RV630XT, RV630 & RV610 Spotted

[Anoobis]

Link.

Posted already??

Edit: There's also a note that the R600 may be 65nm but it sounds a bit shakey.

 

[pauldh]

Hmm, RV630XT is looking nice. I have a hard time imagining a 128-bit card being >= an X1900XTX though.

Following your link, found this. http://www.xtremesystems.org/forums/showpost.php?p=2071080&postcount=134

I like to believe the line: "Compared to the solution of NVIDIA, which will give the R600 in terms of performances?
You will be surprised…"

 

[Anoobis]

Possibly, but according to the VGA charts a 7600GT can lock horns with an X850XT pretty well and is slightly above par against a 6800Ultra so it's not that far fetched. It's still probably a bit of assumption on VR-zone's part though.

 

[sailer]

Keep in mind that its not just the 128 bit interface, but the GDDR4 as well. Now, if I could only get my hands on one, instead of just looking at pictures.

No wait, I used to have the same thought when looking at centerfolds.

 

[sirheck]

Possibly, but according to the VGA charts a 7600GT can lock horns with an X850XT pretty well and is slightly above par against a 6800Ultra so it's not that far fetched. It's still probably a bit of assumption on VR-zone's part though.

i agree. have messed a lot with the 76gt,s x800xl,s and 68gt,s

they are all pretty close performance wise.
except for the x8## series and their lack of sm3.0.

i now have an 8800 so im doin ok :lol:

 

[kaotao]

Hmm, RV630XT is looking nice. I have a hard time imagining a 128-bit card being >= an X1900XTX though.

Yeah, that's what I was thinking.

To the OP: Thanks. Nice find.

 

[bullaRh]

''Compared to the solution of NVIDIA, which will give the R600 in terms of performances?
You will be surprised…''


well i sure hope i will =D haha, thanks for the link :b

 

[prozac26]

Great link, good read.

 

[pauldh]

Keep in mind that its not just the 128 bit interface, but the GDDR4 as well.

Yeah, but lets say 2.0GHz GDDR4 over a 128-bit bus vs. 1.55GHz GDDR3 over a 256 -bit bus. Obvious advantage in mem bandwidth to the X1900XTX. The 7600GT did impress, Just saying it's alot to overcome having half the bus, and the X1900XTX sure ain't a X850XT.

 

[sailer]

Its quite a mix of numbers, and that's why we have to wait until actual benches come out. Until then, its all guess and speculation. It would be bad for AMD if the thing didn't perform as well as my X1900 XTX. If that happened, all the hype and wait will be for nothing.

 

[TheGreatGrapeApe]

Keep in mind that its not just the 128 bit interface, but the GDDR4 as well.

Yeah, but lets say 2.0GHz GDDR4 over a 128-bit bus vs. 1.55GHz GDDR3 over a 256 -bit bus. Obvious advantage in mem bandwidth to the X1900XTX. The 7600GT did impress, Just saying it's alot to overcome having half the bus, and the X1900XTX sure ain't a X850XT.

Yeah it's gotta be damn fast memory to make up for that. And I wonder how much more expensive it would be to use GDDR4 than to put 256bit memory controller.

Now of course Samsung has been talking about using the 1.23(2.5), 1.4(2.8) and 1.6(3.2) ghz GDDR4 for a while, not to mention that 2.0(4.0) stuff I posted last time, but nothing has even the 1.25(2.5) stuff yet, and it's not going to be cheap. In that 2.0/4.0 article they mentioned issue with using the 3.2 on PCBs, so it's unlikely they will be able to match the bandwidth just through fast memory alone.

The thing is I doubt the X2600 will ever equal the X1900XTX at hi resolution, but these numbers they're using for 3Dmark are all at low resolution and no AA, so they aren't memory bound there. That's why the numbers make sense to me, but also don't mean much for games and real-world situations.

I think that the X2600 will remain a primarily 1024x768-1280x1024 range card, just like all the mid-level cards before it, so probably won't matter as much. But, man, I betcha they would just sing on 256bit, like an X1650XT or GF7600GT could likely use some more memory (just look at hyper-clocked GSs). Also for a laptop a slow 256bit memory configuration would be better than a fast and hot 128bit solution. I'd rather have 500mhz 256bit than 1/2ghz 128bit.

I'm just looking forward to the media aspects, especially after all the various sources (including yours), and the confirmation that ATi will go the nV route with dedicated UVD hardware on die. VC1 acceleration never matter much to me (it's less intensive than H.264) but if it helps watch an epic HD-DVD like Blazzing Saddles or Army of Darkness all the better. 8)

edit: WTF is it with the forum and 'you title has too many upper case characters' BS !?! :x

 

[sirheck]

WTF is it with the forum and 'you title has too many upper case characters' BS !?

yeah. i tried to reply to (sailor) earlier and got the same.?

 

[TheGreatGrapeApe]

Yeah I tried replying to the AGP vs PCI-E thread at work and got it and thought it was an IE thing or something since I'd never seen it before and I was getting a division by 0 error today too.

But it seems to be with all editors, so it's a new forum addition.
Mentioned it in the bug section hopefully it'll be fixed soon because the number of acronyms in this area is pretty much all we talk about...

ATi/AMD WDDM2.1 / D3D10 PCIe GDDR4 RV630XT-X2600XT w/ H.264 + VC1 UVD AVIVO DLDVI HDCP HDMI & SPDIF 5.1 DTS HD , ETA & MSRP. :twisted:

 

[sailer]

WTF is it with the forum and 'you title has too many upper case characters' BS !?

yeah. i tried to reply to (sailor) earlier and got the same.?

You mean I'm not the only one who has had troubles? I've been wondering if something was going wrong with my keyboard or something, except the trouble was only on Tom's and nowhere else.

 

[sirheck]

except the trouble was only on Tom's and nowhere else.

seems so.?

i am in the middle of reinstalling windows. 8O

not blaming (thg).

i got a virus or my hdd is failing :?

 

[sailer]

Well, Tom's is the only place the the trouble has occured. It hasn't happened in Word, Excel, Power point, a couple of groups that I'm in or anyplace else. I'm not including my occasional spelling mistakes. I do them all by myself. I'm not sure where the problem is or why it is, only that it is.

 

[rammedstein]

I think according to all of atis plans (including the originals), the rv610 and rv630 were both meant to be 65nm and the r600 was meant to be 80...

 

[TheGreatGrapeApe]

True, but it seems like AMD may have pulled another R520->580 type move where the issues with 80nm (if rumours are true) precipitated a rush on the 65nm production, which if it did well, may explain why they are skipping 80nm.

That would be quite a surprise for the R600 to launch at 65nm. I still think it's doubtful that the launch will be 65nm, but it could mean a very quick refresh to 65nm.

As always, only time will tell if there's any credability to the stories.

 

[Anoobis]

Keep in mind guys that the article states "supports GDDR4". That doesn't mean they will be using it right away. IIRC the G86 supports GDDR4 but so far all info I've seen indicates models with GDDR3 at launch.

EDIT: And TGGA...this might interest you.

 

[sailer]

Keep in mind guys that the article states "supports GDDR4". That doesn't mean they will be using it right away. IIRC the G86 supports GDDR4 but so far all info I've seen indicates models with GDDR3 at launch.

Good point. That leaves us back at the wait and see what the final specs and benches show. I'm not going to get overly excited one way or another until the card is actually out.

 

[darkstar782]

GDDR4 isnt instantly better.

The 8800GTX uses 900MHz/1.8GHz GDDR3, while the x1950XTX uses 1GHz/2GHz GDDR4.

Thats only a 200MHz difference, and lets not forget that GDDR4 has terrible latencies compared to GDDR3.

I know that the 8800GTX has a 50% wider bus, but I'm just showing the cards with the fastest example of GDDR3 and GDDR4 currently availible.

If you dont believe me, compare an x1900XTX to an x1950XTX, the only difference being 775MHz/1.55GHz GDDR3 vs 1GHz/2GHz GDDR4. Due in part to the latencies of GDDR4, the advantage of the x1950XTX over the x1900XTX is tiny when you condiser the 30% memory clock increase!

I'd much rather have 1GHz/2GHz GDDR3 than 1GHz/2GHz GDDR4, and GDDR3 of that speed is attainable on almost every 8800GTX, while some ship with this speed @ stock.

Unless ATi DO use one of the much faster varieties of GDDR4, I'll be as unimpressed as I was by its addition to the x1950XTX.

 

[LAN_deRf_HA]

I believe I read something about the newer gddr4 having much lower latencies... but then again it was the inq.

 

[TheGreatGrapeApe]

EDIT: And TGGA...this might interest you.

Yeah actually I saw that at B3D, and the thing that interested me the most was the mention of Quantum Effects. To me that implies that there will be VPU physics available to just a single card/chip solution, and that you can load balance on the GF8 series so that some portion of the sream processors et al can be used for physics and others for graphics.

It'll be interesting to see how that fleshes out.

I think the GF86 in a laptop will be very interesting. I'm most interested to see whether ATi brings audio over HDMI and UVD stuff to their laptop offerings, but knowing tha the GF86 is there as another option is very comforting. Just need to see these suckers in action.

 

[TheGreatGrapeApe]

I believe I read something about the newer gddr4 having much lower latencies... but then again it was the inq.

Nah darkstar is right, the ones shiping now have higher latencies, but they are not a factor of just GDDR4 itself, but of the modles used and the speeds they are running at.

One of the other advantages of GDDR4 is slightly lower power consumption, which is nice for laptops, but like I said I'd prefer slower/wider memory that uses less power overall.

 

[bitter]

http://www.theinquirer.net/default.aspx?article=38292 all r600 will be 65 nm