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

#327
_roman_
Bomby569Zotac had nothing to do with this
Even if those board partner knew, they rely on the goodwill and contracts to NVIDIA.
Posted on Reply
#328
Vya Domus
_roman_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.
No because those units are not always fused off, that's why in the past you could flash some cards and unlock shaders, because those were never defective or fused off to begin with, so there wouldn't even be a way to probe for disabled shaders. Some defects you also simply cannot probe for as they can cause intermittent issues so there is no reliable way to detect them through software, it would be a very bad idea to rely on something like that.

I don't think you realize how much stuff actually is present on these chips, you're talking about millions/billions of data lines, you simply cannot test everything every time, it's not feasible. It's more reliable to check the GPU die first, detect the defects and configure the firmware.
Posted on Reply
#329
_roman_
modmaxfor their very high price they present several problems first the 5.0 bus and black screen, then the connectors that melt, in some cases, now even the ROP, it doesn't seem to me that they are worth what they cost
No old cuda, no old physx
melted power supply side connectors + melted cable + defective connector on the graphic card
+ possible fire hazard - which everyone denies in this forum.
150°C thermal pictures are enough to state that in certain cases = trouble.

We rely on the power supply unit, because the cable has no fuses and the graphic card has no fuses.

As of now - no mainboard, no nvme, no monitor was harmed (assumption).
I would be very upset when my M2 NVME would be dead because of the mainboard because of a nvidia 5090 graphic card and a asus power supply.
I would be very upset whne my floor would burn and all the furniture because of a nvidia graphic card, asus power supply unit and a cable.
Who wants to make an insurance for 0€ and take all nvidia card owners as insurance takers? I would not bet on that in the long run.

Still no power supply unit company had made any statements? Or am I wrong? There are two sides. The psu side and the nvidia 5090 side. This only takes in consideration normal usage - not abuse or special use like overclocking or other operation modes I have not even thought of.
Posted on Reply
#330
Darmok N Jalad
Vya DomusNo because those units are not always fused off, that's why in the past you could flash some cards and unlock shaders, because those were never defective or fused off to begin with, so there wouldn't even be a way to probe for disabled shaders. Some defects you also simply cannot probe for as they can cause intermittent issues so there is no reliable way to detect them though software, it would be a very bad idea to rely on something like that.

I don't think you realize how much stuff actually is present on these chips, you're talking about millions of data lines, you simply cannot test everything every time, it's not feasible. It's more reliable to check the GPU die first, detect the defects and configure the firmware.
And that's my point. There would have to be intentional selling of disabled hardware if you're knowingly shipping a different firmware, unless every single 5090 that went to ZOTAC (and others) has these ROPs disabled, so they didn't know any better. That, or the full-ROP versions are not using some hardware because they are running the defect-edition firmware. It should bear out in the firmware, but it might be really hard to discover (I know I couldn't figure it out). What I'm getting at is that this couldn't be an "oops" in the supply chain if they are knowingly working around the missing ROPs with firmware. Maybe there's just one firmware that does a resource check and adjusts accordingly, so we'd never know.
Posted on Reply
#331
Bwaze
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.
Reviewers don’t usually comment on reliability of products, only sometimes after issues are public they comment if they had problems.

All manufacturing has some issues, even if you work in much more critical industries, so reporting a problem with a single unit isn’t really news worthy. I wouldn’t really read any conspiracy into that - a single failure is just that.

Many times the reviewers also notice issues with the product, but when notifying the company they are told they have pre-production sample, and the issues should be fixed in retail, and such issues many times aren’t mentioned in reviews. I know some reviewers then don’t publish such reviews since they don’t represent the retail product, others don’t have issues with that.
Posted on Reply
#332
Vya Domus
Darmok N JaladWhat I'm getting at is that this couldn't be an "oops" in the supply chain
Of course, Nvidia knows as every wafer is checked optically for defects, that's how they came up with that 0.5% figure, it's not like they made up some statistical estimation, they know exactly how many chips were defective and for what reason.

I still couldn't find any info if these GPU dies are marked differently, that would clarify a lot of things.
Posted on Reply
#333
Dragokar
Seeing so many pop up in reports it is hard to believe it is only 0.5%.......but since they don't disclose the numbers of shipped GPUs vs. defective ones it is hard to tell.

So now your cards can burn due to “user error”, can face black screens or miss performance targets when you got lucky enough to get one anyway. That's an “epic” launch, but it still benefits the company and the other two are going to screw up again I suppose.....
Posted on Reply
#334
sbacc
DragokarSeeing so many pop up in reports it is hard to believe it is only 0.5%.......but since they don't disclose the numbers of shipped GPUs vs. defective ones it is hard to tell.

So now your cards can burn due to “user error”, can face black screens or miss performance targets when you got lucky enough to get one anyway. That's an “epic” launch, but it still benefits the company and the other two are going to screw up again I suppose.....
It's only 0.5% for dGPU sold through retailer, but what is the percentage for system builder... I mean how many percent of those customer follow the technews and are able to check for those ROP...

Edit : by customer I mean the final one, not the system builder making the PC
Posted on Reply
#335
trsttte
The issue is they got caught :D

Next driver version will stop reporting real active ROP numbers and just have a define with whatever the number is supposed to be, problem solved.

An entire new level of silicon lottery was just unlocked :D
Posted on Reply
#336
W1zzard
Hecate91I 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.
The Zotac review is in progress, I noticed the odd performance and they shipped a second card so that I can compare both, to further debug the issue. We were thinking GPU clock issue, maybe PCIe 5.0, or lanes, or power, or heat.

Until yesterday, nobody noticed that ROPs were missing. When I found out I drafted the news post, informed Zotac, asked for a statement and published our story, nothing was held back.

Also @wNotyarD
trsttteNext driver version will stop reporting real active ROP numbers and just have a define with whatever the number is supposed to be, problem solved.
That won't work .. what about the missing performance?
Posted on Reply
#337
Hecate91
W1zzardThe Zotac review is in progress, I noticed the odd performance and they shipped a second card so that I can compare both, to further debug the issue. We were thinking GPU clock issue, maybe PCIe 5.0, or lanes, or power, or heat.

Until yesterday, nobody noticed that ROPs were missing. When I found out I drafted the news post, informed Zotac, asked for a statement and published our story, nothing was held back.

Also @wNotyarD
Thank you for the clarification.
As you can tell I just don't have much of any trust left for Nvidia, after their doubling down on the power connector, and now trying to downplay missing ROPs, I wonder how Nvidia is actually going to fix this or are the AIBs going to be left dealing with customers.
Posted on Reply
#338
Anc13ntEvil
How would they know it only affects .5% of cards? It's only been a couple of days since this news came out. Maybe they already knew how many cards were affected and sold them anyway? Perhaps I'm seeing a conspiracy theory where there is none? Then again, it is Nvidia.
Posted on Reply
#339
Princeinterweb
I am so glad I am in a near desperate need for a GPU this cycle..........
Posted on Reply
#340
lexluthermiester
And this is one of the reasons why I'm going to be skipping the RTX5000 gen. I usually skip a gen in my personal upgrade cycles, but not this time. Now I'm either going with an RTX4000 or just skip two gens. This has quickly become a mess of incompetence and has been pathetic in nearly every respect. From the blatant performance gap from what was OPENLY promised VS what was actually delivered, to the melting power jacks(AGAIN), to the price gouging bullshit and now this revelation about missing specs that should be present and are not.

What they need to do to at least partly recover is as follows:
A. Reduce prices(a lot) to reflect the piss poor performance compared to previous gen cards.
B. Do away with the dangerous ATX3.0 power jack and bring back the PCIe jacks that are KNOWN to be safe and DON'T melt.
C. Offer replacements for cards that not at the declared specs.
D. Plan for a better future, one where they are honest, competent and fair.

@ NVidia
People, pull your heads out of your asses, get your shit together and quit being incompetent jerks! This showing is piss poor and unacceptable. You REALLY need to pull a GTX1000 series kind of run with RTX6000, in every respect, including performance/power/price ratio.

@ Jensen Huang
All of these circumstances clubbed together make you look like a liar. You claimed specific levels performance, RTX5000 has not delivered. You are either woefully out of touch with reality(likely given your comment about $10,000 PCs) or have lost grip of your senses & mental faculties entirely. Either way, get a grip or step down.


I think AMD and Intel really have a chance to close the gap here.. And yes folks I'm an NVidia guy. However I'm not a fanboy nor someone whose judgment is less than objective.
Posted on Reply
#341
Solaris17
Super Dainty Moderator
lexluthermiesterThis has quickly become a mess of incompetence. This has been pathetic in nearly every respect. From the blatant performance gap from what was OPENLY promised VS what was actually delivered, to the melting power jacks(AGAIN), to the price gouging bullshit and now this revelation about missing specs that should be present and are not.
I already own an arc welder so wifes gonna be pissed. Its a shame from a PCB size standpoint its an amazing feat of engineering.
Posted on Reply
#342
Tek-Check
As somebody said elsewhere, the issues are mounting almost daily:

- consumer brainwashing by marketing that 5070 is "faster" than 4090
- MSRP means "Multiple $hitty Retail Prices"
- more users complain about black screen issue
- 600W power connector is still melting two years later, with no user error (DerBauer)
- no cards to be purchased anywhere in any meaningful volume
- PhysX 32-bit support ended ( list of affected games )
- defective chips with less ROPs that Nvidia knew about before shipping (legal issue?)
- a halo card catches random fire (not enough time for AIBs to test PCBs?)
Posted on Reply
#343
W1zzard
Hecate91I wonder how Nvidia is actually going to fix this or are the AIBs going to be left dealing with customers.
These usual mechanics for cases like this are that merchants or AICs (depending on local laws) collect the defective cards and replace or refund. The GPUs go back to NV and AICs get credit for them. This is what happens with RMAs as well. Any component returned will be repurposed, all these companies have staff and equipment to repurpose components, nothing goes to waste

Of course things could be different in this case, I have no official or inofficial information. Just sharing what's done usually
Posted on Reply
#344
lexluthermiester
Solaris17I already own an arc welder so wifes gonna be pissed. Its a shame from a PCB size standpoint its an amazing feat of engineering.
5090? I shouldn't have to tell you this, but I'll say it for everyone else reading, make sure all of the pins on that power cable are properly & firmly seated in the jack socket. Be extra careful. There is potential for fire.
Posted on Reply
#345
tvshacker
W1zzardThat won't work .. what about the missing performance?
They'll claim temperature constraints or bad airflow in the case, not allowing max boost/throttling
Posted on Reply
#346
W1zzard
tvshackerThey'll claim temperature constraints or bad airflow in the case, not allowing max boost/throttling
That might work with end users who have a single card and no experience with professional testing .. they won't ever notice 5-10% missing performance.. not gonna work here when I have a stack of cards of the same SKU that all perform significantly better. Just take out one card, plop in another, numbers won't lie
Posted on Reply
#347
Tek-Check
tvshackerThey'll claim temperature constraints or bad airflow in the case, not allowing max boost/throttling
No. There is a physical defect on the chip itself that does not fit the official spec.
It can become a legal issue immediately if anyone tries to hide anything.
Posted on Reply
#348
Dragokar
W1zzardThe Zotac review is in progress, I noticed the odd performance and they shipped a second card so that I can compare both, to further debug the issue. We were thinking GPU clock issue, maybe PCIe 5.0, or lanes, or power, or heat.

Until yesterday, nobody noticed that ROPs were missing. When I found out I drafted the news post, informed Zotac, asked for a statement and published our story, nothing was held back.

Also @wNotyarD


That won't work .. what about the missing performance?
That's well organized.
Posted on Reply
#349
chstamos
lexluthermiesterI think AMD and Intel really have a chance to close the gap here.. And yes folks I'm an NVidia guy. However I'm not fanboy nor someone whose judgment is less than objective.
They'd have to be stark raving mad to concede defeat at this point. Even intel, with all its problems. nvidia has been screwing the pooch for multiple generations of product now, either iterative or complete garbage like the current one. Remember when intel was considered invincible, the 1000-pound gorilla that NOONE, noone could be possibly out-R&D and outcompete? We've been fed the line that nvidia could at any moment's notice crush the competition, while all they had in line was the shittiest generation of Geforce since the leafblower thing.

The gaming market is up for grabs. And I don't want to use that kinda sexist glengarry glenn ross line from Blake but.... it's that.
Posted on Reply
#350
MacZ
efikkanPerhaps they thought it would go unnoticed, or that the performance difference would be insignificant.
For most people, this (missing ROPs) will go unnoticed and they won't request a RMA.

From my understanding, to have as much stock as possible, nVidia just allowed much more questionable quality than before.

These gaming GPUs are competing for wafer space with $25,000 AI GPUs that CEOs are having meeting with Jensen Huang to beg for having their order processed in a reasonable time.

People should be happy to have any at all, because they are in no way, shape or form entitled to any GPUs at any price point.

They are otherwise welcome to build their own GPU company. Ask the chinese for pointers, maybe.
Posted on Reply
Add your own comment
Feb 25th, 2025 13:02 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts