I have one quadro k4000 and Tesla c2050. The problem is that when i hook them up the right way, windows cant recognize the Tesla card, but if i hook them up the wrong way, it does recognize them both, but when i run Maximus Configuration, it tells me there is an error and it wont show up the Tesla card.
It’s not a valid Maximus config. You need a kepler-based Tesla card combined with the k4000 to be a valid maximus config.
ah crap… so is this config useless? ANd will any from the K series work fine? CAn i change the quadro instead? Sorry for all the questions
The document I linked (maybe you should read it) suggests that a Kepler K20 card be used for a proper Maximus gen2 config with your Kepler K4000 Quadro card.
There was a Maximum Gen1 config based on Fermi (your c2050 is Fermi generation). I don’t remember the config details, however it probably would have used a Fermi Quadro combined with a Fermi Tesla.
Yeah i saw it. Better yet i saw this CUDA - Wikipedia but it suggest K10 there
Sad, but the funny thing is, it still kinda works. Aside from that weird error, Nvidia control center recognizes them both and gives me all the options. Didnt really assumed that it mattered. Flew right pass my mind. I feel so dumb…
Hi guys! I have a similar issue. I’d like to know if it’s possible to achieve a Nvidia Maximus configuration with a Quadro 6000 and a Tesla M2090 on a Windows 10 system. Another possibility would be a Tesla C2075 but I have a really good deal on the M2090. I know these cards are dated but due to the high price of newer architectures, specially in my country, this is what I can afford at the moment. Regards
You’re going to have a miserable system if you put an M2090 in a workstation. It will not cool the GPU properly. This has been discussed extensively elsewhere on this forum, if you want to do some research. Regarding Maximus on Fermi (which seems to be what you are asking about) I’ve shared above some links that may be worth a read.
Thank you very much for your time. When you say miserable, you mean about performance? Poor cooling with the M2090 is something that I was expecting and I even considered puting a cheap custom cooling kit on it since my case is well cooled with 6 120mm fans (3 in +3 out). But my question was more about the possibility of Nvidia Maximus on a Win 10 system (didn’t find people talking about it), the performance of those configurations I mentioned and if a Fermi GF110 + Fermi GF100 is something possible. When I talk about performance, keep in mind that I plan to use the workstation for light full HD video editing, color correction, light 3D work with After Effects and something like that. I’ve read the “Nvidia GPU Aplications catalog” to see what apps take advantage of multi GPU tecnology and I’m pretty aware of that. I just aim to have a fluid, real time workflow. Thank you again
I mean miserable from the standpoint that the M2090 under heavy load will likely overheat and drop off the bus. That’s why I said “It will not cool the GPU properly.” If you have some scheme to address that, great. You’re welcome to try whatever you wish of course. I’m not sure you are grasping the difference between typical workstation case cooling, and ducted fan forced air cooling of a passive M2090 heatsink, but have at it.
I don’t really know if Fermi Maximus will work under win10. Certainly win10 was not around when Fermi Maximus was conceived, and NVIDIA has made no statements about win10 support for a Fermi Maximus config, AFAIK.
I think you should be able to find an NVIDIA GPU driver that supports e.g. both a C2075 and Q6000 under win10. Beyond that, I don’t make any claims. My statement about cooling is merely a caveat. In a forum like this, I have no idea of the knowledge base, background, or expertise of most posters. So I want to point out common mistakes. And inadequately cooled server GPU processors is a common mistake, as evidenced by various postings on this and other forums.
Thank you for sharing your knowledge, sir! You certainly helped a lot.
Nvidia Maximus is not compatible w10, I already tested