PDA

View Full Version : Nice ATI really nice presentation...


Pages : 1 2 [3]

ChrisW
03-31-04, 03:44 PM
What are the advantges of PS 3.0 over PS 2.0? There are few quality PS 2.0 games available. Let alone at reasonable speed.

My 5900XT just stalls with Tombraider.
I don't know much about it. It will probably not make much difference except to be able to claim support for it. This will give "TWIMTBP" developers the green light to make "special effects only available for NV4x cards" as they will claim the effects just can't run on ATI cards. This will be a huge marketing advantage for nVidia.

volkskrant
03-31-04, 04:06 PM
I don't know much about it. It will probably not make much difference except to be able to claim support for it. This will give "TWIMTBP" developers the green light to make "special effects only available for NV4x cards" as they will claim the effects just can't run on ATI cards. This will be a huge marketing advantage for nVidia.

These same as the Nv3x tech demo's? They do not work until someone release a wrapper (Dawn?)
:rolleyes:

Moose
03-31-04, 04:38 PM
PS 3.0 will come much faster than PS 2.0 because nVidia will encourage their "TWIMTBP" partners to use it. Instead of holding back the industry like they did with the NV3x, they will be doing the opposite this time. Personally, I'm all for progress.


I agree that Nvidia will be pushing it this time but I'm not sure how successful it will be. ATI has gained quite a bit of marketshare this last round especially in the high end market. Developers may not want to alienate all the ATI owners out there quite so readily as before.

I think alot may depend on which card holds the PS2.0 speed crown this time around.

I'm starting to wonder if a special PS3.0 version of Farcry or some other such game may be bundled with the NV40. I guess we'll see soon enough.

Ghola
03-31-04, 09:04 PM
Question: How many games are there that support shader 2.0 right now and would you say that actually make a big difference in game visuals?

How long did it take to reach that number?

(bur)

StealthHawk
03-31-04, 11:11 PM
Both CEO's of both corporations have been investigated one time or another by a regulatory body, so neither companies had angel halos to begin with. The CEO is the company since the company follows his vision.

K.Y. Ho won't be CEO of ATI for much longer.

Hmm,

No one noticed slide 38? Look at the 2nd sentence under the slide.

"Don't go into detail about the granularity of the VM system - just press home the fact that this solves important problems."

ATI is gonna have grainy graphics?

"Granularity" is technical jargon...

What are the advantges of PS 3.0 over PS 2.0? There are few quality PS 2.0 games available. Let alone at reasonable speed.

Possible speed-ups in PS execution compared to PS2.0 code. But even NVIDIA says that you shouldn't use dynamic branching everywhere.

Shamrock
04-01-04, 06:26 AM
"Granularity" is technical jargon...


true, but me being an architect, I know about "granularity" we use it all the time in presentations, it means how fine or how "grainy" certain textures will look, but ATI wants to "evade" the fact it will look "Granular" and just point out the fact that it "solves important problems"?


:screwy:

Razor04
04-01-04, 10:53 AM
true, but me being an architect, I know about "granularity" we use it all the time in presentations, it means how fine or how "grainy" certain textures will look, but ATI wants to "evade" the fact it will look "Granular" and just point out the fact that it "solves important problems"?

:screwy:
Lets start with the definition of Granular:

gran·u·lar
adjective

1. made up of grains: consisting of small grains or particles

2. with texture of granules: appearing to consist of or be covered in small grains or particles

3. divisible: made up of conveniently small and independent parts
a granular interface

gran·u·lar·i·ty noun
gran·u·lar·ly adverb

Definitions 1 and 2 do not apply in this case as we are talking about Virtual Memory. This means that your theory of ATI having grainy textures is blown out of the water as we are dealing with memory and not the textures themselves. Definition 3 however seems to make a lot of sense as it posits that breaking down things into smaller parts makes the VM interface faster. Amazing how people don't take the time to research things they don't understand fully anymore. It doesn't take long to look something up...especially something like this where the confusion is in the definition of the word. This isn't a jab at you Shamrock...but a lot of people seem to be ignoring things lately because they are lazy or the truth wouldn't support their viewpoint (i.e. die-hard fanboys). This seems to be happening an awful lot lately as the launches near...and it isn't just on this board either.

MUYA
04-01-04, 10:58 AM
These same as the Nv3x tech demo's? They do not work until someone release a wrapper (Dawn?)
:rolleyes:
There will be difference this time if NV makes the demo in OpenGL and exposes in their driver SM 3.0 like calls etc...I dunno too much about it . But if things like Sm 3.0 PS equivalent coding was used, no wrapper can help as the r420 will not have (last rumoured) SM 3 capabilties. Also if anyone can explain further or shoot this down, I will be glad to hear it as its just something i am writing with my limited knowledge.

euan
04-01-04, 01:11 PM
Lets start with the definition of Granular:

gran·u·lar
adjective

1. made up of grains: consisting of small grains or particles

2. with texture of granules: appearing to consist of or be covered in small grains or particles

3. divisible: made up of conveniently small and independent parts
a granular interface

gran·u·lar·i·ty noun
gran·u·lar·ly adverb

Definitions 1 and 2 do not apply in this case as we are talking about Virtual Memory. This means that your theory of ATI having grainy textures is blown out of the water as we are dealing with memory and not the textures themselves. Definition 3 however seems to make a lot of sense as it posits that breaking down things into smaller parts makes the VM interface faster. Amazing how people don't take the time to research things they don't understand fully anymore. It doesn't take long to look something up...especially something like this where the confusion is in the definition of the word. This isn't a jab at you Shamrock...but a lot of people seem to be ignoring things lately because they are lazy or the truth wouldn't support their viewpoint (i.e. die-hard fanboys). This seems to be happening an awful lot lately as the launches near...and it isn't just on this board either.


Correct, eh mostly. VM Granularity is the "page size". IE the minimum chunk of data that you can ask for. The larger the grains, the less work a system has to do to keep track of, but lots is wasted. The smaller the more labour intesive they are to look after, but the more effieient the use of resources is. Much, in the same way a harddrive is split into clusters (on some filesystems) that are say for example 4KB. If you want to save a 1kb file, or even a file just a few bytes long, the entire 4KB is used up and unavailable to the system.

eesa
04-01-04, 08:28 PM
hehe, granularity of VM -> grainy textures. lol. Thanks for the break. Back to studying now.

Blacklash
04-01-04, 10:08 PM
That's why I continue to not be a fan boy devoted to one company. Both Ati and Nvidia are all about taking your money, and as much of it as they can get. This is why I actually try to support both companies because the high competition benefits us. Price vs. performance. When one company sits on top too longer they get lazy, and prices climb.

I am on my side, not Nvidia's or Ati's.

eesa
04-02-04, 12:11 AM
Here's my two cents. There's a thing called business ethics. It's that intangible factor that all businesses should value but in many cases there's no sure way to enforce it. In my honest opinion, what you see here isn't unethical, just straight business. How fast an application runs is highly dependent on what architecture it is optimized for. Everyone knows that. Where's the problem with them trying to sway developers to write code that runs better on their stuff? NV does the exact same thing and again, that's totally understandable. Any company that makes hardware tries to get devs to write optimal code for their platform. Why do you think companies are so eager to get devs to run their next gen cards in their development machines? If this is not how a company thinks, then I question how it's possible for them to stay in business so long.
And people seem to have totally ignored the fact that they're insisting that speed is important but don't sacrifice quality for it. Leave that choice to the users.
And for what it's worth, I wouldn't know about NV so I can't say anything, but having worked at ATI at least those that I've come in contact with, here are some of the nicest people I've ever had a chance of knowing and working with in a business environment. And as for the whole quality over speed thing, let me just say it really does seem like something the company believes in. I'm not one who usually buys into polished marketing.

StealthHawk
04-02-04, 03:23 AM
true, but me being an architect, I know about "granularity" we use it all the time in presentations, it means how fine or how "grainy" certain textures will look, but ATI wants to "evade" the fact it will look "Granular" and just point out the fact that it "solves important problems"?


:screwy:

Noooo, that is not the computer science definition of the term granularity. That is why I said that you were not using the correct "technical jargon" definition of the word. Where technical referred to computer science, not architecture ;)

Razor04 and Euan are correct, granularity refers to size/divisibility.

Shamrock
04-02-04, 05:34 AM
I see, only trying to learn here

Spiritwalker
04-02-04, 11:46 AM
Here's my two cents. There's a thing called business ethics. It's that intangible factor that all businesses should value but in many cases there's no sure way to enforce it. In my honest opinion, what you see here isn't unethical, just straight business. How fast an application runs is highly dependent on what architecture it is optimized for. Everyone knows that. Where's the problem with them trying to sway developers to write code that runs better on their stuff? NV does the exact same thing and again, that's totally understandable. Any company that makes hardware tries to get devs to write optimal code for their platform. Why do you think companies are so eager to get devs to run their next gen cards in their development machines? If this is not how a company thinks, then I question how it's possible for them to stay in business so long.
And people seem to have totally ignored the fact that they're insisting that speed is important but don't sacrifice quality for it. Leave that choice to the users.
And for what it's worth, I wouldn't know about NV so I can't say anything, but having worked at ATI at least those that I've come in contact with, here are some of the nicest people I've ever had a chance of knowing and working with in a business environment. And as for the whole quality over speed thing, let me just say it really does seem like something the company believes in. I'm not one who usually buys into polished marketing.

eesa has your 'internship' there ended already? And what hints may you be able to feed us about upcoming hardware ;)

eesa
04-02-04, 11:55 AM
well, what sucks is that while I'm there I'm technically not even allowed to post. As a result, my activities in the forums are reduced to general mocking and trolling and other things that are completely unproductive and gets me into trouble...

And since I've been away for 4 months, I don't really know what's going on anymore. But that's a good thing, cause I can start speculating again. Going back as soon as finals are over and you can be sure I'll do my best to figure out what 420 is all about... you can also be sure I won't breathe a word about it once I do if I know what's good for me though. Oh well, just the way things work I guess.