The Official RV630/RV610 Rumours & Speculation Thread

Status
Not open for further replies.
It's not exactly normal to see a dual slot cooler come stock on a mainstream card. That card has to be packing something under that hood.

If you want to make a case for RV630 running at high frequency based on the cooler size (which for all we know could have something to do with heat output or be a one-off design like dual-slot 8600 GTS/X1650XT HSFs), go ahead. But if you are saying that the HSF in that one picture is an indicator of memory bus width.... Cue :runaway:
 
So they just went with a two slot cooler for the fun of it?

No, each extra 128 bits of memory bus width add an extra slot to the cooler size. The reason for R600 delay had nothing to do with the chip itself but rather the drastic lack of 4-slot cooler availability.
 
I am sorry; we will return to the serious business of determining memory bus width from card's length momentarily.

BTW, this IS B3D, right? Just checking....
 
ahh, I'm going in circles. What I mean is that their must be alot transistors under that hood and it's not everyday you see a dual slot cooler on a mainstream card.

Now this I agree with. I an personally dying to learn T-counts for RV6x0 series.
 
Nothing, i just hope, we need strong competition and low prices :smile:

Don't worry about that, I think that all the card segments (value, mainstream, performance, enthuasiast, etc) from $100-$600 will be hotly contested by both companies over the next few months. The DX10 cards from both sides should be really good all around.
 
That's what I was wondering before when I saw the pics of the RV630 line at VR-Zone (?). The RV630XT has different PCB, it looks very much like X1950Pro, but there might be some other reason for it.

Edit: here are the pics

Hmm, if the HD 2600XT does indeed end up being faster than a X1900XT then I imagine it will be one heck of a value over time, actually it might just start of a amazing battle in the mid range, such performance easily jumps over Nvidia's 8600GTS.
 
So, in terms of unit counts (die area) G84 is 1/4 of G80 (SPs and TMUs), but with 1/2 the texture addressing and 1/3 ROP/MCs.

Can anyone explain what exactly is calculated by a texture addressing unit?
I assume its input is an x,y coordinate (and z for a 3D texture). Does it simply convert to a linear address and fetch the data or will it fetch 4 or 8 samples to bi- resp. tri-linear filtering?
 
You think 128bit card need ~22cm PCB?



aopen-5900xt-scan-front-small.jpg

gffx5800u-scan-front-small.jpg



One of the above cards has a 256-bit bus. The other one has a 128-bit bus. Both have 128MB of RAM. Using you bus width-to-PCB length correlation formula, would you please tell me which one is which?
 
Status
Not open for further replies.
Back
Top