DeltaChrome preview at ExtremTech

It was also my understanding that DeltaChrome has no FP16 abilities, and certainly nobody else has mentioned it - I wouldn't be suprised if it was a mistake on ExtremeTech's part. Can anyone else confirm 100% either way?
 
I thought it was strictly FP24 as well. Let me see what I can find out......

Pixel shaders are a little more interesting, because there are all sorts of color precision issues and the like. The S3 folks said they tried hard to avoid cutting precision internally in this chip, and they did not—I repeat, did not—include a bunch of 12-bit integer shaders plus a couple of DX9 pixel shaders, like a certain other graphics chip company. This puppy is eight pipes with eight real, DX9 pixel shaders. The max color precision in S3's pixel shaders is 96 bits per pixel, or 24 bits each for color channel, just like ATI's R3x0 chips.

from TR's preview
 
ByteMe said:
I don't see were you get this. Maybe there is more to this idea than I see. Do you have some XGI or S3 info that is not posted yet?

You don't necesarily see everything. However, something that certainly is easy for you to see is to search by username and spot trends...
 
parhelia said:
Tridam : no I didn't know that. I'll check it out later on.

This fact has been clearly explained on this forum... You've also read it in French on Hardware.fr in the news and in the forum (where you're using another nickname : Magicman).

Don't take us for fools. You know very well what we're talking about. If it isn't the case it means that you're maintaining some BS without knowing what you're talking about. You're saying everywhere you can that XGI isn't cheating. You're even accusing Hardware.fr of being biased. And now you're saying that you don't know what we're talking about ?? What's that ???
 
Tridam said:
I'm sure you know well enough that changing the exe is not what you have to do. You have to edit the driver file and change the name inside the driver file....

Hmm... I'll try that later on my Volari V8 Ultra as well.

However, something that certainly is easy for you to see is to search by username and spot trends...

Tridam said:
This fact has been clearly explained on this forum... You've also read it in French on Hardware.fr in the news and in the forum (where you're using another nickname : Magicman).Don't take us for fools.

There's a guy called Magicman here? I couldn't seem to find him while searching through user name :?:
 
Hanners said:
It was also my understanding that DeltaChrome has no FP16 abilities, and certainly nobody else has mentioned it - I wouldn't be suprised if it was a mistake on ExtremeTech's part. Can anyone else confirm 100% either way?

DeltaChrome have some limited FP16 abilities. Some shadercode works faster if the pp hint is used. It depend on the instructions that were used in the shader. There was not enough time to figure out which instructions can use the pp hint to work faster.
 
Demirug said:
DeltaChrome have some limited FP16 abilities. Some shadercode works faster if the pp hint is used. It depend on the instructions that were used in the shader. There was not enough time to figure out which instructions can use the pp hint to work faster.

Limited? Surely it either supports it or it doesn't?
 
Hanners said:
Demirug said:
DeltaChrome have some limited FP16 abilities. Some shadercode works faster if the pp hint is used. It depend on the instructions that were used in the shader. There was not enough time to figure out which instructions can use the pp hint to work faster.

Limited? Surely it either supports it or it doesn't?

I believes we have different definitions of full fp16 support. IMHO full fp16 means the chip can calculate and store internal with the fp16 format.

DeltaChrome can calculate some (at least one) operations with fp16 but all other calculations are done allways with fp24. That is the reason why iI used the word limited.

IMHO the NV3X Chips have also only limited fp16 abilities. All operations (one exception) are calculated with fp32. fp16 is only the storage format for the result.
 
xGL said:
Tridam said:
This fact has been clearly explained on this forum... You've also read it in French on Hardware.fr in the news and in the forum (where you're using another nickname : Magicman).Don't take us for fools.

There's a guy called Magicman here? I couldn't seem to find him while searching through user name :?:

Magicman is parhelia's nickname on the HFR forum.
 
Let me ask a question regarding the two new players (Volari and DeltaChrome). Do they have *any* redeeming values? Volari has terrible drivers, poor performance, obscene power requirements, (at this point) terrible IQ, and is way too expensive. DeltaChrome just seems all-around weak. They offer nothing that makes me want to sit up and say, "Damn." It's just another current-generation card. It's not magically fast, it's not near-free AA, it's not really anything in particular. Both are just bland.

Plus, neither card is available in the States yet, and Volari only has limited availability in Europe. By the time we see DC and Volari in the States, we'll be a month or so away from NV40 and about two months away from R420.

So, basically, what is the point? The two new players are following a near-identical release schedule, and it doesn't match up to ATI or NVIDIA at all. It just seems like they're going to get blown out of the water completely as the current high-end cards drop in price. Hell, I don't even know that the new higher-end models would match a 9800 non-Pro (definitely not a 9800np, actually) or a 5900XT. The ET preview of the DC is with a 5600 Ultra and a 9600 Pro, not even the two newer models. I'd be surprised if that DC won *any* benches against a 9600XT or a 5700n Ultra.

Are we just curious for the sake of curiousity (and the burning desire for a third option)? Do we just want to see if they offer anything new in the architecture? It's just like the old days, when you had three options: 3dfx, NVIDIA, and everything else.

Now that we've seen Volari and DeltaChrome, what else do we have to look forward to in the way of a third option? I know Series 5 is supposedly coming in the next few months, but information on that has been scarce, to say the least. Is there anything else on the horizon that actually might not suck?
 
I'm too bored to look up Wavey's exact comments concerning the small players, but it's essentially true. In their majority they seem to wait for the API to be defined more or less and then start development.

Roadmaps of XGi and S3 show architectures with PS/VS3.0 compliance or beyond, but I'm afraid that we'll be waiting again for ATI/NV's "DX-Next" compliant products when those hit the shelves.

I know Series 5 is supposedly coming in the next few months, but information on that has been scarce, to say the least. Is there anything else on the horizon that actually might not suck?

Architectures look impressive on paper, yet we all know that it can be misleading, and I really don't think there are any exeptions until an architecture gets thoroughly tested/analyzed by independant sources. IMG gives me the feeling for quite some time now, that they're following in relative terms a "put up or shut up" logic; here considering their track record of delays or recent product cancellations, it makes sense to me.

It'll "suck" if it doesn't make it to shelves on time.

Are we just curious for the sake of curiousity (and the burning desire for a third option)? Do we just want to see if they offer anything new in the architecture? It's just like the old days, when you had three options: 3dfx, NVIDIA, and everything else.

Part of it must be curiousity. Prior to S3/XGi there were a year ago or more also hopes for Matrox and 3Dlabs, yet the landscape seems pretty blunt there too.
 
The Baron said:
Volari has terrible drivers

Well I started up our internal build of our upcoming game on Volari and everything worked.
Including the PS2.0 shadow buffering high-precision buffers.
Until now this only worked on R3xx cards.
So I'm impressed.

But I agree with anything else you said.
 
Reverend said:
The Baron said:
So, basically, what is the point?
I think the point is that we're seeing two companies trying to at least penetrate a $4Billion industry.
Hm, you think I'm being overly negative? I want to like the Volari, but it seems clear that the multichip consumer card concept just isn't the best way to go right now. Even assuming that the driver issues could be worked out, dual Molex connectors? Sorry--that's not a way to penetrate any market (unless, of course, you have die-hard fanboys already, a la 3dfx). DeltaChrome is just simply.... bleh. It's got eight pipelines on a mainstream card--wooh. It has a 128-bit memory interface--okay, so what's the point again? It's being benched against cards that are over six months old, and it's still not winning. There's no way that the higher-end DC can win anything against the 9800XT or the 5950 Ultra if it doesn't have a 256-bit memory bus.

Just seems to me like these aren't The Mythical Third Option.
 
if you take a look at the perfornace of S8 vs Nvidia's 5700 card (not ultra) and lower) you will see that S8 is NOT as weak as you think in DX9....actually it is very impressive for the state the card was in....


One other thing to remember Ati picked up most of S3's driver team when they went joint venture with Via (SirEric and OGL Guy just to name a couple)
 
Back
Top