The challenges, rewards, and realities of a two tier console launch

I mean when Anaconda Pro hits in 2 years...

 

Then it will look bad for Sony except your last question.

Then it will compete with PS5 Pro+ or PS5 Pro 2

 

But what if the Lockhart Pro is stronger than the Anaconda Slim?

 

I hated the idea of a 2 tier console launch, but then I remembered how this past year, interviews with Matt Booty and Phil Spencer, both of them talked about supporting Xbox One consoles (One, S, and X) through Scarlet's launch and how games would run on both platforms so they weren't reserving games from internal studios for Scarlet's launch. That got me thinking, what if Lockhart is really just a new One X, with maybe a few "modern" features (newer HDMI, USB, Wifi standards) and slightly boosted performance along the lines of One to S, and maybe some cost reductions. It isn't uncommon for companies to release a new version of an old console after a new console launch. There Xbox 360E, PSone, NES Top Loader, Genesis 3 are all examples of this. Genesis 3 was just a year before Dreamcast! It makes a lot of sense for MS to want to clean up their product stack, so if Lockhart is essentially a replacement for X level performance and they sunset One S/SAD/X while they launch a true next gen console that would do that.

So what is Lockhart is just an Xbox One X2.

 

I think if Lockhart is anything but a cut down Scarlet (same CPU, SSD, memory, and only cut down GPU), I think it will be a failure. I don't buy MS supporting Xbone S and X much longer, the CPU and hard drive is just too much of a hinderance. I don't think it will be possible to deliver a next-gen experience on a Jaguar level CPU. Playstation would have too much of an advantage.

One thing that I thought of when contemplating Lockhart/Anaconda, is that it should work fine as long as Anaconda is hitting 4k resolution in a game then Lockhart could deliver 1080p-1440p given the rumored specs. However, late in the gen, what happens when Anaconda can't achieve 4k, but some devs are pushing graphics so much that resolutions drop to 1440p with reconstruction on the high tier platform. Is Lockhart going to be rendering those games at 720p?

I like the idea of different tiers, but the more I think about it, planning and building out support for a higher tier that will launch 3 years from now seems like a better idea.

 

What I'm saying is that it might not deliver a next gen experience because it isn't designed to. It may just be the late release, cost reduced type console to give late adapters somethings to purchase. Microsoft seams much more inclined to get people subscribing to live and game pass than they are doing anything else, because the real end game of set top boxes is to lock people into an ecosystem.

 

Other than it being different storage sizes like has been done before I don't see how anything good can come out of it. How much do you really save by having a smaller GPU part in the APU? They can't really cut down on the CPU because that would gimp the other machine.

The only way I see it working is if the cheap machine is the main console that competes with the competition and the other is a much more expensive machine that's just there for the people who must have the best and it will have high margins much like the elite controller.

 

Or putting it another way, Anaconda would be the next-gen system, but Lockhart isn't. Lockhart is a current gen system.

Not an unreasonable idea, but it clashes with the rumours that place Lockhart firmly as a next-gen system.

 

What's the point of having Lockhart when MS already has 1X? I'm sure the stronger GPU in the X would compensate for the weaker CPU and slower SSD to an extent that the difference would almost be redundant.
I'm not jumping into any conclusions yet about how much Lockhart could potentially hold back Anaconda's graphics output, but it does seems like developers are sharing the sentiment.
Here's a quote from Jason on Splitscreen podcast taken from Era.

 

I am in line with iroboto thinking.
Which ratio of good Anaconda APUs you have?
If is good enough, then all Anacondas full ahead. If not, a caped, underclocked APU on a less populated motherboard in a slim box (less heat) and Lockhart is born.
Game code is the same for both, one with high settings and the other with medium settings or higher/lower resolution.

 

The whole idea of 2 SKU's @ launch seems strange anyway's, for both ms and sony. One powerfull machine where games will be optimized for will do for me. Maybe a pro later on, if even that, as the only upgrade with the current are resolution, mostly, if anything.

 

Isn't bytewise compatible, lacks features, lacks user experience (SSD instant play) and lacks CPU power. Devs don't want to target another architecture. They won't even want another target spec either. Designing a next-gen game for a 10 TF next-gen console and also a 4 TF one is a pointless burden for devs who gain nothing from the added workload. Though I suppose the idea is they're doing that anyway as every game is cross-platform with PC. But those wanting to make a next-gen only experience shouldn't be hampered with a weak target spec foisted upon them and should be allowed to go Anaconda only in the case of Lockhart being a real thing.

 

In that case, MS could opt to not release anything new next year, or a One X upgrade (lockheart).

 

Personally I'm more in favour of the idea of an updated XB1 of some sort for the low end and XBNext as a separate new generation. That or the bottom tier being the base, next-gen console and there being a slighter better high-end tier, with devs just targeting the bottom end and the upper end getting extra gravy. I don't see any sense at all in a next gen starting with two SKUs, one of which is 40% the capabilities of the other. How do you manage production in favour of which SKU without over-stocking one and under-producing the other? How do you develop and showcase games for the top-end SKU?

I suppose a legitimate explanation would be this as the start of MS ending console generations and going with a rolling model. They'd kickstart with a $300 low-end box and a $500 high end, and then after a few years, the $500 box is now the $300 box and there's a new $500 box, and MS keeps rolling updates every few years. If the hardware is suitably abstracted now, devs will get a fairly transparent development environment and can simply have a bunch of flag on performance, and will be able to choose which 'gen' of hardware to release to.

 

I'm not sure I know what "next-gen" means anymore. The rumors also suggest a 4TF GPU, which based on Eurogamer's / Digital Foundry's flop vs flop benchmarks, Navi is about 30% faster. That would put Lockhart at 5-6PTF (Polaris TeraFLOPs) if it's 4 to 4.5 Navi TeraFLOPs. Or right in the ballpark of One X's GPU. Even with a faster CPU, One X is consistently held back by it's GPU because of the resolutions it's rendering at. Lockhart would be as well.

Although, thinking about it more, if Lockhart can match One X in performance for One X titles, and then new titles use VRS and hardware ray tracing, that might provide a graphical increase that's acceptable. But launching 2 consoles day and date where the premium console is 2.5x more powerful (in shaders anyway) isn't something that I can understand, unless one of them isn't considered of the same generation.

 

Graphical increase of Lockhart over X1X is immaterial as X1X owners will be buying Anaconda. If Lockhart is seen as an early-upgrade path for XB1 owners, it just needs be notably better than XB1 in games. RT, better CPU, SSD experience, and better efficiencies may be enough, although it'd be the lamest console upgrade in history. Perhaps because the cost of true next-gen machine is considered prohibitive, at $500+, and they aren't willing to take a loss, the plan is a $300 base model just for a next-gen that's not priced out of starting? Juggling two SKUs is still a business nightmare though IMO. I'd think it better to go $500 launch of one SKU and just let that gen grow in its own time, letting last-gen fill out the lower end of the market like usual.

 

X often targets 4k but we may assume Lockhart will target 1080 which means 4times the flop per pixel. This seems a fairly decent jump even before the new features.

A cheap box might get fast market penetration which is a benefit to developers especially at launch, maximising the potential market.
Same APIs , same CPU,.same GPU features , that seems fairly minimal extra work.

 

Microsoft reveal event in june:
- two glass cases placed side by side and both covered with a sheet
- enter phil spencer
- the road so far
- the crowd roars
- phil spencer grab both sheets
- the crowd start to orgasm
- phil spencer removes the sheets and reveal in one a console in the other a controller
- the crowd cancels the gold subscription

 

If the rumors are true that MS would like the baseline to shift to 60 FPS for games (1440p for Lockhart and 4k for Anaconda) then the CPU has to be a lot more powerful than what is in XBO-X. NOTE: this doesn't mean that developers have to shift to a 60 FPS baseline or that they will, just that it's possible that MS would like that to happen.

If you look at the developer interviews for Gears 5, the biggest challenge was having to go through and somehow make 60 FPS in Gears 5 SP possible with the weak CPU that current gen consoles use. Most of the effort was put on moving as much stuff as possible off of the CPU or finding alternate ways of doing things to reduce how reliant those effects are on the CPU.

That's effort that a lot of developers either can't afford to do or aren't capable of doing. Having an excess of CPU power ensures that it's far easier to hit a 60 FPS target. And for really talented and/or ambitious developers, it opens up the possibility of 120 FPS for use with temporal reconstruction for some amazing graphics.

Regards,
SB

 

Plus, long term Lockhart will probably be cheaper to make than the X1X. No hdmi in, no optical, simpler mobo for a narrower bus and less power draw, cheaper cooler etc etc.

Biggest worry though, imo, is that pre E3 there were rumours of 2x faster than current gen for Lockhart, 4x faster for Anaconda. The 4x figure ended up being real, but turned out to be only about the cpu.

Would this mean Lockhart only has 4 cpu cores in a single Zen 2 ccx? Let's hope not!