It is not strange to see some end users submitting issues when a new GPU or CPU launches, but there’s early information suggesting that some RTX 3080 and RTX 3090 GPUs have a steadiness issue when they press in close proximity to-to or over 2GHz. Most stories have targeted on 2GHz, but at the very least one particular user reported his GPU was under that clock.
Experiences have started popping up on-line of steadiness issues very likely tied to GPU increase frequencies. Recognised-impacted types contain the Zotac RTX 3080 Trinity and the MSI RTX Ventus 3X OC. MSI’s Gaming Trio is pointed out, as is EVGA’s RTX 3080 XC. These stories make roughly the similar assert: The GPU crashes in one particular or more titles, generally at all-around 2GHz. Cutting down clock speeds can take care of the issue.
Igor of Igor’s Lab thinks he has an clarification for the issue. After referencing Nvidia’s files for the RTX 3080’s PCB structure, he writes:
The BoM and the drawing from June leave it open up no matter if massive-spot POSCAPs (Conductive Polymer Tantalum Strong Capacitors) are utilised (marked in red), or rather the relatively more highly-priced MLCCs (Multilayer Ceramic Chip Capacitor). The latter are more compact and have to be grouped for a better ability.
The regions blocked off with bins and the set of 10 green rectangles are all energy rails. If the RTX 3080 and RTX 3090 are getting fed soiled energy, it would clarify why the playing cards destabilize and crash at higher frequencies.
Zotac, for illustration, utilised POSCAPS for all six rails:
Nvidia utilised four POSCAPS and two MLCCs, as shown under. Igor notes he’s been unable to crash his Founders Version GPU, implying that this may perhaps be the trigger of the crashing bugs. We don’t know for selected that the 3080 FE doesn’t have this issue, but so much no one reporting an issue seems to have an RTX 3080 FE.
We don’t know how numerous playing cards are impacted by this issue. Even if employing POSCAPs in its place of MLCCs is the trigger of the issue, it doesn’t instantly observe that just about every POSCAP-equipped product has a issue. It appears to be more very likely that a selected share of POSCAP products would have troubles than that Nvidia would fail to observe it experienced misinformed all of its AIBs and handed them terrible production files. The former could trigger some elevated restore prices and grumbling, when the latter would call for the restore of just about every RTX 3080 and RTX 3090 created to-date.
If you operate into this issue, the initially matter to test is lowering the base clock on your GPU. Customers are reporting that 80-100MHz usually performs. While overclocking your GPU is dangerous if you don’t know what you’re undertaking, it is challenging to muck up a portion by functioning it slower.
This condition is going to evolve about the following several days as Nvidia discusses it with AIBs and investigates how ideal to deal with the issue. I recommend maintaining an open up thoughts as much as what the potential trigger could be, not for the reason that I question Igor, but for the reason that from time to time production evaluation reveals extra issues that weren’t earlier recognized and couldn’t be seen from surface area-stage assessment.
No subject what the issue turns out to be, Nvidia or its partners have a mess they’ll will need to clear up when also working with severe card shortages. We’ll continue to keep you posted on regardless of what the trigger turns out to be.