My friends and I got a few doubts about SLI, so I hope someone can enlighten us with some answers.
Our discussion is primarily focused on exactly what is the definition of the "same type of card"?
I guess it is obvious they have to be the same GPU, I doubt few people would expect to be able to hookup an Ultra to a NU.
Then the question gets more interesting.
1. Do the cards have to be from the same vendor?
2. Even if it is the same vendor, chances are the cards might change from time to time due to costdown or just supply of raw materials. In that case, are these still considered the same card? I wouldn't want to buy one card expecting to hook up a second one next year, then realize I have to look on eBay for a second hand XXX brand XXXX card with Memory modules xxxxxxxxx or with whatever specific chip of TMDS, VIVO.
3. Under what conditions might nVidia abandon SLI? We are afraid by the time we get the second card, nVidia thinks SLI is stupid or unsuccessful for whatever reasons and stop supporting the architecture.
4. Does SLI put extra strains on the power supply? It is obvious that two cards are definitely going to consume more than 1, but does the PCI-E architecture provide any power supply improvements so we don't need to connect 2/4 MOLEX connectors to the PSU?
Thanks for any input.
Our discussion is primarily focused on exactly what is the definition of the "same type of card"?
I guess it is obvious they have to be the same GPU, I doubt few people would expect to be able to hookup an Ultra to a NU.
Then the question gets more interesting.
1. Do the cards have to be from the same vendor?
2. Even if it is the same vendor, chances are the cards might change from time to time due to costdown or just supply of raw materials. In that case, are these still considered the same card? I wouldn't want to buy one card expecting to hook up a second one next year, then realize I have to look on eBay for a second hand XXX brand XXXX card with Memory modules xxxxxxxxx or with whatever specific chip of TMDS, VIVO.
3. Under what conditions might nVidia abandon SLI? We are afraid by the time we get the second card, nVidia thinks SLI is stupid or unsuccessful for whatever reasons and stop supporting the architecture.
4. Does SLI put extra strains on the power supply? It is obvious that two cards are definitely going to consume more than 1, but does the PCI-E architecture provide any power supply improvements so we don't need to connect 2/4 MOLEX connectors to the PSU?
Thanks for any input.