groups.

If nvidia really know it is 0.5% than recall all those chips. If products are made than put a sticker on the box, flash a different firmware and sell those cards as nvidia 5090 G = Good Value Garbage edition (D is already used for D=defect)
We have identified a rare issue affecting less than 0.5% (half a percent) of GeForce RTX 5090 / 5090D and 5070 Ti GPUs which have one fewer ROP than specified. The average graphical performance impact is 4%, with no impact on AI and Compute workloads. Affected consumers can contact the board manufacturer for a replacement. The production anomaly has been corrected.
Personally I would not call 0.5 % a small value. Far too large. I doubt that only one in every 200 cards is effected. I think it's a higher number.
That issue popped up quite fast. Not everyone reads, not everyone uses or has access that special read out software. Not everyone runs that software after the driver for correct readout is installed. Not everyone uses windows.
Vya DomusIt's not the driver which ensures correct hardware functionality but the firmware, aka VBIOS, these cards have to be shipped with a different firmware for them to work correctly, there is no self correcting software, the firmware needs to know to which units it can issue micro ops.
I highly doubt that. The hardware has some mechanism to know itself which ones are defect. I assume those defect areas do not exists anymore because the connections are lasercut in the factory. The firmware just probes the hardware and see it.

Anyway - the firmware needs to be programmed for such a mechanism beforehand. The hardware needs to be prepared for that mechanism beforehand. There is no random we have less rop issue.
Posted on Reply
#324
JustBenching
Hecate91I don't see news of this in the OP.
Edit- I see it now, they had a Zotac card for review.
I find it weird TPU didn't say anything about the missing ROPs sooner.
So your handpicked theory goes out of the window, right? Unless you meant they are handpicking for faulty cards.
Posted on Reply
#325
_roman_
Super XPThis is a perfect opportunity for AMD to screw up a perfect opportunity.
They already screwed up. But hey, they can screw it up many times more with the next graphic card release.
Posted on Reply
Add your own comment
Feb 25th, 2025 13:38 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts