What are the chinks in XB360's armour?

[Shifty Geezer]

Btw Shifty...I realized after posting that you may have wanted this to be HW related only... and in that case... sorry man

Well I was looking for hardware chinks as opposed to market chinks, though I didn't necessarily make that clear, but it makes no odds really. I agree with quite a few points you raised, at least in arguments if not in 'chinkiness rating'

Regards RAM latency, how does DDR compare with XDR and XeCPU's caching? MS's PS3 comparison rated Cell wayyyyy down because of no true caching, but I think XeCPU is in the same boat - main memory accesses are going to be rather costly and a degree of memory management is needed for the devs to maximise data throughput. [/muse]

 

[blakjedi]

12-16 megs of eDRAM
Would have liked 33% higher bandwidth between CPU RAM AND GPU [ bump from 22GB to 33]
Larger L2 cache - 2MB minimum 3 MB preferable
VMXs that can handle their own threads much like the SPEs in Cell
Faster GPU ~ 700 MHZ

 

[koldfuzion1]

Acert93,
Well, I read the whole analysis, interesting read. Thanks. You should go work for McKinsey & Co. or something if you don't already. I think it's pretty much on the money, maybe I'd quibble with some rankings but that's basically how I see things shaking out as well. Anyway Shifty, sorry for hogging anymore of this technical thread, just wanted to give props to that market analysis which must have taken a 2 hours to write and is as good analytically as any Harvard MBA case study I've read. Now getting back to the technical stuff... [tries to push worms back in can]

 

[AlNom]

12-16 megs of eDRAM

I think 10MB was a good compromise on cost (and maybe die space)


Question: for L2 caches, is it just not possible to have 1.5MB? Some sort of thing against compromising between 1MB and 2MB?

 

[Shifty Geezer]

A couple of megs wouldn't have made any difference. 720p would still need two tiles. Enough eDRAM for 720p + AA would be the next step up.

I did laugh at the 700 MHz GPU though. Some people aren't happy with anything!

 

[darkblu]

12-16 megs of eDRAM

I think 10MB was a good compromise on cost (and maybe die space)

i have to second blakjedi here - 10MB is not sufficient _given_ the mandatory targeted res of 720p, despite all the clever tiling. expect latency issues :?

 

[Thowllly]

VMXs that can handle their own threads much like the SPEs in Cell

Argh! When will people stop comparing the vmx units to SPEs? The VMX units couldn't "handle their own threads", then they wouldn't be VMX units anymore. The VMX unit is just a small part of the core, like the SSE unit in the P4 (I've never heard anybody suggest the SSE units should "handle their own threads")

It makes about as much sense as saying "I really wish the Opterons 16K bimodal global branch prediction counter could handle its own threads..."

sorry for ranting...

 

[Tacitblue]

12-16 megs of eDRAM
Would have liked 33% higher bandwidth between CPU RAM AND GPU [ bump from 22GB to 33]
Larger L2 cache - 2MB minimum 3 MB preferable
VMXs that can handle their own threads much like the SPEs in Cell
Faster GPU ~ 700 MHZ

Won't a larger L2 also have a higher latency though? There's compromises in every choice a designer makes. Not to mention how much extra die space the extra 1-2MB would take up. It would be bigger than Cell. And producability wouldn't be very nice then.

 

[Acert93]

i have to second blakjedi here - 10MB is not sufficient _given_ the mandatory targeted res of 720p, despite all the clever tiling. expect latency issues :?

ATI is on record as saying there wont be any performance penalties for 2x AA (2 tiles) and only 1-5% for 4x AA (3 titles... of course it seems the GPU can write to the eDRAM as it is also flushing the framebuffer to the main memory).

If those estimates from ATI are accurate it would seem that the process is pretty effecient. As tradeoffs go, I am sure there are other places of concern than a 1-5% drop at 720p with 4x AA.

 

[pc999]

I did laugh at the 700 MHz GPU though. Some people aren't happy with anything!

Meybe Fast14 anuncement had gave to us to much hope, for ,at least, 2Ghz ALUs (at least the ALUs).

 

[ShootMyMonkey]

If you're assuming 4x FSAA at 720p, then you're talking around 18 MB. MSAA 2x, well, then 12 MB. Sans tiling, 10 MB is insufficient for antialiazing of any order.

1.5 MB of cache at least would have been good -- since it's shared among 3 cores, I think having only the equivalent of 340k of L2 cache per core isn't so much. At the very least, with 1.5 MB, you'd have at least 512k per core if it were evenly partitioned. Personally, I don't see what's the big deal about sharing the L2 cache, anyway.

 

[Nite_Hawk]

Perhaps this has been mentioned (though I didn't see it), but I think the lack of a DVI port is a pretty major chink.

The xbox360 will end up connecting to most LCD or plasma HDTVs using component output, and some will probably end up using svideo or (dare I say it) composite. Having DVI means that there is atleast one more highend option for users with expensive TVs. There will be a greater chance that they will happen to actually use a high quality connection as opposed to something like composite.

Especially if you throw 1080p into the mix, DVI is going to play a big role.

Nite_Hawk

 

[darkblu]

i have to second blakjedi here - 10MB is not sufficient _given_ the mandatory targeted res of 720p, despite all the clever tiling. expect latency issues :?

ATI is on record as saying there wont be any performance penalties for 2x AA (2 tiles) and only 1-5% for 4x AA (3 titles... of course it seems the GPU can write to the eDRAM as it is also flushing the framebuffer to the main memory).

yes, i'm aware of what ATI are saying*. regardless, the only thing that can be carried out concurrently wrt EDRAM access is the zbuffer pass for the next frame while the last tile of the present frame is being flushed. unfortunately, the flushing of any prior tiles from the current frame is going to incur full penatly, and i can see absolutely no way to mask that.

If those estimates from ATI are accurate it would seem that the process is pretty effecient. As tradeoffs go, I am sure there are other places of concern than a 1-5% drop at 720p with 4x AA.

* allow me to express my scepticism re ATI's claims: how can they claim 1-5% penaly at a certain tiling scheme if the tiling logic penaly is inherently _fixed_cost_ while the overall frame latency is arbitrary? how do they come up with the percentage figure? based on what frame timing?

 

[Dave Baumann]

unfortunately, the flushing of any prior tiles from the current frame is going to incur full penatly, and i can see absolutely no way to mask that.

That is opposed to what they have said. I distinctly asked and they are saying that writes can occur whilst the previous frame / tile is being resolved.

As for GDDR3 for system RAM - there is no more latency for GDDR3 than DR2, which is presently being employed by Intel systems (and AMD sooner or later). The system RAM on Xenos is alo running over twice as fast as current DDR2 system memories which would reduce the latency in relation.

 

[Edge]

Eight CPU's (ala Sun's Niagara chip) with 512 KB of exclusive cache for each CPU (4 MB total), and a segmented memory design with 512 MB for the CPU, and 512 MB for the GPU, with higher bandwidth (50 MB/sec) between the CPU and GPU would have been much better, so memory accesses would be less contentious. And of course a much higher capacity optical disk format.

Microsoft could have easily afforded that, and would have put them in a better position against it's rivals.

 

[gosh]

Xbox 360 has quite a few chinks, if not more accurately, potential chinks in the ol' armour. I think there are three major types of "chinks"

Type A. Potential Chinks (i.e. conjecture about potential shortfalls like design failures)
Type B. Chinks compared to their competitors (i.e. shortfalls when comparing them head-to-head with their competitors)
Type C. Market Adoption Chinks (i.e. concepts that just do not pan out/under supported/ill concieved)

So here are some of my thoughts on possible "chinks" the X360 armour. These are all "thoughts" and conjectures, just things to fuel the fire. I actually think all three have issues and also think MS has done a good job minimizing holes. But there are always some... even if they never materialize. So my thoughts:

1. 10/100 base-T Ethernet instead of Gbit LAN.
Type: A & B.
Why?: If the Xbox 360 is aimed to be a media center extender, the limited bandwidth of a standard ethernet port may affect the transfer of large files. Their prime competitor has 3 Gbit LAN ports in comparison. While MS is promoting the X360 as a hub, it appears its competitor is posititioning itself better. The ethernet limitation may never materialize, but there is an issue of mindshare.
My guess: No impact on game play (which is most relevant), may be an issue with transfering HD TV media from a Media Center PC.
Weight: 2 out of 10 (10 being most)


2. Poor launch software
Type: A, B & C
Why?: The poor E3 showing with very little playable and with about half of the software looking like nothing more than upgraded ports does not inspire confidence. Yet this is true of every pre-launch E3 for the most part. What I did not see at E3 was a killer launch app. Gears of War is coming in Spring 2006 (I think it will be delayed). Nothing else shown looked like it will wow us come Fall 2005. That may very well change, but E3 did nothing to change that perception. That is chink Type A which would lead to Type C. If X360 does not show consumers a good reason to migrate over they wont. And Sony can reinforce this chink by showing much better software, i.e. Type B. Software is where it is at, and MS has not shown anything that blows people away... yet. This is typical, but if MS wants to be in contention for market leadership they need to be making moves... like Sony already has with their E3 reel.
My guess: Xbox 360 will have 2 solid launch titles: PDZ and PGR3, with a (delayed) summer killer app in Gears of War. Kameo will be an inspiring, if not mainstream, solid title. Everything else will be hit or miss, your average console releases. Wild Card: Madden 2006. If it looks like the artists concept footage shown at E3 it will sell systems like crazy. If it looks like the media at IGN it wont move hardware. The second wild card is what Sony can show, in playable form, come Fall 2005 and Spring 2006. If MotorStorm and KillZone 2 are like the expectations Sony have set MS is in some trouble. If they are noticably less than what has been shown MS will be in a position to strike the first blow with a killer app... like Halo 3.
Weight: 6 out of 10 (10 being most)

3. Consumers feel nickled and dimed
Type: B & C
Why?: Pay-to-play full access for Live, wireless controllers may not be standard, wireless networking extra, face plates, micro transactions, etc... Sony/Nintendo are offering, it seems, mainly free internet play. LIVE has better features/service, but you have to pay. Some consumers will enjoy the quality, many others will feel the sting of, "I pay $300 for the machine, $50 for the game, and now they want more?!" Face plates are a moot point since they are just a side addon and uneccary, and wireless networking is minor IMO (I know I would never think of using wireless!) as it is more of a convienence that is not normally required. Nice, of course, but not a deal breaker in most cases. Where I can see some sting, especially if MS does not keep it in control, is micro transactions. No one wants to buy a "game" that is really a pretty engine and some demos and the real "game" levels you need to download for extra. No one wants to buy a 5hr game and have the publisher offer the real substance as an extra.
My guess: Live will continue to be the cream of the crop as Sony continues to allow most developers to offer their own services the way they want. The extra expense will be negligable to most online gamers. Micro-transactions will be a nice perk but abused by some. But most of the tension will be directed toward publishers.
Weight: 3 out of 10 (10 being most)

4. 1080p HD
Type: B
Why?: MS went all out proclaiming the HD era with 720p being standard and supporting 1080i... and Sony one upped it with 1080p support. Even if Sony never provides any meaningful quantity of 1080p games they have rained on MS's parade some (similar to how MS rained on Sony's Teraflop parade).
My guess: With only 10M American homes with HD TVs and Europe a mess I cannot see 1080p being a major issue. More significantly I have pessimism that the PS3 will be able to deliver 1080p games with 4x AA and 128bit HDR at 60fps (or even 30fps which seems to be the current standard, which is sure to make money scream). This is one of those bulletpoints GPU consumers are so familiar with: Yeah, it can do it, but not at any meaningful performance. Call me a pessimist here.
Weight: 1 out of 10 (10 being most)

5. Hi Def optical
Type: A, B, & C
Why?: Sony has a clear progression they have evangelized: PS=CD, PS2=DVD, PS3=BR. They have given consumers a mindset that 1. next gen games always need more space and 2. next gen consoles coincide with a format change. I do not believe either of these is necessarily true, but in any industry where casual consumers are the prime target market perception is all that is important. And thus Sony has size on its size. Further, having BR wont hurt consumers necessarily. If BR is not the dominant format it could hurt Sony's bottomline, but having some BR movies and a lot of space wont hurt the console. On the other hand it could hurt MS. In the least Sony will portrary MS as "non-progressive" and lacking a key feature.
My guess: It wont be a big issue where games are concerned in 98% of games. As for movies, HD adoption is very low. DVD's looked better than VHS on standard TVs that have like 380M units in the US. HD TVs have 10M. This will be more relevant down the road, but not now. What matters now is that Sony characterizes MS as failing in this area. It will sway some, but the true benefit of BR/HD-DVD will be recordable media and that seems to be a few years off and wont be a PS3 featurte. This is a mindshare win for Sony, but I do not think it will impact the games much.
Weight: 3 out of 10 (10 being most)

6. Main memory latency
Type: A
Why?: GDDR3 is great for GPUs because it has high bandwidth. The high latency is not an issue for a GPU. But this may not be the case for a CPU, especially one with 1MB of cache for 3 CPU cores and that can processes 115GFLOPs (compred to top end PCs in the mid 20s). There was a reason why CELL went with low latency XDR...
My guess: I think this is one of those overlooked factors. Everyone wants to focus on the GFLOPs, Polygons, Bits, etc... I think a lot of that is a wash. I don't think this will be a huge problem, but I see it as a potential bottleneck. Conversely, preventing the framebuffer from eating away at the main memory's bandwidth is a big win as the 512MB can be fully exploited for storing--and USING--textures and other assets. The procedural synthesis design of the CPU-GPU and their use of HOS will be another space/bandwidth saving features. This pluses will probably offset the latency negatives a bit.
Weight: 3 out of 10 (10 being most)

7. Dev support
Type: A & B
Why?: We have already seen a few vocal developers go "Pro-Sony". While it is pretty clear MS has a lot of support this round from the get go (especially compared to 2001) it is never good to have quality devs espousing the wins of your biggest competitors. I think software is vital, and these areas are very sensative in the consumer realm. But what is more scary for MS is not locking up key devs. There is absolutely no reason there should be rumors that NJ or Bioware games will be PS3 bound. MS is weaker than Sony on exclusive content, and losing some of their big names could hurt their overall progress goals. They wont prevent it from selling, but it could kill their goal for being #1 in NA.
My guess: I think MS has done a wonderful job with support. They also have good tools and a solid HW design that should be a big win. They also are being more aggressive this time around about porting PC titles at launch (something they were careful not to do too much of at the Xbox launch). That said, losing a Ninja Guidan, a Bioware game, etc... stuff that makes Xbox special and stick out will only result in customers going to the competitors product. With no clear killer app at launch (I can name 5 Sony demo reels that scream killer app... none may make launch and as usual we will be lucky if 1 is a killer app, but the perception is that Sony is loaded) not losing the marquee titles they have is important. My guess is they will win some and lose some--which is not good enough to be the market leader.
Weight: 7 out of 10 (10 being most)

8. Not as flexible as devs would want
Type: A
Why?: In order execution, limited cache, limited branching, trimmed down cores compared to the 970. While no one will mistake the XeCPU for CELL, it ain't no P4 or AMD64 either when it comes to GP code.
My guess: Closed box system with a fairly familiar lineage. Most devs will overcome this aspect because both consoles with announced CPUs require them to. As some devs struggle with GP code on CELL's 1 PPC the XeCPU will seem like a dream with 3. It is all in the eye of the beholder. That being said, I expect some growing pains early, especially with some PC devs. The limited cache wont will be negligable because consoles do a lot of streaming and it is a closed box not being asked to be a PC that has to run CAD, Outlook, IE, Winamp, and calculator all at the same time with all that legacy baggage.
Weight: 2 out of 10 (10 being most)

9. MS has sent a Mixed Message
Type: C
Why?: E3. MS had the stuff to rock. It came out that neither Kameo nor PDZ looked as bad as MS had shown at MTV, yet they did not trump either very well at E3. Ditto some other stuff. PGR3 was absent. Heck, the games just stunk. And MS seemed defensive about their specs and scattershot. They had the momentum and freely gave it away. While preorders are doing well, MS needs to evangelize that this IS the thing to get this year. That starts with the games, not stupid hip hop MTV specials. Early adopters are about the games, and so far no sale. Potential, but no sale. MS has also really hit Live hard, but without the SOFTWARE to show it shine it is almost pointless going indepth on it. Features, Services, Software... that is MS's mantra. Yet their weak software showing has not demonstrated HOW those will benefit us in game.
My guess: As final hardware arrives and real games start showing up I think this will change. But right now this is a HUGE bugger of a problem for some. I see it being resolved with actual products (X360 looks very solid, I would agree with MS that it should be fine next to the competition, and it has a lot of software in the tube). But right now, I just don't feel the love. MS needs to change that.
Weight: 8 out of 10 (10 being most)

10. MS is far too focused on protecting the PC
Type: A, B, & C
Why?: MS has a multi-Billion dollar business to protect. The game console is an extension of their dominance--not a replacement. MS took a shot at Sony'b bow last gen by entering the console race. This gen, Sony is not being shy: PS3 IS A COMPUTER. Media ports, IP cameras, Linux and a HDD, large optical storage, fairly standard high end GPU, lots of memory, and a ton of games. If Sony can offer a killer game machine + offer basic PC services like email, browser support, music, video editing, movie watching, etc... they could convince some consumers to opt NOT to get a cheap PC for those same purposes. Make no mistake: PS3 and CELL is a platform that will evolve to the PS4 and beyond.
My guess: I don't think it will pan out exactly how Sony wants. Specifically their online service will be a weak point in my guestimation. But that said, a PS3 that can do email, browse the net, and other basic PC tasks + play games and watch HD movies could pull some PC users. MS, on the other hand, seems to be careful not to blur the distinction. Xbox 360 is a media extender but NOT the core. The problem is Sony is moving into this territory. MS, who needs to preserve their billion dollar industry, has no interest in selling a lot of Xbox 360s--at a lost--and including a nice OS when they can just sell an OS for profit on the PC. Ultimately I believe MS will give some in this area and the Xbox 360 will get a larger HDD and more basic PC tasks. They have already talked about Video Chat, MSN like chat services, etc... Email, a browser, and some other basic PC tasks I think are destined to the device. I think Sony will force MS's hand here. Also, I believe MS and Sony are talking up their goals. MS does not want to talk about PC-tasks too much because of image. It must be about the games. On the other hand, Sony wants to get out of the "games only" image so they are talking about their other goals (again.... ahem). While I do not think Sony will reach all their goals, on the other hand I believe MS will have to be VERY VERY VERY careful. I think they will blunt the charge from Sony, but I do not see them, at least not yet, closing the door. I think MS will wait until Xbox 3, when they can come up with a profitable financial model, to begin targeting more PC like tasks. But I see Sony very dangerous here, and it may be MS biggest weakness. Odd how a companies biggest strength could also be a hinderance.
Weight: 10 out of 10 (10 being most)


Other things that could be chinks:

1. Too early... developers seem to be aiming more for a full transistion in 2006. 2005 looks slow for games, but on the other hand it looks like 2005 is the year they are spending their resources getting ready for 2006.

2. To expensive... Will the X360 have software to justify a $300+ console, an extra controller, wireless adapter, and a couple $60 games?

1) Most people dont need higher BW than that on thier internet. its useless for a console in most cases

2) You have got to kidding me about launch software. It has probably one of the strongest launch software lineup in gaming history.

3) Consumers feel nickled and dimeD at paying 299? I wonder how they will feel when paying 475 for PS3

4) 1080p is not needed until XBox 3

5) MS is playing it smart. Its primarily a software company and wants Windows to support whichever format will win, so its not supporting any outright

6) Developers always complain about memory latency, it doesnt effect system at all

7) MoreDevelopers than Xbox 1 are supporting Microsoft, Plain and Simple

8) No comment, thats stupid

9) Thats even more stupid

10) The Console is not a PC, and Console users want it that way, so the stupidest was left at the end
[/b]

 

[ERP]

unfortunately, the flushing of any prior tiles from the current frame is going to incur full penatly, and i can see absolutely no way to mask that.

That is opposed to what they have said. I distinctly asked and they are saying that writes can occur whilst the previous frame / tile is being resolved.

As for GDDR3 for system RAM - there is no more latency for GDDR3 than DR2, which is presently being employed by Intel systems (and AMD sooner or later). The system RAM on Xenos is alo running over twice as fast as current DDR2 system memories which would reduce the latency in relation.

This is how I read the article

Having said that you should work out what a flush costs (it hits peak bandwidth to main RAM, and how many you could do in a frame. Remember you need 3 of them for 4AA at 720P.

One of the benchmarks it would be interesting to run is whether the double buffering the EDRAM is worth the additional tiling overhead it incurs. I'd guess it probably wouldn't be, but it'd be worth running the test.

 

[darkblu]

unfortunately, the flushing of any prior tiles from the current frame is going to incur full penatly, and i can see absolutely no way to mask that.

That is opposed to what they have said. I distinctly asked and they are saying that writes can occur whilst the previous frame / tile is being resolved.

previous frame or tile? that's essential. if it's frame, then that is in line with what i said - they can carry out zbuffer writes while the last tile of the prev frame is being resolved. if it's tile, and more imporantly not the last one of the frame, then i'd be very curious to know how they do it assuming that Xenos is still an IMR, i.e. its access within a tile is still random.

 

[Shifty Geezer]

3) Consumers feel nickled and dimeD at paying 299? I wonder how they will feel when paying 475 for PS3
4) 1080p is not needed until XBox 3
5) MS is playing it smart. Its primarily a software company and wants Windows to support whichever format will win, so its not supporting any outright etc.

gosh - at the risk of sounding rude, read the first post in this thread, and then push off

 

[ERP]

unfortunately, the flushing of any prior tiles from the current frame is going to incur full penatly, and i can see absolutely no way to mask that.

That is opposed to what they have said. I distinctly asked and they are saying that writes can occur whilst the previous frame / tile is being resolved.

previous frame or tile? that's essential. if it's frame, then that is in line with what i said - they can carry out zbuffer writes while the last tile of the prev frame is being resolved. if it's tile, and more imporantly not the last one of the frame, then i'd be very curious to know how they do it assuming that Xenos is still an IMR, i.e. its access within a tile is still random.

Just double buffer the tile......

half the memory is resolving/clearing the other half is being used as a tile. As I said in my previous post I wouldn't bet oin this being a win from a performance standpoint, but it might be.