• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

X800 pic

Its nothing special :( . Hopefully third party companys will make things a bit more exciting
 
I'm waiting for the
"I smell ATI and it smells..." quote at the top of the [H] main page.
 
I would not be surprised if that was the reference design. If that is the case, then the card is definitely not overclocked (IE: overvolted at stock)

No cooling on memory, just like most reference ATi cards. Conservative one slot cooling, with a nice big fan that pushes out decent air with very little noise.

One molex sure would be nice. It might leave the recommended powersupplies in the under 430 or maybe even under 400 watt range.


Things don't need to be hot and powerhungry to be good, as anyone with a 30 watt Opteron EE or HE will tell you that it kicks butt on a P4, especially if you've got a dual or quad system.
 
Dang it. The wait for the NDA to expire on the R420 is killing me. I want to see some numbers, damn it!
 
its the same damn heatsink on a 9800xt, but the card is a little longer so the heatsink wont reach out far enough to cool the memory
 
Im guessing that its the midrange card. You would think that the ram would need to be cooled at a high clock speed. Or the ram is heat sensitive, the core might get too hot and overheat the ram!
 
i think the card have one molex and one fdd conector...

the yellow one is the fdd...
 
i suggest some pplz take a trip over to B3D if ya want a tad more info than the speculation going on over here...granted some of them over there are prolly speculating as well, but i find the read up over there much more interesting ;p

http://www.beyond3d.com/forum/viewtopic.php?t=11781&postdays=0&postorder=asc&start=0

some seem to think that its not even a Pro or XT but the SE version. also the yellow connector next to the molex is suggested to be "something else" =P anyway head on over there and have a read. interesting to say the least
 
also the yellow connector next to the molex is suggested to be "something else"
Click to expand...

It's not suggested to be something else, it is something else, Dave B has the card you know
 
Originally posted by evilchris
It's the SLI connector!
Click to expand...

One can only hope. ;)

One can also only hope that that is the SE model with it's ~ 800MHz 128bit ram.....which is the only way I can really jusitfy the lack of ramsinks or a larger cooler. I guess in a sense they're devoting the entire cooler to the VPU without transfering any heat back and forth.....but that ram is gonna need SOME cooling on the XT. ;) Seeing as the ram on the SE has been continuously said to be 800MHz, I'm thinkin' that it might go like this: SE = 800MHz, Pro = 1GHz, XT = 1.2GHz. That would set ATI up to be running the core in sync with the ram at 600MHz, which will be interesting in itself.

I'm really glad at used their Low-K technology on the core for the X800, it helped keep it under 200 million transistors, and maintained the single slot cooling solution AND only one molex connector.

Just because of the physical asthetics and functionality of the card I have a feeling that the X800 is gonna hit it big, but a lot of people are gonna be upset if they don't throw in that second DVI port for them. ;) :p

Just remember how much Low-K helped out the 9600XT.....look at the core clock.....now tack on another 100MHz and imagine it "powering" a 16x1 pipeline, backed by 1.2GHz 256bit GDDR3 ram. :D I can hardly wait.....it's so exciting. :D
 
Originally posted by SnakEyez187
It's not suggested to be something else, it is something else, Dave B has the card you know
Click to expand...


oops guess i shoulda made myself more clear...its been a long day before i posted that =P anyway, certainly interesting reading over @ B3D for sure. and i didnt quite realize DaveB had the card in hand...im not a regular over there :\ certainly these last few days before the cards come out from behind their NDA cloak shall be interesting :D
 
Originally posted by B1zz
certainly these last few days before the cards come out from behind their NDA cloak shall be interesting :D
Click to expand...

For sure, but I'll be glad when it's all finally over and there's no more argueing over what the card "will" be or what will beat what. :rolleyes: ;)
 
Originally posted by cornelious0_0


I'm really glad at used their Low-K technology on the core for the X800, it helped keep it under 200 million transistors, and maintained the single slot cooling solution AND only one molex connector.
Click to expand...
low-k has absolutely nothing to do with number of transistors. it is about cutting down the internal leakage, to lower heat and power use, but it has downsides, Nvidia thinks it is dangerous, IBM no longer offers it to customers because in their fabs it kills yeilds, and we all know how well intel did with the 90 nano strained silicon low-k process on prescott. the fact that ATI manages good yeilds at TSMC with low-k is incredible.
 
Originally posted by Merlin45
low-k has absolutely nothing to do with number of transistors. it is about cutting down the internal leakage, to lower heat and power use, but it has downsides, Nvidia thinks it is dangerous, IBM no longer offers it to customers because in their fabs it kills yeilds, and we all know how well intel did with the 90 nano strained silicon low-k process on prescott. the fact that ATI manages good yeilds at TSMC with low-k is incredible.
Click to expand...

well i also liken the 9600XT run as a "test run" for the low-k process, that card has been out in retail for awhile now and the fact that its done so well makes me look forward to the R420 offerings...as mature as the process is by now over at TSMC im sure yield is prolly no longer an issue as much anymore. It also tends to make me believe the _possibilty_ of 600MHz core freq's on the new X800XT coming up before long...the only way to know the truth is to wait it out a bit longer and we'll all know =P
 
Originally posted by Merlin45
low-k has absolutely nothing to do with number of transistors. it is about cutting down the internal leakage, to lower heat and power use, but it has downsides, Nvidia thinks it is dangerous, IBM no longer offers it to customers because in their fabs it kills yeilds, and we all know how well intel did with the 90 nano strained silicon low-k process on prescott. the fact that ATI manages good yeilds at TSMC with low-k is incredible.
Click to expand...

There are differences between the low-k TSMC is using compared to IBM and Intel
 
So many armchair "engineers" in here tossing low-k around when they don't even know wtf it is/means. lol. :rolleyes:
 
and we all know how well intel did with the 90 nano strained silicon low-k process on prescott.
Click to expand...
Prescott does use a "low-k" (whether it's really got a dielectric constant of <3.0 I'm not sure) dielectric with strained silicon, but the newer revisions appear to be running significantly cooler. Intel's sucess with the process is going to be determined over the long term not based on the intro review samples.

Comparing Prescott to Northwood is not representative of a process shrink because Prescott is much larger than Northwood in terms of transistor count. There is enough availibility of P4E's in the wild to suppose that yields are quite acceptable, and the
2.4A article indicates thermal output is down considerably.

well i also liken the 9600XT run as a "test run" for the low-k process, that card has been out in retail for awhile now and the fact that its done so well makes me look forward to the R420 offerings...as mature as the process is by now over at TSMC im sure yield is prolly no longer an issue as much anymore. It also tends to make me believe the _possibilty_ of 600MHz core freq's on the new X800XT coming up before long...the only way to know the truth is to wait it out a bit longer and we'll all know =P
Click to expand...

Right. The TSMC black diamond low-k is used in conjunction with "fluorinated glass" on their 130nm process. Still, 600MHz seems like an upper bound to me with a die size of 180 million transistors. Possibly yields will be improved by crippling one or two of the 4 pipeline subunits to sell as Pro and SE versions.
 
Originally posted by evilchris
So many armchair "engineers" in here tossing low-k around when they don't even know wtf it is/means. lol. :rolleyes:
Click to expand...

Not to mention the people critiquing posts while adding no actual commentary pertinent to the thread
 
Originally posted by evilchris
So many armchair "engineers" in here tossing low-k around when they don't even know wtf it is/means. lol. :rolleyes:
Click to expand...

I'm not saying I know everything about this stuff, but enough to get by. I may have been wrong about one thing I said but please.....if you're so learned on the subject then educate us. I'm not meaning for this to sound like a threat but really.....anyone can walk in and said what you just said. Got anything to back it up??? I don't mean anything by this of course, I actually hope you do have some actual facts to back yourself up.....I like learning new things, even on smaller scales.
 
Originally posted by cornelious0_0
I'm not saying I know everything about this stuff, but enough to get by. I may have been wrong about one thing I said but please.....if you're so learned on the subject then educate us. I'm not meaning for this to sound like a threat but really.....anyone can walk in and said what you just said. Got anything to back it up??? I don't mean anything by this of course, I actually hope you do have some actual facts to back yourself up.....I like learning new things, even on smaller scales.
Click to expand...

Exactly. Well said. I would like to learn, too.
 
Originally posted by planetary
Exactly. Well said. I would like to learn, too.
Click to expand...

Heh, it's been awhile since I was in school and since we've had a major card release (other then earlier this month) and I haven't really had to work my brain at all.....it's just been the same facts and information for the past few months, and now ATI and Nvidia have to bring all this "new stuff" to the table.....bah.

Ah well, you can't coast on forever. :cool:
 
Prescott does use a "low-k" (whether it's really got a dielectric constant of <3.0 I'm not sure) dielectric with strained silicon
Click to expand...
Update:

Apparently Intel is using ASM's "Aurora" dielectric, which is a carbon doped silicon oxide (organosilicate). It appears to be a genuine low-k material.

According to chipworks, TSMC and Sony/Toshiba (The PSX Emotion Engine is being produced on 90nm low-k) use Applied Materials' "Black Diamond" dielectric while TI uses Novellus' "Coral".
 
I have to agree with what some people have been saying here...the lack of direct RAM cooling and the presence of only a single molex makes this picture questionable. I first surmised that it was the SE, although only time will tell. Two peripheral reasons for this: the SE is speculated to use DDR (not GDDR3), which doesn't require extra cooling, and its power profile matches the need for a single molex.

As I have a physics degree, I'll attempt to shed some basic light on what low-k is all about. It refers to low-k dielectrics, a technology which was indeed used on the 9600XT. K is the dielectric constant, usually in reference to K = 1 (vacuum/air). For instance, water has a dielectric constant of 80...a shoddy cable might be around 6. It's important to realize that both low-k and high-k technologies will be needed. As oxide gates diminish in size, they will eventually lose the ability to function as an insulator (due basically to the an overabundance of current tunneling through the space), requiring high-k (>~4) dielectrics for proper capacitance coupling. On the other hand, there are instances were you want less capacitance and cross-talk (signal mixing)...mainly when dealing with multi-layer (PCBs) and metal lines, in terms of insulation. As should be obvious, a vacuum (K = 1) is not conductive at all; hence, low-k dielectrics act as insulators. Essentially, low-k dielectrics allow signals to move faster and to make tighter fits, while high-k dielectrics can be used for the gates themselves. The low-k technology basically means faster speeds and lower prices, it should really take off with the 90 nm process...but when you get down a good deal below that, you will be seeing more of high-k in the tech news.
 
Originally posted by planetary
Dang it. The wait for the NDA to expire on the R420 is killing me. I want to see some numbers, damn it!
Click to expand...

When does it expire?
 
Originally posted by cornelious0_0
I'm not saying I know everything about this stuff, but enough to get by. I may have been wrong about one thing I said but please.....if you're so learned on the subject then educate us. I'm not meaning for this to sound like a threat but really.....anyone can walk in and said what you just said. Got anything to back it up??? I don't mean anything by this of course, I actually hope you do have some actual facts to back yourself up.....I like learning new things, even on smaller scales.
Click to expand...

Here are the basics:

http://nepp.nasa.gov/index_nasa.cfm/934/

I didn't say *I* was learned on the subject. I can spot BS a mile away though and am sick of people getting hard-ons over "low-k" as if it is going to make their games better. Are you one of the people where if someone says a game has bugs you say " I'd like to see you code better" ?
 
Originally posted by lorcani
When does it expire?
Click to expand...

yeah, anybody have an idea? I think I heard Apr 26th and May 4th as 2 possible dates, but I'm not at all sure about those.
 
Originally posted by ZenOps
DDR3 runs much cooler than DDR, due to it using 1.8Volts instead of 2.5
Click to expand...

This is true, but only when making the comparison at identical frequencies. It's fairly common for 9800 Pro DDR1 to hit 730 MHz without any cooling, even considering the adjacent heat from the pitifully-cooled stock core, while only being rated for about 714 MHz and run at 680 MHz default. DDR2/3 is innately less efficient than DDR1, even though it can hit higher speeds. Besides that, the details are still up in the air...DDR1 can hit 1 GHz though, albeit not very well.

But the big point for me is that DDR1 is widely available and inexpensive, making it far more fitting for the SE (and has been slated for it). Also, the SE is supposedly using only 128-bit RAM, which possibly extends the clock range. GDDR3 running at lower clocks might not need extra cooling, but it definitely does at higher clocks (XT anyway, probably Pro as well). The whole point of GDDR3 is for higher memory clocks with less power usage...it seems counter-productive to use that for your budget enthusiast card.

It is possible that it could be GDDR3, but it would still NOT be on the XT...the lack of cooling still indicates lower memory clocks. But it just may be that this picture is inaccurate in some fashion, but I suppose we shall find out and then discuss it.
 
Thank you for what has been said by those of you who have said it.....it's much appreciated.

Are you one of the people where if someone says a game has bugs you say " I'd like to see you code better" ?
Click to expand...

No, not at all.
 
Originally posted by Sandman
yeah, anybody have an idea? I think I heard Apr 26th and May 4th as 2 possible dates, but I'm not at all sure about those.
Click to expand...

I just googled it, and this is what I came up with:

http://www.theinquirer.net/?article=15377

"This means that you should be able to buy one on the 4th of May but let's wait and see if ATI can deliver on that day."

So, the NDA has to be lifted (well, hopefully) before then.
 
You must log in or register to reply here.
Back
Top