Saturday, February 22nd 2025

NVIDIA GeForce RTX 5090 Spotted with Missing ROPs, NVIDIA Confirms the Issue, Multiple Vendors Affected, RTX 5070 Ti, Too

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".
Add your own comment

426 Comments on NVIDIA GeForce RTX 5090 Spotted with Missing ROPs, NVIDIA Confirms the Issue, Multiple Vendors Affected, RTX 5070 Ti, Too

#351
lexluthermiester
chstamosEven intel, with all its problems.
Intel's got game going on! The B580/B570 are surprisingly robust and if they can kick it up a few levels with the B750/B770, then they will be a very serious competitor for both NVidia and AMD. I'm crossing my fingers they pull the rabbit out of that hat and kick some GPU market butt! Of course, I hope AMD does the same thing with RDNA4 and they both really rip NVidia a new one.
Posted on Reply
#352
Legacy-ZA
lexluthermiesterIntel's got game going on! The B580/B570 are surprisingly robust and if they can kick it up a few levels with the B750/B770, then they will be a very serious competitor for both NVidia and AMD. I'm crossing my fingers they pull the rabbit out of that hat and kick some GPU market butt! Of course, I hope AMD does the same thing with RDNA4 and they both really rip NVidia a new one.
Would have bought one to tinker with, but hey, they decided $250 is magically $400, so,. no.
Posted on Reply
#353
Tropick
MacZPeople 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.
Holy smokes what a take. I don't know about you but when I shell out the money that a company is asking for a product they're selling I feel pretty damn entitled to it. Considering I, you know, paid for it?
Posted on Reply
#354
MacZ
lexluthermiesterThis mental drivel is why companies like NVidia think they can charge outlandish prices. :shadedshu: Why? Why did you even bother with that nonsense? :rolleyes:
What nonsense ?

You forgot to provide any counterargument.

If you don't like reality, maybe you should ask that someone stops the planet from rotating so that you could get off.
Posted on Reply
#355
lexluthermiester
MacZWhat nonsense ?

You forgot to provide any counterargument.
No counterargument is required. Your statement was so lacking of all merit and logic that any counterargument would be completely superfluous. Your statement argues itself into the realms fanciful delusion, not worthy of any consideration. It is worthy only of being mocked.
Posted on Reply
#356
JustBenching
MacZFor 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.
That's not how it works. Chips are made months in advance. Many many months. This is a complete failure on nvidias part. Whatever the reason it's not a great look.
Posted on Reply
#357
MacZ
lexluthermiesterNo counterargument is required. Your statement was so lacking of all merit and logic that any counterargument would be completely superfluous. Your statement argues itself into the realms fanciful delusion, not worthy of any consideration.
No argument because you have none.

Selling something at $25,000 rather than at $2,000 is what all companies on the planet will do 100% of the time.
Posted on Reply
#358
lexluthermiester
MacZNo argument because you have none.
Yes, THAT'S got to be it..
MacZSelling something at $25,000 rather than at $2,000 is what all companies on the planet will do 100% of the time.
Only if their goal is to be out of business. Are you done gracing these forums with embarrassing nonsense?
Posted on Reply
#359
trsttte
W1zzardThat won't work .. what about the missing performance?
Chuck it down to silicon lottery and run to run variance ¯\_(ツ)_/¯
They can also adjust boost behaviour to make the issue even less pronounced, make the loser cards boost slighlty higher to compensate and no one will be the wiser.

10% is noticeable but even with your stack of cards would you notice a 5% difference (or 4% as claimed by nvidia)?
lexluthermiesterif they can kick it up a few levels with the B750/B770
The one that reportdly was in development hell and has yet to tape out? It would be great if Intel got it's house in order but that doesn't seem in the cards short term.
MacZSelling something at $25,000 rather than at $2,000 is what all companies on the planet will do 100% of the time.
That's a great way to become the next Intel. They have multiple market segments to serve, with different volumes and profit margins involved. Selling a 2000$ consumer GPU is not charity, it's a business decision, just like AMD selling 500$ CPU. Just like both of them pretty much not selling any new product under 200$ leaving that to previous generations. It's business.
Posted on Reply
#360
MacZ
lexluthermiesterYes, THAT'S got to be it..

Only if their goal is to be out of business. Are you done gracing these forums with embarrassing nonsense?
At that point, when you are so much out of touch with reality ...

pcmasterrace/comments/1awtso6
Gaming has become a rounding error for nVidia.

Maybe after cryptocurrencies and AI, you should understand that GPUs, the 'G' notwisthanding, are very useful SIMD processors that are and will be used for much more than just gaming.

This will continue and amplify.
Posted on Reply
#361
Tropick
MacZSelling $2,000 GPUs when you could make and easily sell $25,000 GPUs instead is charity at this point. And not insulting the future.

You should be grateful that you are offered any GPU for gaming at all is what I'm saying.
Okay so let me make sure I understand this perspective. Let's say you buy an car from a company that also sells larger fleet vehicles to other businesses. You then learn that the car you bought has a misfire on one of the cylinders that reduces it's performance. The car still works, but does not have the performance that was advertised by the manufacturer. You'd feel lucky to have been sold a defective product just because the company had the opportunity to put more resources towards making more money doing something else? It was their choice to sell and support the lower end product!! They're obligated to keep their word if they're accepting payment for it! I don't understand, what's wrong with being upset that you paid full price and then received a thing that wasn't as advertised? They should have reduced the price, or better yet fixed the issue, if they knew it didn't meet the specifications they advertised it as having! That feels pretty reasonable!
Posted on Reply
#362
W1zzard
trstttewould you notice a 5% difference (or 4% as claimed by nvidia)?
Absolutely, the delta between the FE and the fastest factory OC'd card, the ASUS Astral is 4-5% only, so yes, this will definitely raise eyebrows, especially when a card drops to below FE levels
Posted on Reply
#363
brianinengland
This could fall foul of UK/EU consumer laws, which state that a product must be 'fit for purpose' and work 'as described'

If it's clearly stated that the 5090 has 176 ROPs, then that's what the buyer can legally expect.
Buyer would be entitled to a full refund under our Sale of Goods Act
Posted on Reply
#364
MacZ
TropickOkay so let me make sure I understand this perspective. Let's say you buy an car from a company that also sells larger fleet vehicles to other businesses. You then learn that the car you bought has a misfire on one of the cylinders that reduces it's performance. The car still works, but does not have the performance that was advertised by the manufacturer. You'd feel lucky to have been sold a defective product just because the company had the opportunity to put more resources towards making more money doing something else? It was their choice to sell and support the lower end product!! They're obligated to keep their word if they're accepting payment for it! I don't understand, what's wrong with being upset that you paid full price and then received a thing that wasn't as advertised? They should have reduced the price, or better yet fixed the issue, if they knew it didn't meet the specifications they advertised it as having! That feels pretty reasonable!
I'm saying that you should be grateful to buy a $1,000 car when this company can sell $25,000 cars all day.

If you are not happy with the quality you get, you should stop buying the $1,000 car.

See how that will do for you.
lexluthermiesterSo no, you're not done then. Citing Reddit does not help your silliness. Cite something with actual merit and credibility or just stop. Additionally, sidestepping the issue will result only in more mocking.. Just stop.
nVidia being a public company, its information is accessible all over the internet.

www.visualcapitalist.com/nvidia-revenue-by-product-line/

Your desperate struggle against reality is pathetic, really.
Posted on Reply
#365
JustBenching
MacZAt that point, when you are so much out of touch with reality ...

pcmasterrace/comments/1awtso6
Gaming has become a rounding error for nVidia.

Maybe after cryptocurrencies and AI, you should understand that GPUs, the 'G' notwisthanding, are very useful SIMD processors that are and will be used for much more than just gaming.

This will continue and amplify.
None of that is an excuse for advertising x amounts of rops and delivering (to 0.5% of products, or however much they claimed) a numbers of rops that is smaller than x.

Yes, we are blessed that Jensen still makes gaming gpus, but that's completely irrelevant to the issue at hand.
Posted on Reply
#366
lexluthermiester
MacZYour desperate struggle against reality is pathetic,really.
So is your high school level of Economic theory, which is not how the real world works. As stated above, your argument is meritless and lacking all logic. It's not worth further debate. Bye.
Posted on Reply
#367
MacZ
lexluthermiesterSo is your high school level of Economic theory, which is not how the real world works. As stated above, your argument is meritless and lacking all logic. It's not worth further debate. Bye.
You forgot your arguments ... ah, yes : you hadn't any to begin with.
Posted on Reply
#368
QUANTUMPHYSICS
Nvidia can do whatever it wants.
AMD is laughable competition.
Nvidia has a waiting army of incels with credit cards to spend whatever they can just to get one of these cards on launch day.
Posted on Reply
#369
trsttte
MacZAt that point, when you are so much out of touch with reality ...

pcmasterrace/comments/1awtso6
Gaming has become a rounding error for nVidia.

Maybe after cryptocurrencies and AI, you should understand that GPUs, the 'G' notwisthanding, are very useful SIMD processors that are and will be used for much more than just gaming.

This will continue and amplify.
I'd hardly call more than 15% a rounding error but what you're not smart enough to see - and nvidia clearly is - is that the consumer market is always there. Through several ups and downs on the other business areas, consumers are there buying their stuff. They also don't require 24h 7days a week support, or custom made solutions, or long term payment plans, or any number of other requirements. They just buy their stuff, use it for a while, and then buy it again.

The data center market is booming right now - and that's not just gpu's, they also own mellanox among other providers they bought along the way - but it will dry up eventually when the AI buble finally bursts (no one seems to have any idea how to monetize it) and they have competitors mounting from all sides. They need the consumer market to keep going or they'll die when that happens.
Posted on Reply
#370
MacZ
JustBenchingNone of that is an excuse for advertising x amounts of rops and delivering (to 0.5% of products, or however much they claimed) a numbers of rops that is smaller than x.

Yes, we are blessed that Jensen still makes gaming gpus, but that's completely irrelevant to the issue at hand.
No it's not because :
1 - AI GPUs are prioritized big time.
2 - Therefore gaming GPUs aren't
3 - Therefore there is much less gaming GPUs produced
4 - As a result the quality suffers and/or the price of gaming GPUs increase.
Posted on Reply
#371
lexluthermiester
MacZYou forgot your arguments ... ah, yes : you hadn't any to begin with.
No, I did not. They were, and try to follow along here, deliberately not present. :rolleyes:
Posted on Reply
#372
MacZ
lexluthermiesterNo, I did not. They were, and try to follow along here, DELIBERATELY not present. :rolleyes:
Yes, I understand you have DELIBERATELY no arguments, because you CAN'T think of any.
Posted on Reply
#373
JustBenching
MacZNo it's not because :
1 - AI GPUs are prioritized big time.
2 - Therefore gaming GPUs aren't
3 - Therefore there is much less gaming GPUs produced
4 - As a result the quality suffers and/or the price of gaming GPUs increase.
I said nothing about the price. If they can't make enough 5090s with 176 rops then cut them all down and sell them as 168 rops, problem solved. You don't ship products with different specs under the same name. Why are we even arguing about something as fundamental as that?
Posted on Reply
#374
MacZ
trsttteI'd hardly call more than 15% a rounding error but what you're not smart enough to see - and nvidia clearly is - is that the consumer market is always there. Through several ups and downs on the other business areas, consumers are there buying their stuff. They also don't require 24h 7days a week support, or custom made solutions, or long term payment plans, or any number of other requirements. They just buy their stuff, use it for a while, and then buy it again.

The data center market is booming right now - and that's not just gpu's, they also own mellanox among other providers they bought along the way - but it will dry up eventually when the AI buble finally bursts (no one seems to have any idea how to monetize it) and they have competitors mounting from all sides. They need the consumer market to keep going or they'll die when that happens.
As I explained futher up, GPUs, notwithstanding the 'G', are SIMD processors that are very useful in a lot of different applications. After crypto and AI, there are lots of place in research and production where they are used. And there is NO case where they will somehow revert to just being used in gaming.

As for AI being a bubble, did the internet cease because the dotcom bubble burst ? If you're not sure, I'm asking on the internet right now.

And AI is much more than just LLMs.

Crypto calculations on GPUs were unnecessary, but GPUs as SIMD processors will continue to be used more and more for other things than gaming. Dreams of its usage drying up are just that : dreams.
JustBenchingI said nothing about the price. If they can't make enough 5090s with 176 rops then cut them all down and sell them as 168 rops, problem solved. You don't ship products with different specs under the same name. Why are we even arguing about something as fundamental as that?
Just another fact that aligns with my theory.

As I said, if you are unhappy with the quality, you always have the option not to buy said GPU.

I don't think nVidia will be bothered that much.
Posted on Reply
#375
MxPhenom 216
ASIC Engineer
bearClaw5If people are dumb enough to pay $2000+ for a 70 series card, more power to nvidia to take advantage of them.
A 5090 is a 70 series card?
Posted on Reply
Add your own comment
Feb 23rd, 2025 09:03 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts