Originally posted by: Zebo
Virtual Larry: Re: price/performance & SLi
I know a little about price/performance and you're forgetting the initial qualifier in your analysis of SLi and that is: the minimum level of performance would satisfy thier requirements before applying any kind of price/performance matrix. w/o that starting point you should be using on-board video since it offeres supreme price/performance.
Well, you're right, I'm using my own internal metric for the minimum performance baseline, basically the minimum (lowest-cost)
gaming card, that is
capable of playing current game software, at at least minimally-reasonable frame-rates and feature levels. Not necessarily the highest levels of eye-candy and resolution and frame-rates though. (Note that implicitly excludes "non-gaming" cards and integrated video solutions.)
And to Ben's credit, he kept talking about the high-end, and I was kind of ignoring that, but he does have a point in that, in cases in which there
is no single-GPU/card consumer solution, then at least at the high-end, SLI is the only way to fly.
So to break it down, now that we have current-gen PCI-E based SLI tech in the market, it has caused a diversification, really, and added on another level of performance, above high-end (gaming) single-GPU/card solutions, lets call it ultra-high-end (UHE).
It just so happens that NV enabled SLI for mid-range (6600-series) cards too, so that creates a bifurication of the mid-range gaming hardware level too, splitting it and adding on an "ultra-mid-range" solution, which just happens to be similar in performance to current single-card high-end solutions, more or less.
Originally posted by: Zebo
No, you don't. Why? Because you had a rough idea what type of performance would satisfy your requirements, compared competing products, purchaesed the one with the best p/p ratio, and on-board video was'nt even in the picture. Similarly, for some, a single XTPE/Ultra is'nt either. Perhaps they want 20x15 4x8x, as thier starting point. SLi, not only offers a good value, it's the only game in town.
Yes, but in that case, price/performance doesn't even really figure into the picture in terms of hardware selection - the hardware was chosen on a performance basis alone, since there are no other consumer-level gaming card solutions at the UHE level. You can't really compare a P/P graph, if there is only one data-point, and it is located at the top of the chart. That's basically what I was saying to Ben. I think (correct me if I'm wrong), that his argument (involving a $10K budget), was to point out that there aren't any single-GPU/card solutions in the consumer UHE market space.
What I'm basically arguing against, here, is the possibility (due to pressure from NV, both on the mobo and software companies), that the former single-card mid-range market segment will fade away, supplanted by the ultra-mid-range instead, as the lowest-end hardware platform that modern games will support. This theory is seemingly supported by the lack of development and product diversity currently available in the mid-range gaming market segment, and is why I brought up with Ben the fact that, performance-wise at least, the prior-gen high-end cards were having to fill this market niche, because of the lack of new products in the pipeline for this segment.
I feel that will shortchange the consumer, really, because that means that NV has them "by the short hairs", to extract more money from them, and at the same time, allows them to spend less on R&D, to advance the tech for single-GPU/card solutions. (Is this NV's solution to remain profitable, and slow the incredible GPU "arms race" caused by one-upmanship between two major rivals, in the face of a six-month product-refresh cycle?) I know that it may be taking a toll on the companies' resources, but by the same token, up until now at least, customers have benefited from the intense competition. Well, if you ignore the fact that it also tends to create more "paper launches" than products, when said companies' can't deliver product on time, and still have to create the market illusion that they are competive in spite of that.
I do agree, the market will (eventually) sort this all out, perhaps consumers will vote with their wallet and not buy into SLI systems, at least not at the ultra-mid-range. At the ultra-high-end though, there really isn't anything else, so I can't really advocate against someone spending their money where they want to. If they want to pay extra for that sort of performance, feel free, but I don't want to be dragged along with the rest of the mid-range market segment, if ultra-mid-range supplants mid-range, and SLI becomes a de-facto requirement for gaming. Thus the mention of the "SLI tax".