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

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

#276
Bwaze
JustBenchingThe samsung driver in question (980 pro) had no hardware issues actually, it was just the software misreporting. I have one btw, haven't updated the software, it's misreporting as 80% health but it's just a software reading error, the driver is fine (have it since 2021) As I've said - stop reading headlines and youtube clickbaits.
Yes, that’s the firmware fix. But if you let the misreporting reach a certain level, drive was inoperable and unfixable, for some it was working in read-only mode, for others it was bricked and unreadable. But clearly you have done your research, and in all samples you tested it was a 100% statistical non issue!
Posted on Reply
#277
modmax
for 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
Posted on Reply
#279
Vayra86
JustBenchingYouve been doing nothing BUT bashing nvidia. Every single damn of your posts. Ever. Still I haven't seen you mention the RX480 ruining your mobo and causing fires, the 5700xt black screen drivers issue, the 7900xtx cooler issue. But here you are, in an nvidia thread to crap about nvidia, and then complain that people are bashing AMD. :banghead:
And 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
Posted on Reply
#280
10tothemin9volts
https://www.techpowerup.com/332884/nvidia-geforce-rtx-5090-spotted-with-missing-rops-nvidia-confirms-the-issue-multiple-vendors-affected-rtx-5070-ti-tooUpdate 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."
In this case GPU-Z and other tools who rely on an internal database if the NV driver is not installed, should fix this by e.g. not displaying the actual ROP count until the NV driver is installed and display it on the GUI in the field, so the user knows, but if they chose to still display the ROP count, it should be clearly noticeable to the user by e.g. a different field color and/or a mouse over should say that to see the actual ROP count, the NV driver must be installed.
Posted on Reply
#281
R0H1T
There's probably no way to do that & why should GPU-z not do its work if NVidia messed up their hardware :wtf:
Posted on Reply
#282
Bomby569
Good on nvidia for this response, we can only hope 99% of the problems had this fix
holly shit they simply don't check anything, it was such an easy thing to detect
Zotac had nothing to do with this
Posted on Reply
#283
john_
NVIDIA
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.
Yeah, sure 0.5%. 0.5% connectors melting, 0.5% cases where ROPs are missing and as I said above, the option of a quick RMA. Of course who would RMA their card when the replacement unit could be 2-6 weeks away? If there is one available. Imagine people getting their money back and then having to pay the new much higher prices compared to MSRP. I mean who ever got that card will convince themselves in less than a minute to say "Fk it" and keep that card.

Well played Nvidia, well played.
And the same people are those who see a chart where Intel or Nvidia wins by 1% and rush to post "HAHAHA AMD slowwwwww!!!!!".
Posted on Reply
#284
Contra
10tothemin9voltsIn this case GPU-Z and other tools who rely on an internal database if the NV driver is not installed, should fix this by e.g. not displaying the actual ROP count until the NV driver is installed and display it on the GUI in the field, so the user knows, but if they chose to still display the ROP count, it should be clearly noticeable to the user by e.g. a different field color and/or a mouse over should say that to see the actual ROP count, the NV driver must be installed.
I would suggest specifying the value Undefined without a driver, and for this label to work as a link to the TPU database so that the user can see the reference values.
Posted on Reply
#285
R0H1T
Contrathis label to work as a link to the TPU database
Bad idea, I know I wouldn't want to open a link in a browser just for this.

The only difference I'd have is to make it clear that these are "expected" values & the actual number can be determined only after installing the drivers.
Posted on Reply
#286
Prima.Vera
This is absolutely unacceptable for the cards of those price level.
nVidia should offer a discount of at least 30% for future products to those affected.
Posted on Reply
#287
Jeager
Look like some customers were lucky here and got their hands on the real 5080 :) A quick bios update will solve this !

As for the money dif, no refund is needing since they were able to spend that much in the first place (and they are nGreedia best friend/supporter after all)
Posted on Reply
#288
notoperable
BwazeI’m talking about how some users applaud the exemplary Nvidia’s response, which is a correct one - but in their position perhaps the only one possible. And I have written a possible other way out without admitting their fault - as we have seen other companies try before.

A cynical take on how tech companies deal with such issues.

Wanna hear another such story? Largest SSD maker, Samsung, had a drive that was failing unless you updated it’s firmware. No ”if’s” or “why’s”, just “when”. But rather than making a pubic statement that users should do that, they issued their solution quietly in tech help, which nobody sought until it was too late - how many users lost their valuable personal data so that Samsung maintained their perfect public image?
Wtf? Correct one? Its their f*****g responsibility, not a privilege - its called warranty and you say as if I need to bend over and be thankful that the all mighty blessed me with divine intervention
Posted on Reply
#289
Quicks
Trillion Dollars company, but can't get the basic quality control right. But seems like a norm these days, all around the world. They just push out products and see what's a hit or a miss.
Posted on Reply
#290
notoperable
QuicksTrillion Dollars company, but can't get the basic quality control right. But seems like a norm these all around the world. They just push out products and see what's a hit or a miss.
Who gives a shit about Q/A if you have enterprise wait list that's a year into the future
Posted on Reply
#291
efikkan
BwazeSo an issue they haven’t noticed before and which requires some knowledge from the user to even identify it (and has to actively look for it) already got an assessment on how many products it affects, and it’s of course extremely low number?

The fact that right now there is a very small number of these cards around, but affected cards were found from several AIBs and models shows you that 0.5% number is pulled from the arse.
Let 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.
Posted on Reply
#292
LittleBro
1. Since QA failed badly in case of dies for $2k GPU, their 0.5% of affected units is just rough estimation. Anyway, it seems user must be lucky enough to get a card that runs up to the specification these days. Shitshow.

2. Nvidia sold out of spec dies to AIBs. Maybe whole scarcity of RTX 5000 lineup in retail is connected to this problem. RTX 5070 amd 5060 are delayed because of performance bug. There might be a correlation.

3. So missing 8 rops kills 5% of raster perf. With all that physical space that AI units occupy in die, how about they used that space for another ROPs instead to see another meaningful increase of perf. in native without AI workarounds?
Posted on Reply
#293
Vya Domus
"Anomaly" lol, like it's something supernatural.

"Sorry guys we shipped out chips with portions of them disabled and strangely enough they had the correct firmware on them so the thing doesn't crash, weird right, don't know how that happened."
Posted on Reply
#294
azrael
NVIDIA's statement downplaying the issue reminds me a bit of a game I play, Elder Scrolls Online. Every so often the servers have issues and then you see a banner saying "Some people are experiencing issues..." when in fact ALL people are. Some companies have turned downplaying problems into an art form. It seems it's better to mislead than taking the actual blame.
Posted on Reply
#295
LittleBro
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.
They knew and they will tell users to don't mind, just turn on DLSS.
Posted on Reply
#296
efikkan
LittleBroThey knew and they will tell users to don't mind, just turn on DLSS.
It would be interesting to know whether these cards even have separate device IDs too. Having multiple iterations branded as the same product is completely normal for Nvidia, and most of the time the performance should be the same or even slightly better.


This is what's listed in ListDevices.txt provided with the 572.47 driver:
DEV_2B85 "NVIDIA GeForce RTX 5090"
DEV_2B87 "NVIDIA GeForce RTX 5090 D"
DEV_2B85&SUBSYS_53051028 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_A7611025 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_530717AA "NVIDIA GeForce RTX 5090"
DEV_2B87&SUBSYS_530817AA "NVIDIA GeForce RTX 5090 D"
DEV_2B85&SUBSYS_53001462 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_53011462 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_53021462 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_53031462 "NVIDIA GeForce RTX 5090"
DEV_2B85&SUBSYS_53061462 "NVIDIA GeForce RTX 5090"
DEV_2B87&SUBSYS_53001462 "NVIDIA GeForce RTX 5090 D"
DEV_2B87&SUBSYS_53011462 "NVIDIA GeForce RTX 5090 D"
DEV_2B87&SUBSYS_53021462 "NVIDIA GeForce RTX 5090 D"
DEV_2B87&SUBSYS_53031462 "NVIDIA GeForce RTX 5090 D"
Posted on Reply
#297
Hecate91
Nvidia calling this issue an "anomaly" is downplaying the issue, of course they knew of it the whole time as Nvidia are the ones selling dies to the AIB's.
But they've already gotten away with it by calling it some weird mistake, when it isn't possible for out of specs dies to go out on accident, and they know 5090 users won't send their cards back because the supply has been artificially limited.
Posted on Reply
#298
Super XP
This is a perfect opportunity for AMD to screw up a perfect opportunity.
I'll give them the benefit of the doubt, lets see if the RX 9070XT and non XT are priced accordingly. To take advantage of Nvidia's screw-ups.
Hecate91Nvidia calling this issue an "anomaly" is downplaying the issue, of course they knew of it the whole time as Nvidia are the ones selling dies to the AIB's.
But they've already gotten away with it by calling it some weird mistake, when it isn't possible for out of specs dies to go out on accident, and they know 5090 users won't send their cards back because the supply has been artificially limited.
I wouldn't be surprised if Nvidia was actually doing this. lol
Posted on Reply
#299
JustBenching
Super XPI wouldn't be surprised if Nvidia was actually doing this. lol
To what end? There is nothing competing with the 5090 anyways. Scratch that, there is nothing competing with the 4090 anyways. Nevermind, scratch that, there is nothing competing with the 5080. Actually, keep scratching until we go through the 4080S, the 4080 and the 5070ti. The last one, nvidia might have some competition.
Posted on Reply
#300
Bwaze
Are 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.“
Posted on Reply
Add your own comment
Feb 23rd, 2025 02:53 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts