Anyone still using a HD 2900XT or Pro?

[Rebo&Zooty]

Except nVidia's products far from suck in the current situation. Hell, nVidia's generation old product outperforms all but the very top ATi card from the current generation...I would hardly consider that sucking.

yeah but they are all defective and prone to dieing after repeted heat cycling(google it urself) the g84/g86 and g92 all are FLAWED not the cards the CHIPS/GPU packege, something to do with the bonding or some such, fail rates are climbing along the bell curve :/

sad but fact is that perf means dick if the card dies and you gotta rma it!!!!

 

[Aceman.au]

I think my friends ran hot because he had it overclocked to run crysis... But he had a big PSU to run his PC...

 

[OzzmanFloyd120]

yeah but they are all defective and prone to dieing after repeted heat cycling(google it urself) the g84/g86 and g92 all are FLAWED not the cards the CHIPS/GPU packege, something to do with the bonding or some such, fail rates are climbing along the bell curve :/

sad but fact is that perf means dick if the card dies and you gotta rma it!!!!

They've isolated the problem, though and have started steps to correct it... in fact if you have it happen once, you'll get it replaced (and most likely upgraded) in the RMA with a new GPU that won't have that issue.

 

[Rebo&Zooty]

but the current batch are ALL flawed, and i have had multi 8800gt's replaced finnely got a referbed 8800gts as replacement, but since they are all flawed, im on borrowed time waiting for this one to fail so i can rma yet again.......wouldnt have been a problem if there had been a reasonable alterntive to the 8800gt when i got my card( the 3870 was not out yet)

seems that companys like XFX are moving to ATI or at least now partialy moving, I can see them replacing nvidia cards that have failed multi times with ati cards if the users willing to take one, since it would endup being cheaper then continuing to rma the g92 cards over and over.

and nv still is blaming everybody but themselves, when it was THEIR FAULT they screwed up, if they would just admit that, do a optional recall, exchange all the flawed cards for something thats not flawed then maby i wouldnt be so dissapointed(and maby the rest of the ppl that know about this issue wouldnt be dissapoited either.)

 

[qubit]

ChipGate scandel

yeah but they are all defective and prone to dieing after repeted heat cycling(google it urself) the g84/g86 and g92 all are FLAWED not the cards the CHIPS/GPU packege, something to do with the bonding or some such, fail rates are climbing along the bell curve :/

sad but fact is that perf means dick if the card dies and you gotta rma it!!!!

Indeed. Charlie D from The Inquirer has uncovered the whole sordid truth; he even cut up a brand new MacBook to prove that nvidia are lying! Hence nvidia's products suck. I linked to his articles in this forum post: http://forums.techpowerup.com/showpost.php?p=1138421&postcount=10

It's certainly possible that my 9800 GX2 may die because of this problem. At least I paid about half price for it and it will likely last until it's time to replace with a better card next year sometime.

 

[imperialreign]

just a mention from a 4870 user here - anyone complaining about high-temps with a 4870 . . . either adjust your fan speeds with ATT or re-flash the card with better fan defaults.


Sure, they ran warm right out of the box, but with the fan only blowing at 30% and only 20% of the time, they'll get warm. Re-flash to a default of 50% at 100% of the time, and they'll stay much cooler, and still be barely audible. Hell, with my current fan mapping, neither GPU has ever broken 70C, even pushing the maximum OC for this BIOS with no hard-modding (845/1051) for hours of gaming at 1920x1200.

Their predecessors, though, were heat-deamons. The HD3870s (which supposedly ran cooler than the 2900XTs) that I had constantly rode 100C no matter what I could come up with to cool them (aside from liquid cooling, which I never tested).

 

[mrw1986]

but the current batch are ALL flawed, and i have had multi 8800gt's replaced finnely got a referbed 8800gts as replacement, but since they are all flawed, im on borrowed time waiting for this one to fail so i can rma yet again.......wouldnt have been a problem if there had been a reasonable alterntive to the 8800gt when i got my card( the 3870 was not out yet)

seems that companys like XFX are moving to ATI or at least now partialy moving, I can see them replacing nvidia cards that have failed multi times with ati cards if the users willing to take one, since it would endup being cheaper then continuing to rma the g92 cards over and over.

and nv still is blaming everybody but themselves, when it was THEIR FAULT they screwed up, if they would just admit that, do a optional recall, exchange all the flawed cards for something thats not flawed then maby i wouldnt be so dissapointed(and maby the rest of the ppl that know about this issue wouldnt be dissapoited either.)

Indeed. Charlie D from The Inquirer has uncovered the whole sordid truth; he even cut up a brand new MacBook to prove that nvidia are lying! Hence nvidia's products suck. I linked to his articles in this forum post: http://forums.techpowerup.com/showpost.php?p=1138421&postcount=10

It's certainly possible that my 9800 GX2 may die because of this problem. At least I paid about half price for it and it will likely last until it's time to replace with a better card next year sometime.

Yep because the Inquirer is always right...lol

They aren't all defective, I've ran 2x 8800GT in SLI for months, an 8800GTX, and now a GTX280. No problems. My friend runs Tri-SLI 8800GTX's and hasn't had a problem. I just think you guys believe everything you read.

 

[newtekie1]

yeah but they are all defective and prone to dieing after repeted heat cycling(google it urself) the g84/g86 and g92 all are FLAWED not the cards the CHIPS/GPU packege, something to do with the bonding or some such, fail rates are climbing along the bell curve :/

sad but fact is that perf means dick if the card dies and you gotta rma it!!!!

That only affected G84M cards(8400M and 8600M). No G92, G84, G86, or G94 cards were affected by that heat issue.

It also wasn't all the batches, just few of them, and the issue was only an issue if the core was improperly cooled.

 

[LittleLizard]

using a 2900xt is like using a 3870x2 on a game that has not a crossfire profile.

no, acutally, the 3870 has better performance

 

[JC316]

I was using onea little whil back, but I sold it. Kinda miss it too. I would like to find one at a bargain price since I am looking to move up from my 2600's.

 

[Rebo&Zooty]

Yep because the Inquirer is always right...lol

They aren't all defective, I've ran 2x 8800GT in SLI for months, an 8800GTX, and now a GTX280. No problems. My friend runs Tri-SLI 8800GTX's and hasn't had a problem. I just think you guys believe everything you read.

um, i normaly take all news sites, specly the inq with a huge grain of salt, but the sheer ammount of evedance and even apple backing it......and things i have seen MYSELF, well they are defective, and nvidia are trying to pretend nothings happned and nothing was ever defective by their fault.

and newtekie
http://www.theinquirer.net/category/friction/charlie-vs-nvidia

read those articals, the problem is they ALL USE THE SAME PACKAGE PROCESS, so they ARE ALL DEFECTIVE, i have gone thru 3 8800gt's(all cooked themselves) and have had friends with same problems, most have moved to ati again just to get away from the constant rma's.

i got an 8800gts as replacement for my last card, the coolers better so it probably wont die from heat cycling as quickly, but its still susceptible due to the component's used.

from some of what i understand its the bonding agent/process used, after heating/cooling/heating/cooling it breaks down/wares out be it from expansion/contraction or from the heat/cool process, either way its still a flaw in the design, one that HAS effected returns, why do you think you are now starting to see previously nvidia only company's like XFX offering ATI based products now?
an easy to

 

[newtekie1]

So your only evidence is a site write by an anti-nvidia author?

Don't believe everything you read.

None of my G92, G94, or G84 cards have failed. Of course I don't let them overheat either.

All of them might use the same package process, but the desktop cards do not get hot enough to cause an issue. If your cards failed, it wasn't because of this issue. The only reason the laptop cards had the issue is because the laptop manufactures screwed up and let the cards overheat. It isn't uncommon for a laptop card to hit 100C, it is rare for a desktop card to see that temp.

And you have to be one big fanboy to believe XFX and others are moving to ATi because of this problem. They are moving to ATi because ATi finally has a competitive product worth selling.

Anyway, this is all off topic, so lets get back to the original topic, the HD2900 series.

 

[[I.R.A]_FBi]

So your only evidence is a site write by an anti-nvidia author?

Don't believe everything you read.

None of my G92, G94, or G84 cards have failed. Of course I don't let them overheat either.

And you have to be one big fanboy to believe XFX and others are moving to ATi because of this problem. They are moving to ATi because ATi finally has a competitive product worth selling.

and nvidia doesnt?

 

[newtekie1]

and nvidia doesnt?

Obviously they do, otherwise XFX wouldn't keep selling them.

 

[[I.R.A]_FBi]

ok, checking.

 

[mrhuggles]

i have a 2900xt 1gb, i like it but i had to make a few changes to the bios

backed off the memory to 900mhz [was artifacting in RO, and crashing in fallout 3 but that was the only problems ever]

upped the gpu a little bit, to 750mhz in 3d mode

droped the clocks in 2d mode to 200mhz core and 300mhz ram to save power [tried to drop the 2d voltage to .7v but that made windows not able to boot lol so default voltage but it prolly uses way less power that way]

and finaly, i changed the fan speeds so that the fan speeds would steadily increase with a peak of 70c [under 100% load in normal ambient temps it hits about 80, which imho is alot healther than the way they had it, it would stay around 90c]

i dont think i really lost any performance droping the ram to 900mhz and if i did i definatly made it up upping the gpu to 750mhz yay, i dont get why people keep saying the 3870 is faster, the 3870 is slower per clock, and its default clocks are higher, but its slower per clock!
also the 2900xt has 700 million transistors, 3870 has 666 million transistors... wtf?

 

[a_ump]

gotta say i support Rebo&Zooty on his statements or at least point, i bought 2 XFX 8800GT's last january 1 died by march the other in august cooling wasn't an issue cause i had always checked the temps since they had single slot coolers, 46-60 celcius was where they always were. i was pissed but hey shit happens.

 

[qubit]

Believe what Charlie writes

Yep because the Inquirer is always right...lol

They aren't all defective, I've ran 2x 8800GT in SLI for months, an 8800GTX, and now a GTX280. No problems. My friend runs Tri-SLI 8800GTX's and hasn't had a problem. I just think you guys believe everything you read.

Charlie certainly is. Did you read that article where they cut up that MacBook? In short, nvidia claimed that none of those had chips made with the high lead process; eutectic instead. However, after cutting up the graphics chip in it and testing it with special equipment, they found out that it was high lead and that nvidia are lying. You can bet that nvidia would nail them for libel if it wasn't true. How much more evidence do you need?

I'm glad your nvidia cards haven't failed and I hope they don't for your sake, but according to the evidence, they have this manufacturing defect. If I remember correctly, the GTX 280 is clear of this problem.

I wonder how much that investigation must have cost The Inq? Can't have been cheap, that's for sure.

 

[qubit]

Off Topic / On Topic

Just for the record, I don't mind if my thread goes off topic. Conversations are like that anyway when people talk face to face.

 

[qubit]

i have a 2900xt 1gb, i like it but i had to make a few changes to the bios...
....................

...i dont get why people keep saying the 3870 is faster, the 3870 is slower per clock, and its default clocks are higher, but its slower per clock!
also the 2900xt has 700 million transistors, 3870 has 666 million transistors... wtf?

Interesting tweaks there mrhuggles. A stock 3870 generally benchmarks like a stock 2900XT within 5% or so. Just a little under at times, especially at high resolution, that's why I bought the 2900XT 1GB instead of a 3870 at the time (little did I know that the noise would be such a serious problem).

ATI tweaked the core to produce less heat (hence quieter fan and the chip can take more heat) and also reduced the data bus from 512-bit to 256-bit to reduce cost and complexity. That's why the 3870 is slower clock for clock than the 2900XT and has less transistors.

A 2900XT can perform quite respectably if it can be water cooled and clocked very high, like another member on this thread did.

 

[Atnevon]

Interesting tweaks there mrhuggles. A stock 3870 generally benchmarks like a stock 2900XT within 5% or so. Just a little under at times, especially at high resolution, that's why I bought the 2900XT 1GB instead of a 3870 at the time (little did I know that the noise would be such a serious problem).

ATI tweaked the core to produce less heat (hence quieter fan and the chip can take more heat) and also reduced the data bus from 512-bit to 256-bit to reduce cost and complexity. That's why the 3870 is slower clock for clock than the 2900XT and has less transistors.

A 2900XT can perform quite respectably if it can be water cooled and clocked very high, like another member on this thread did.

Wait, isn't that the guy that has a story of how his water cooling leaked and fried his compy? Yet, under some miracle, his card still worked? That was a good laugh worthy of a LuLz.

 

[Wile E]

Could it beat its nemesis, the 8800 GTX running at stock, with those clocks? The heat output must be terrific at those speeds...

The heat output is fine with water cooling. At 1.35V, it sees a max of 45C on a single 120mm rad. And yes, at those speeds, it beats a GTX. But still, it's nemesis was not the GTX anyway. It was built to compete with the GTS 640, and it did that very well (just don't use a lot of AA. lol).

Wait, isn't that the guy that has a story of how his water cooling leaked and fried his compy? Yet, under some miracle, his card still worked? That was a good laugh worthy of a LuLz.

Yeah, that was me, and it was a TEC setup. You can read the whole story here: http://forums.techpowerup.com/showthread.php?t=38153

The card was not water cooled at that time.

 

[Rebo&Zooty]

only you WileE could destroy a perfectly good 6000(if u hated it that much u could have just given it to me!!!), then get a nice shiny new cpu and burn it to cinders because he couldnt wait for a relay kit

 

[Wile E]

only you WileE could destroy a perfectly good 6000(if u hated it that much u could have just given it to me!!!), then get a nice shiny new cpu and burn it to cinders because he couldnt wait for a relay kit

You'd cry at the amount of stuff I have blown up since then. lol. You'll be happy to know that the 6400+ is still living happily to this day tho. lol

 

[Rebo&Zooty]

what u didnt rip its core out or burn it up to