RSX = Stream Processor!?!

It would help if I knew the largest texture size supported by RSX and whether or not you are including mipmaps in the total.
 
I'd say 16 MB mip mapped, or 22 MB w/ alpha for a 2048 x 2048 dx1 texture, assuming that's the largest supported texture size.
 
ROG27 said:
I'd say 16 MB mip mapped, or 22 MB w/ alpha for a 2048 x 2048 dx1 texture, assuming that's the largest supported texture size.

so the rsx has upto 22mb of texture cache....is that the sames as a 7800gtx?
 
Nerve-Damage said:
WHAT!?!

When did graphic cards get cache???

Vram (texture) memory yes...cache no.........

Don't you remember the good ol' days of texture cache and framebuffer memory...

Diamond Monster 3d w/ 2MB texture cache and 2 MB framebuffer memory

3dfx for the win!
 
Nerve-Damage said:
WHAT!?!

When did graphic cards get cache???

Vram (texture) memory yes...cache no.........

barbarian said this:

barbarian said:
Here's an interesting tidbit - a DX1 texture is always guaranteed to fit in RSX's texture cache. How many megs would that be?

if he was talking about memory i am sorry....
 
Psssh! Whatever - Barbarian's given us more than any other dev so far; if anything the others should be looking to his example! With some of these comments lately, we can actually start to paint a picture, blurry though it might be.
 
mrdarko said:
barbarian said this:



if he was talking about memory i am sorry....

No, No, No,

I got what Barbarian was saying!! ;)

That (Cache) would make RSX unique to its PC brethren (i.e. 7800 GTX).

Barbarian was hinting that RSX will contain some cache along with the reported 256MB of Vram Sony stated at 2005 E3.

Most (if not all) todays graphic cards do not contain any cache…from what I can remember.

RSX having cache guarantees to a certain existent faster texture swapping, less redundant data processing, faster post processing routines, ECT…
 
Last edited by a moderator:
Nerve-Damage said:
No, No, No,

I got what Barbarian was saying!! ;)

That (Cache) would make RSX unique to its PC brethren (i.e. 7800 GTX).

Barbarian was hinting that RSX will contain some cache along with the reported 256MB of Vram Sony stated at 2005 E3.

Most (if not all) todays graphic cards do not contain any cache…from what I can remember.

RSX having cache guarantees to a certain existent faster texture swapping, less redundant data processing, faster post processing routines, ECT…

this "feature" sounds like a very good thing to me.....but do you believe it to be viable on rsx?
 
mrdarko said:
this "feature" sounds like a very good thing to me.....but do you believe it to be viable on rsx?

Honestly, only time will tell….

If the cache becomes a bottleneck for some odd reason for a developer; will they have the option of bypassing it by hardware/software control? That’s my only concern...anyway.

Anyhow, there’s plenty of advantages of having on die cache (proper amount of it anyway) compared too the disadvantages of not having it.
 
The G70 architecture has 24 KB of texture cache per pipeline. Now times that by 24 pipelines. Up from the 16 KB texture cache from the previous generation (the 6800 series).

Saying their is no cache in a GPU is just silly. There is lots of cache, but for whatever reason the amount of cache is very rarely mentioned by the GPU manufacturers.
 
Most (if not all) todays graphic cards do not contain any cache…from what I can remember.
Umm, that's not remotely true. Today's cards have plenty of cache. Texture cache is there, but if you write your pixels and access your data intelligently you don't need a whole lot. More on-chip memory is found in fifos/register cache/pixel cache/whatever you want to call it, and there's a ton in any graphics chip. This is what makes them good at absorbing texture latency and sustaining their peak throughput. You also have caches for storing information about Z and colour compression, and in the case of ATI, for Hi-Z data.

RSX having cache guarantees to a certain existent faster texture swapping, less redundant data processing, faster post processing routines, ECT…
That's pretty much all wrong. Only in the worst cases (e.g. tons of textures, or very non-coherent dependent texture access) is texture swapping an issue, in which case more cache isn't the solution. Post processing has very predictable memory access, and again more cache won't help much. Don't know where you could save redundant data processing.

I guarantee you that G70 has already optimized the cost/benefit ratio of caches to the point where you'll only get a few percent more performance with bigger caches. Any tricks to maximize memory usage are already in use now to minimize costs and maximize profits for the low bandwidth budget cards in the PC sector.

All cache does is help you achieve perfect efficiency. For bilinear texture filtering you need the nearest 4 texels, and under magnification many neighbouring pixels will need the same 4 texels, and that's what the cache is for. Even under minification, adjacent pixels need 2 of those texels. If you draw your pixels in an intelligent order (and believe me, these chips have done so since the GF3/Radeon days), you're unlikely to need to load many of those texels again.

There's a fundamental limit to how little bandwidth you'll need, and modern GPU's from ATI and NVidia aren't far from achieving that.

The only time cache will actually save you significant bandwidth is if a whole texture can fit in the cache and that texture is tiled and repeated, which we'll see less of in next-gen graphics, or if you're doing tiled rendering like Xenos or PVR's TBDR chips do.
 
now i am confused....

mintmaster,could you try to explain what barbarian is hinting at in his last post.

thanks
 
Edge said:
Saying their is no cache in a GPU is just silly. There is lots of cache, but for whatever reason the amount of cache is very rarely mentioned by the GPU manufacturers.

Well I learned something new today!!

My eVGA(s) 7800 GTX have itty bitty cache modules on the GPU die :oops:
 
Nerve-Damage said:
Your opinion anyway.....
Well this opinion is based on hard data while working at ATI a few years back, and my extrapolation of that data to current GPU's based on public synthetic tests. I am also very familiar with 3D coding, which you do not seem to be judging by your assertions.
 
Barbarian said:
Here's an interesting tidbit - a DX1 texture is always guaranteed to fit in RSX's texture cache.

Haha. Funny. :D

DirectX 1 (back then called GameSDK) didn't have Direct3D in it. Way back in 1995, it was just DirectDraw, DirectInput, DirectPlay, and DirectSound.

Direct3D didn't make it into DirectX until DirectX 2. :devilish:

Barbarian said:
How many megs would that be?

The correct answer is zero megs, since they don't exist they always fit. ;)
 
Last edited by a moderator:
aaaaa00 said:
Haha. Funny. :D

DirectX 1 (back then called GameSDK) didn't have Direct3D in it. Way back in 1995, it was just DirectDraw, DirectInput, DirectPlay, and DirectSound.
Maybe he meant DXT1?
 
Back
Top