Slashdot Mirror
NVIDIA GeForce GRID Cloud Gaming Acceleration
Vigile writes "NVIDIA today announced a new technology partnership with Gaikai, an on-demand gaming company that competes with OnLive, to bring GeForce GRID to the cloud gaming ecosystem. GRID aims to increase both the visual quality and user experience of cloud gaming by decreasing latencies involved in the process — the biggest hindrance to acceptance for consumers. NVIDIA claims to have decreased the time for game stream capture and decode by a factor of three by handling the process completely on the GPU, while also decreasing the 'game time' with the power of the Kepler GPU. NVIDIA hopes to help both gamers and cloud streaming companies by offering 4x the density currently available and at just 75 watts per game stream. The question remains — will mainstream users adopt the on-demand games market as they have the on-demand video market?"
Not as long as Comcast controls my bandwidth.
I have to say, some of their comparisons seem... unfair.
Their main chart compares three things: regular "console connected to display", "current cloud systems", and "GRID cloud rendering".
First off, they cite 66ms latency just at the display level, which is definitely at the higher end of the spectrum. But at least they use the same latency in each.
Their cloud/cloud comparisons are also quite suspicious. Reducing encode by 60%, yeah, I can buy that. Reducing *decode* - which, I remind you, is done client-side - by the same amount is also suspicious in light of their "this does not require an nVidia client, it will work with any h.264 decoder" claim.
Then they claim to have reduced network latency, and significantly (75ms to 30ms). Now, I can vaguely see how they *could* - if they can reduce bandwidth usage significantly, they might eke out a bit less latency, but I highly doubt they can more than *double* their compression efficiency. Unless they're doing something crazy like putting a network interface directly on the GPU (one image they show contradicts that theory), I think this claim is also pretty dubious.
The worst one is the "game pipeline" time. While I can believe that a newer, more powerful graphics card can definitely perform *twice* as well as an older one, I can also state definitively both that "you can put that new card in the home console as well" and "new games will expand to use that power, leaving you back where you started at 100ms render times". The former I can state because they've *already* released Kepler-based cards (to rave review, although my own seems to be backordered); the latter I can state because that's how the industry has worked since at least the late 70s.
Long story short, they seem to be doing some *extremely* unscientific, biased comparisons. Do they probably have something here? Yeah. Is it literally going to be as good as an actual console (or better yet, PC) connected directly to the console, as they claim? No.
Movies are not games. With movies, you often tend to watch once in a few hours and then not touch again, so services like Netflix are much more efficient. You also don't need powerful hardware to run movies in real time or stream them.
Games need hardware, they need good latency, but most of all games are fundamentally different from movies. You're interacting with the game, often for extremely long stretches of time. I can't seriously think for a second that services like Gaikai or OnLive would be cheaper than buying the game straight away, unless you're uncertain you'll be playing much. I can see them being good demo-like services, but not full-blown gaming services.
Finally, there's one critical element that makes PC gaming, which is what's targeted by these services primarily, unique: the games can be modified. Mods have breathed life into games that deserved it, fixed games that were broken, improved games to perfection, kept games alive for years. They're the one thing that PC gaming has as a crucial advantage over just about any other (closed) model. You can't mod Gaikai games. Say goodbye to those amazing Half-Life 2 or UT2004 mods.
Oh yeah, and say hello to gatekeepers getting to choose which games are available. Dominance of streaming game services would be bad news for indie developers, since hardware would slowly get deprecated and not replaced. With the current market, you don't have to be on Steam to be competitive; the Humbe Bundle more than proves that.
I really hope this won't happen. We don't need more centralization than we already have.
I predict that "GRID Cloud Gaming" will end up being mainly a way to make gaming more expensive. Somehow, someway, it's going to mean paying more to get the same product - with extra DRM goodness.
Now, I say that without reading the article or knowing one thing about GRID or clouds or what's required to decode...whatever it is they're trying to decode.
And that does not reduce my confidence in my prediction one bit. Nor does it reduce the probability that it is absolutely correct.
You are welcome on my lawn.
Cloud assumes bandwidth is free
Agreed. It also assumes that latency is faster than a typical satellite Internet connection.
on-demand games market
Seriously, I hate seeing the PC become prostituted and sold short. I've watched its decline since last century, where it has gone from number-crunching powerhouse to (almost) just a dumb terminal. Why bother getting a PC at all under these conditions? Let's go back to coin-operated video games. I'm sure the "market" will love that. Except of course it's going to cost you much more than $0.25 per go this time. $10 bills only.
Seven puppies were harmed during the making of this post.
The fear is that certain major publishers will make even their single-player games exclusive to these cloud gaming services, leaving people living outside the service area of wired broadband without games to play.