Next Gen Intel Gfx ( PowerVR 4 ?? ) - SM3.0 and HW T&L

[iwod]

Phew...... whenever i think about PowerVR the first i thought of is Beyond 3D .
And whenever i thought of Intel Gfx i thought of Power VR ^^

The next generation IGFX from Intel will support Directx 9.0 and SM 3.0 as well as Open GL 1.5 . These info were posted on www.hkepc.com

Now my question is that ever since Intel has a cross lisence argeement with PowerVR i always thought there iGfx has something to do with them. But i also remember the Tile Based Rendering itself dont work with T&L. ( Back in the days of Power VR 2 )

The argument back then was that when SM3.0 alike comes out T&L will no longer be needed as most geometry operation can be done. So why is Intel bringing it back in?

And if this is not PowerVR4........... where the hell is it LOL.

 

[iwod]

Sorry for the spelling and gramatical errors.... i was properly too excited

( * Why cant i edit my post..... ?? )

 

[Rootax]

PowerVR Series 4 was canceled. SGX is Series 5.

 

[Joe DeFuria]

I should give you negative rep for even mentioning PowerVR for a PC based solution...but I'll be nice.

Unfortunately, I can't read anything that's not English...but is this actually a built-by-Intel IGP, or just more of what they've been doing recently...using ATI IGPs for integrated graphics solutions....

 

[Dave Baumann]

The notion of "T&L" (or any hardware vertex processing") can't be done with PowerVR's TBDR is a falacy - in fact PowerVR had a T&L processor, Elan, for NAOMI2 arcade machines.

 

[_xxx_]

You can't edit or do some other stuff until you have 1000 posts or so, read it in the appropriate forum:

http://www.beyond3d.com/forum/showthread.php?t=23045

As for PowerVR, I think this was discussed in at least five topics already

EDIT: Rootax, how do you come up with that? AFAIK series 5 was never actually produced in any form, SGX should have some similarities, but it isn't series 5 IMHO.

 

[Rootax]

You can't edit or do some other stuff until you have 1000 posts or so, read it in the appropriate forum:

http://www.beyond3d.com/forum/showthread.php?t=23045

As for PowerVR, I think this was discussed in at least five topics already

EDIT: Rootax, how do you come up with that? AFAIK series 5 was never actually produced in any form, SGX should have some similarities, but it isn't series 5 IMHO.

http://www.imgtec.com/News/Release/index.asp?ID=415

"The PowerVR SGX cores for wireless applications are part of Imagination’s PowerVR Series5 scalable and fully programmable multi-threaded universal shader graphics and video core family, previously codenamed Eurasia."

Ok, it's not Series 5, but a "part of" :smile:

Sorry for my short answer, english is not my mother tong

 

[Mariner]

In theory at least a unified shader core such as PowerVR SGX ought to be an ideal fit for integrated graphics - this assumes a similarly-performing unified core would take up less die space than the traditional separated pixel and vertex shaders.

Whether or not the new Intel IGP uses it's own in-house implementation as in the past or not I've simply no idea. Time will tell I suppose.

 

[_xxx_]

http://www.imgtec.com/News/Release/index.asp?ID=415

"The PowerVR SGX cores for wireless applications are part of Imagination’s PowerVR Series5 scalable and fully programmable multi-threaded universal shader graphics and video core family, previously codenamed Eurasia."

Cool, must have missed that

 

[Ailuros]

In theory at least a unified shader core such as PowerVR SGX ought to be an ideal fit for integrated graphics - this assumes a similarly-performing unified core would take up less die space than the traditional separated pixel and vertex shaders.

Whether or not the new Intel IGP uses it's own in-house implementation as in the past or not I've simply no idea. Time will tell I suppose.

I'd say that Intel's "house own" IGPs are quite a few paces apart from D3D10 compliance.

 

[Mintmaster]

The notion of "T&L" (or any hardware vertex processing") can't be done with PowerVR's TBDR is a falacy - in fact PowerVR had a T&L processor, Elan, for NAOMI2 arcade machines.

We were talking about this at length in another thread.

I have no doubt that hardware vertex processing is viable on a TBDR. The question is whether the techniques and graphics engines prevalent in the IMR dominated PC space will cause any headaches for TBDR vertex processing.

It seems possible, but there are a lot of end cases to consider, like high vertex load, framebuffer readbacks, occlusion queries, changing render targets often, etc. Dealing with these on a closed development platform is much simpler.

 

[Ailuros]

SGX is a USC.

 

[Xmas]

It seems possible, but there are a lot of end cases to consider, like high vertex load, framebuffer readbacks, occlusion queries, changing render targets often, etc. Dealing with these on a closed development platform is much simpler.

How are framebuffer readbacks tied to vertex processing? They will break parallelism on any architecture. And what's the problem with changing render targets often?

 

[Dave B(TotalVR)]

the question is, when is this IGFX chipset from intel gonna be out.

 

[max-pain]

summer

 

[Dave B(TotalVR)]

Excellent, is that concrete or inquirer style quicksand evidence?;P

 

[darkblu]

the current generation of integrated video by intel is already a deferred renderer with SM 2.0 and OGL 1.4

 

[Ailuros]

the current generation of integrated video by intel is already a deferred renderer with SM 2.0 and OGL 1.4

The current generation of Intel's IGPs is a tile based IMR with only PS2.0 and no geometry unit whatsover

 

[Pete]

Excellent, is that concrete or inquirer style quicksand evidence?;P

A little from column A, a little from column B.

 

[darkblu]

The current generation of Intel's IGPs is a tile based IMR with only PS2.0 and no geometry unit whatsover

it's a deferred renderer. and yes, it's ps2.0 and cpu emulated vertex shaders, but there's nothing wrong with that.

ed: just to clarify - by current i mean the gma 900/950.